Index: stable/10/contrib/libarchive/cat/test/test_empty.zst.uu =================================================================== --- stable/10/contrib/libarchive/cat/test/test_empty.zst.uu (nonexistent) +++ stable/10/contrib/libarchive/cat/test/test_empty.zst.uu (revision 324418) @@ -0,0 +1,4 @@ +begin 644 test_empty.zst +-*+4O_010`0``F>G840`` +` +end Index: stable/10/contrib/libarchive/cat/test/test_empty_zstd.c =================================================================== --- stable/10/contrib/libarchive/cat/test/test_empty_zstd.c (nonexistent) +++ stable/10/contrib/libarchive/cat/test/test_empty_zstd.c (revision 324418) @@ -0,0 +1,41 @@ +/*- + * Copyright (c) 2017 Sean Purcell + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES + * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. + * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, + * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT + * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF + * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ +#include "test.h" + +DEFINE_TEST(test_empty_zstd) +{ + const char *reffile = "test_empty.zst"; + int f; + + extract_reference_file(reffile); + f = systemf("%s %s >test.out 2>test.err", testprog, reffile); + if (f == 0 || canZstd()) { + assertEqualInt(0, f); + assertEmptyFile("test.out"); + assertEmptyFile("test.err"); + } else { + skipping("It seems zstd is not supported on this platform"); + } +} Property changes on: stable/10/contrib/libarchive/cat/test/test_empty_zstd.c ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: stable/10/contrib/libarchive/cat/test/test_expand.zst.uu =================================================================== --- stable/10/contrib/libarchive/cat/test/test_expand.zst.uu (nonexistent) +++ stable/10/contrib/libarchive/cat/test/test_expand.zst.uu (revision 324418) @@ -0,0 +1,4 @@ +begin 644 test_expand.zst +J*+4O_010Z0``8V]N=&5N=',@;V8@=&5S=%]E>'!A;F0N>G-T+@J;23#F +` +end Index: stable/10/contrib/libarchive/cat/test/test_expand_zstd.c =================================================================== --- stable/10/contrib/libarchive/cat/test/test_expand_zstd.c (nonexistent) +++ stable/10/contrib/libarchive/cat/test/test_expand_zstd.c (revision 324418) @@ -0,0 +1,41 @@ +/*- + * Copyright (c) 2017 Sean Purcell + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES + * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. + * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, + * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT + * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF + * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ +#include "test.h" + +DEFINE_TEST(test_expand_zstd) +{ + const char *reffile = "test_expand.zst"; + int f; + + extract_reference_file(reffile); + f = systemf("%s %s >test.out 2>test.err", testprog, reffile); + if (f == 0 || canZstd()) { + assertEqualInt(0, f); + assertTextFileContents("contents of test_expand.zst.\n", "test.out"); + assertEmptyFile("test.err"); + } else { + skipping("It seems zstd is not supported on this platform"); + } +} Property changes on: stable/10/contrib/libarchive/cat/test/test_expand_zstd.c ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: stable/10/contrib/libarchive/cpio/bsdcpio.1 =================================================================== --- stable/10/contrib/libarchive/cpio/bsdcpio.1 (revision 324417) +++ stable/10/contrib/libarchive/cpio/bsdcpio.1 (revision 324418) @@ -1,431 +1,436 @@ .\" Copyright (c) 2003-2007 Tim Kientzle .\" All rights reserved. .\" .\" Redistribution and use in source and binary forms, with or without .\" modification, are permitted provided that the following conditions .\" are met: .\" 1. Redistributions of source code must retain the above copyright .\" notice, this list of conditions and the following disclaimer. .\" 2. Redistributions in binary form must reproduce the above copyright .\" notice, this list of conditions and the following disclaimer in the .\" documentation and/or other materials provided with the distribution. .\" .\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE .\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF .\" SUCH DAMAGE. .\" .\" $FreeBSD$ .\" .Dd September 16, 2014 .Dt CPIO 1 .Os .Sh NAME .Nm cpio .Nd copy files to and from archives .Sh SYNOPSIS .Nm .Fl i .Op Ar options .Op Ar pattern ... .Op Ar < archive .Nm .Fl o .Op Ar options .Ar < name-list .Op Ar > archive .Nm .Fl p .Op Ar options .Ar dest-dir .Ar < name-list .Sh DESCRIPTION .Nm copies files between archives and directories. This implementation can extract from tar, pax, cpio, zip, jar, ar, and ISO 9660 cdrom images and can create tar, pax, cpio, ar, and shar archives. .Pp The first option to .Nm is a mode indicator from the following list: .Bl -tag -compact -width indent .It Fl i Input. Read an archive from standard input (unless overridden) and extract the contents to disk or (if the .Fl t option is specified) list the contents to standard output. If one or more file patterns are specified, only files matching one of the patterns will be extracted. .It Fl o Output. Read a list of filenames from standard input and produce a new archive on standard output (unless overridden) containing the specified items. .It Fl p Pass-through. Read a list of filenames from standard input and copy the files to the specified directory. .El .Pp .Sh OPTIONS Unless specifically stated otherwise, options are applicable in all operating modes. .Bl -tag -width indent .It Fl 0 , Fl Fl null Read filenames separated by NUL characters instead of newlines. This is necessary if any of the filenames being read might contain newlines. .It Fl A (o mode only) Append to the specified archive. (Not yet implemented.) .It Fl a (o and p modes) Reset access times on files after they are read. .It Fl B (o mode only) Block output to records of 5120 bytes. .It Fl C Ar size (o mode only) Block output to records of .Ar size bytes. .It Fl c (o mode only) Use the old POSIX portable character format. Equivalent to .Fl Fl format Ar odc . .It Fl d , Fl Fl make-directories (i and p modes) Create directories as necessary. .It Fl E Ar file (i mode only) Read list of file name patterns from .Ar file to list and extract. .It Fl F Ar file , Fl Fl file Ar file Read archive from or write archive to .Ar file . .It Fl f Ar pattern (i mode only) Ignore files that match .Ar pattern . .It Fl H Ar format , Fl Fl format Ar format (o mode only) Produce the output archive in the specified format. Supported formats include: .Pp .Bl -tag -width "iso9660" -compact .It Ar cpio Synonym for .Ar odc . .It Ar newc The SVR4 portable cpio format. .It Ar odc The old POSIX.1 portable octet-oriented cpio format. .It Ar pax The POSIX.1 pax format, an extension of the ustar format. .It Ar ustar The POSIX.1 tar format. .El .Pp The default format is .Ar odc . See .Xr libarchive-formats 5 for more complete information about the formats currently supported by the underlying .Xr libarchive 3 library. .It Fl h , Fl Fl help Print usage information. .It Fl I Ar file Read archive from .Ar file . .It Fl i , Fl Fl extract Input mode. See above for description. .It Fl Fl insecure (i and p mode only) Disable security checks during extraction or copying. This allows extraction via symbolic links, absolute paths, and path names containing .Sq .. in the name. .It Fl J , Fl Fl xz (o mode only) Compress the file with xz-compatible compression before writing it. In input mode, this option is ignored; xz compression is recognized automatically on input. .It Fl j Synonym for .Fl y . .It Fl L (o and p modes) All symbolic links will be followed. Normally, symbolic links are archived and copied as symbolic links. With this option, the target of the link will be archived or copied instead. .It Fl l , Fl Fl link (p mode only) Create links from the target directory to the original files, instead of copying. .It Fl Fl lrzip (o mode only) Compress the resulting archive with .Xr lrzip 1 . In input mode, this option is ignored. .It Fl Fl lz4 (o mode only) Compress the archive with lz4-compatible compression before writing it. In input mode, this option is ignored; lz4 compression is recognized automatically on input. +.It Fl Fl zstd +(o mode only) +Compress the archive with zstd-compatible compression before writing it. +In input mode, this option is ignored; zstd compression is recognized +automatically on input. .It Fl Fl lzma (o mode only) Compress the file with lzma-compatible compression before writing it. In input mode, this option is ignored; lzma compression is recognized automatically on input. .It Fl Fl lzop (o mode only) Compress the resulting archive with .Xr lzop 1 . In input mode, this option is ignored. .It Fl Fl passphrase Ar passphrase The .Pa passphrase is used to extract or create an encrypted archive. Currently, zip is only a format that .Nm can handle encrypted archives. You shouldn't use this option unless you realize how insecure use of this option is. .It Fl m , Fl Fl preserve-modification-time (i and p modes) Set file modification time on created files to match those in the source. .It Fl n , Fl Fl numeric-uid-gid (i mode, only with .Fl t ) Display numeric uid and gid. By default, .Nm displays the user and group names when they are provided in the archive, or looks up the user and group names in the system password database. .It Fl Fl no-preserve-owner (i mode only) Do not attempt to restore file ownership. This is the default when run by non-root users. .It Fl O Ar file Write archive to .Ar file . .It Fl o , Fl Fl create Output mode. See above for description. .It Fl p , Fl Fl pass-through Pass-through mode. See above for description. .It Fl Fl preserve-owner (i mode only) Restore file ownership. This is the default when run by the root user. .It Fl Fl quiet Suppress unnecessary messages. .It Fl R Oo user Oc Ns Oo : Oc Ns Oo group Oc , Fl Fl owner Oo user Oc Ns Oo : Oc Ns Oo group Oc Set the owner and/or group on files in the output. If group is specified with no user (for example, .Fl R Ar :wheel ) then the group will be set but not the user. If the user is specified with a trailing colon and no group (for example, .Fl R Ar root: ) then the group will be set to the user's default group. If the user is specified with no trailing colon, then the user will be set but not the group. In .Fl i and .Fl p modes, this option can only be used by the super-user. (For compatibility, a period can be used in place of the colon.) .It Fl r (All modes.) Rename files interactively. For each file, a prompt is written to .Pa /dev/tty containing the name of the file and a line is read from .Pa /dev/tty . If the line read is blank, the file is skipped. If the line contains a single period, the file is processed normally. Otherwise, the line is taken to be the new name of the file. .It Fl t , Fl Fl list (i mode only) List the contents of the archive to stdout; do not restore the contents to disk. .It Fl u , Fl Fl unconditional (i and p modes) Unconditionally overwrite existing files. Ordinarily, an older file will not overwrite a newer file on disk. .It Fl V , Fl Fl dot Print a dot to stderr for each file as it is processed. Superseded by .Fl v . .It Fl v , Fl Fl verbose Print the name of each file to stderr as it is processed. With .Fl t , provide a detailed listing of each file. .It Fl Fl version Print the program version information and exit. .It Fl y (o mode only) Compress the archive with bzip2-compatible compression before writing it. In input mode, this option is ignored; bzip2 compression is recognized automatically on input. .It Fl Z (o mode only) Compress the archive with compress-compatible compression before writing it. In input mode, this option is ignored; compression is recognized automatically on input. .It Fl z (o mode only) Compress the archive with gzip-compatible compression before writing it. In input mode, this option is ignored; gzip compression is recognized automatically on input. .El .Sh EXIT STATUS .Ex -std .Sh ENVIRONMENT The following environment variables affect the execution of .Nm : .Bl -tag -width ".Ev BLOCKSIZE" .It Ev LANG The locale to use. See .Xr environ 7 for more information. .It Ev TZ The timezone to use when displaying dates. See .Xr environ 7 for more information. .El .Sh EXAMPLES The .Nm command is traditionally used to copy file hierarchies in conjunction with the .Xr find 1 command. The first example here simply copies all files from .Pa src to .Pa dest : .Dl Nm find Pa src | Nm Fl pmud Pa dest .Pp By carefully selecting options to the .Xr find 1 command and combining it with other standard utilities, it is possible to exercise very fine control over which files are copied. This next example copies files from .Pa src to .Pa dest that are more than 2 days old and whose names match a particular pattern: .Dl Nm find Pa src Fl mtime Ar +2 | Nm grep foo[bar] | Nm Fl pdmu Pa dest .Pp This example copies files from .Pa src to .Pa dest that are more than 2 days old and which contain the word .Do foobar Dc : .Dl Nm find Pa src Fl mtime Ar +2 | Nm xargs Nm grep -l foobar | Nm Fl pdmu Pa dest .Sh COMPATIBILITY The mode options i, o, and p and the options a, B, c, d, f, l, m, r, t, u, and v comply with SUSv2. .Pp The old POSIX.1 standard specified that only .Fl i , .Fl o , and .Fl p were interpreted as command-line options. Each took a single argument of a list of modifier characters. For example, the standard syntax allows .Fl imu but does not support .Fl miu or .Fl i Fl m Fl u , since .Ar m and .Ar u are only modifiers to .Fl i , they are not command-line options in their own right. The syntax supported by this implementation is backwards-compatible with the standard. For best compatibility, scripts should limit themselves to the standard syntax. .Sh SEE ALSO .Xr bzip2 1 , .Xr tar 1 , .Xr gzip 1 , .Xr mt 1 , .Xr pax 1 , .Xr libarchive 3 , .Xr cpio 5 , .Xr libarchive-formats 5 , .Xr tar 5 .Sh STANDARDS There is no current POSIX standard for the cpio command; it appeared in .St -p1003.1-96 but was dropped from .St -p1003.1-2001 . .Pp The cpio, ustar, and pax interchange file formats are defined by .St -p1003.1-2001 for the pax command. .Sh HISTORY The original .Nm cpio and .Nm find utilities were written by Dick Haight while working in AT&T's Unix Support Group. They first appeared in 1977 in PWB/UNIX 1.0, the .Dq Programmer's Work Bench system developed for use within AT&T. They were first released outside of AT&T as part of System III Unix in 1981. As a result, .Nm cpio actually predates .Nm tar , even though it was not well-known outside of AT&T until some time later. .Pp This is a complete re-implementation based on the .Xr libarchive 3 library. .Sh BUGS The cpio archive format has several basic limitations: It does not store user and group names, only numbers. As a result, it cannot be reliably used to transfer files between systems with dissimilar user and group numbering. Older cpio formats limit the user and group numbers to 16 or 18 bits, which is insufficient for modern systems. The cpio archive formats cannot support files over 4 gigabytes, except for the .Dq odc variant, which can support files up to 8 gigabytes. Index: stable/10/contrib/libarchive/cpio/cmdline.c =================================================================== --- stable/10/contrib/libarchive/cpio/cmdline.c (revision 324417) +++ stable/10/contrib/libarchive/cpio/cmdline.c (revision 324418) @@ -1,381 +1,382 @@ /*- * Copyright (c) 2003-2007 Tim Kientzle * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer * in this position and unchanged. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "cpio_platform.h" __FBSDID("$FreeBSD$"); #ifdef HAVE_ERRNO_H #include #endif #ifdef HAVE_GRP_H #include #endif #ifdef HAVE_PWD_H #include #endif #include #ifdef HAVE_STDLIB_H #include #endif #ifdef HAVE_STRING_H #include #endif #include "cpio.h" #include "err.h" /* * Short options for cpio. Please keep this sorted. */ static const char *short_options = "0AaBC:cdE:F:f:H:hI:iJjLlmnO:opR:rtuVvW:yZz"; /* * Long options for cpio. Please keep this sorted. */ static const struct option { const char *name; int required; /* 1 if this option requires an argument */ int equivalent; /* Equivalent short option. */ } cpio_longopts[] = { { "b64encode", 0, OPTION_B64ENCODE }, { "create", 0, 'o' }, { "dereference", 0, 'L' }, { "dot", 0, 'V' }, { "extract", 0, 'i' }, { "file", 1, 'F' }, { "format", 1, 'H' }, { "grzip", 0, OPTION_GRZIP }, { "help", 0, 'h' }, { "insecure", 0, OPTION_INSECURE }, { "link", 0, 'l' }, { "list", 0, 't' }, { "lrzip", 0, OPTION_LRZIP }, { "lz4", 0, OPTION_LZ4 }, { "lzma", 0, OPTION_LZMA }, { "lzop", 0, OPTION_LZOP }, { "make-directories", 0, 'd' }, { "no-preserve-owner", 0, OPTION_NO_PRESERVE_OWNER }, { "null", 0, '0' }, { "numeric-uid-gid", 0, 'n' }, { "owner", 1, 'R' }, { "passphrase", 1, OPTION_PASSPHRASE }, { "pass-through", 0, 'p' }, { "preserve-modification-time", 0, 'm' }, { "preserve-owner", 0, OPTION_PRESERVE_OWNER }, { "quiet", 0, OPTION_QUIET }, { "unconditional", 0, 'u' }, { "uuencode", 0, OPTION_UUENCODE }, { "verbose", 0, 'v' }, { "version", 0, OPTION_VERSION }, { "xz", 0, 'J' }, + { "zstd", 0, OPTION_ZSTD }, { NULL, 0, 0 } }; /* * I used to try to select platform-provided getopt() or * getopt_long(), but that caused a lot of headaches. In particular, * I couldn't consistently use long options in the test harness * because not all platforms have getopt_long(). That in turn led to * overuse of the -W hack in the test harness, which made it rough to * run the test harness against GNU cpio. (I periodically run the * test harness here against GNU cpio as a sanity-check. Yes, * I've found a couple of bugs in GNU cpio that way.) */ int cpio_getopt(struct cpio *cpio) { enum { state_start = 0, state_next_word, state_short, state_long }; static int state = state_start; static char *opt_word; const struct option *popt, *match = NULL, *match2 = NULL; const char *p, *long_prefix = "--"; size_t optlength; int opt = '?'; int required = 0; cpio->argument = NULL; /* First time through, initialize everything. */ if (state == state_start) { /* Skip program name. */ ++cpio->argv; --cpio->argc; state = state_next_word; } /* * We're ready to look at the next word in argv. */ if (state == state_next_word) { /* No more arguments, so no more options. */ if (cpio->argv[0] == NULL) return (-1); /* Doesn't start with '-', so no more options. */ if (cpio->argv[0][0] != '-') return (-1); /* "--" marks end of options; consume it and return. */ if (strcmp(cpio->argv[0], "--") == 0) { ++cpio->argv; --cpio->argc; return (-1); } /* Get next word for parsing. */ opt_word = *cpio->argv++; --cpio->argc; if (opt_word[1] == '-') { /* Set up long option parser. */ state = state_long; opt_word += 2; /* Skip leading '--' */ } else { /* Set up short option parser. */ state = state_short; ++opt_word; /* Skip leading '-' */ } } /* * We're parsing a group of POSIX-style single-character options. */ if (state == state_short) { /* Peel next option off of a group of short options. */ opt = *opt_word++; if (opt == '\0') { /* End of this group; recurse to get next option. */ state = state_next_word; return cpio_getopt(cpio); } /* Does this option take an argument? */ p = strchr(short_options, opt); if (p == NULL) return ('?'); if (p[1] == ':') required = 1; /* If it takes an argument, parse that. */ if (required) { /* If arg is run-in, opt_word already points to it. */ if (opt_word[0] == '\0') { /* Otherwise, pick up the next word. */ opt_word = *cpio->argv; if (opt_word == NULL) { lafe_warnc(0, "Option -%c requires an argument", opt); return ('?'); } ++cpio->argv; --cpio->argc; } if (opt == 'W') { state = state_long; long_prefix = "-W "; /* For clearer errors. */ } else { state = state_next_word; cpio->argument = opt_word; } } } /* We're reading a long option, including -W long=arg convention. */ if (state == state_long) { /* After this long option, we'll be starting a new word. */ state = state_next_word; /* Option name ends at '=' if there is one. */ p = strchr(opt_word, '='); if (p != NULL) { optlength = (size_t)(p - opt_word); cpio->argument = (char *)(uintptr_t)(p + 1); } else { optlength = strlen(opt_word); } /* Search the table for an unambiguous match. */ for (popt = cpio_longopts; popt->name != NULL; popt++) { /* Short-circuit if first chars don't match. */ if (popt->name[0] != opt_word[0]) continue; /* If option is a prefix of name in table, record it.*/ if (strncmp(opt_word, popt->name, optlength) == 0) { match2 = match; /* Record up to two matches. */ match = popt; /* If it's an exact match, we're done. */ if (strlen(popt->name) == optlength) { match2 = NULL; /* Forget the others. */ break; } } } /* Fail if there wasn't a unique match. */ if (match == NULL) { lafe_warnc(0, "Option %s%s is not supported", long_prefix, opt_word); return ('?'); } if (match2 != NULL) { lafe_warnc(0, "Ambiguous option %s%s (matches --%s and --%s)", long_prefix, opt_word, match->name, match2->name); return ('?'); } /* We've found a unique match; does it need an argument? */ if (match->required) { /* Argument required: get next word if necessary. */ if (cpio->argument == NULL) { cpio->argument = *cpio->argv; if (cpio->argument == NULL) { lafe_warnc(0, "Option %s%s requires an argument", long_prefix, match->name); return ('?'); } ++cpio->argv; --cpio->argc; } } else { /* Argument forbidden: fail if there is one. */ if (cpio->argument != NULL) { lafe_warnc(0, "Option %s%s does not allow an argument", long_prefix, match->name); return ('?'); } } return (match->equivalent); } return (opt); } /* * Parse the argument to the -R or --owner flag. * * The format is one of the following: * - Override user but not group * : - Override both, group is user's default group * : - Override user but not group * : - Override both * : - Override group but not user * * Where uid/gid are decimal representations and groupname/username * are names to be looked up in system database. Note that we try * to look up an argument as a name first, then try numeric parsing. * * A period can be used instead of the colon. * * Sets uid/gid return as appropriate, -1 indicates uid/gid not specified. * TODO: If the spec uses uname/gname, then return those to the caller * as well. If the spec provides uid/gid, just return names as NULL. * * Returns NULL if no error, otherwise returns error string for display. * */ const char * owner_parse(const char *spec, int *uid, int *gid) { static char errbuff[128]; const char *u, *ue, *g; *uid = -1; *gid = -1; if (spec[0] == '\0') return ("Invalid empty user/group spec"); /* * Split spec into [user][:.][group] * u -> first char of username, NULL if no username * ue -> first char after username (colon, period, or \0) * g -> first char of group name */ if (*spec == ':' || *spec == '.') { /* If spec starts with ':' or '.', then just group. */ ue = u = NULL; g = spec + 1; } else { /* Otherwise, [user] or [user][:] or [user][:][group] */ ue = u = spec; while (*ue != ':' && *ue != '.' && *ue != '\0') ++ue; g = ue; if (*g != '\0') /* Skip : or . to find first char of group. */ ++g; } if (u != NULL) { /* Look up user: ue is first char after end of user. */ char *user; struct passwd *pwent; user = (char *)malloc(ue - u + 1); if (user == NULL) return ("Couldn't allocate memory"); memcpy(user, u, ue - u); user[ue - u] = '\0'; if ((pwent = getpwnam(user)) != NULL) { *uid = pwent->pw_uid; if (*ue != '\0') *gid = pwent->pw_gid; } else { char *end; errno = 0; *uid = (int)strtoul(user, &end, 10); if (errno || *end != '\0') { snprintf(errbuff, sizeof(errbuff), "Couldn't lookup user ``%s''", user); errbuff[sizeof(errbuff) - 1] = '\0'; free(user); return (errbuff); } } free(user); } if (*g != '\0') { struct group *grp; if ((grp = getgrnam(g)) != NULL) { *gid = grp->gr_gid; } else { char *end; errno = 0; *gid = (int)strtoul(g, &end, 10); if (errno || *end != '\0') { snprintf(errbuff, sizeof(errbuff), "Couldn't lookup group ``%s''", g); errbuff[sizeof(errbuff) - 1] = '\0'; return (errbuff); } } } return (NULL); } Index: stable/10/contrib/libarchive/cpio/cpio.c =================================================================== --- stable/10/contrib/libarchive/cpio/cpio.c (revision 324417) +++ stable/10/contrib/libarchive/cpio/cpio.c (revision 324418) @@ -1,1474 +1,1478 @@ /*- * Copyright (c) 2003-2007 Tim Kientzle * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer * in this position and unchanged. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "cpio_platform.h" __FBSDID("$FreeBSD$"); #include #include #include #ifdef HAVE_SYS_MKDEV_H #include #endif #ifdef HAVE_SYS_STAT_H #include #endif #ifdef HAVE_SYS_TIME_H #include #endif #ifdef HAVE_ERRNO_H #include #endif #ifdef HAVE_FCNTL_H #include #endif #ifdef HAVE_GRP_H #include #endif #ifdef HAVE_LOCALE_H #include #endif #ifdef HAVE_PWD_H #include #endif #ifdef HAVE_SIGNAL_H #include #endif #ifdef HAVE_STDARG_H #include #endif #ifdef HAVE_STDINT_H #include #endif #include #ifdef HAVE_STDLIB_H #include #endif #ifdef HAVE_STRING_H #include #endif #ifdef HAVE_UNISTD_H #include #endif #ifdef HAVE_TIME_H #include #endif #include "cpio.h" #include "err.h" #include "line_reader.h" #include "passphrase.h" /* Fixed size of uname/gname caches. */ #define name_cache_size 101 #ifndef O_BINARY #define O_BINARY 0 #endif struct name_cache { int probes; int hits; size_t size; struct { id_t id; char *name; } cache[name_cache_size]; }; static int extract_data(struct archive *, struct archive *); const char * cpio_i64toa(int64_t); static const char *cpio_rename(const char *name); static int entry_to_archive(struct cpio *, struct archive_entry *); static int file_to_archive(struct cpio *, const char *); static void free_cache(struct name_cache *cache); static void list_item_verbose(struct cpio *, struct archive_entry *); static void long_help(void) __LA_DEAD; static const char *lookup_gname(struct cpio *, gid_t gid); static int lookup_gname_helper(struct cpio *, const char **name, id_t gid); static const char *lookup_uname(struct cpio *, uid_t uid); static int lookup_uname_helper(struct cpio *, const char **name, id_t uid); static void mode_in(struct cpio *) __LA_DEAD; static void mode_list(struct cpio *) __LA_DEAD; static void mode_out(struct cpio *); static void mode_pass(struct cpio *, const char *); static const char *remove_leading_slash(const char *); static int restore_time(struct cpio *, struct archive_entry *, const char *, int fd); static void usage(void) __LA_DEAD; static void version(void) __LA_DEAD; static const char * passphrase_callback(struct archive *, void *); static void passphrase_free(char *); int main(int argc, char *argv[]) { static char buff[16384]; struct cpio _cpio; /* Allocated on stack. */ struct cpio *cpio; const char *errmsg; int uid, gid; int opt; cpio = &_cpio; memset(cpio, 0, sizeof(*cpio)); cpio->buff = buff; cpio->buff_size = sizeof(buff); #if defined(HAVE_SIGACTION) && defined(SIGPIPE) { /* Ignore SIGPIPE signals. */ struct sigaction sa; sigemptyset(&sa.sa_mask); sa.sa_flags = 0; sa.sa_handler = SIG_IGN; sigaction(SIGPIPE, &sa, NULL); } #endif /* Set lafe_progname before calling lafe_warnc. */ lafe_setprogname(*argv, "bsdcpio"); #if HAVE_SETLOCALE if (setlocale(LC_ALL, "") == NULL) lafe_warnc(0, "Failed to set default locale"); #endif cpio->uid_override = -1; cpio->gid_override = -1; cpio->argv = argv; cpio->argc = argc; cpio->mode = '\0'; cpio->verbose = 0; cpio->compress = '\0'; cpio->extract_flags = ARCHIVE_EXTRACT_NO_AUTODIR; cpio->extract_flags |= ARCHIVE_EXTRACT_NO_OVERWRITE_NEWER; cpio->extract_flags |= ARCHIVE_EXTRACT_SECURE_SYMLINKS; cpio->extract_flags |= ARCHIVE_EXTRACT_SECURE_NODOTDOT; cpio->extract_flags |= ARCHIVE_EXTRACT_SECURE_NOABSOLUTEPATHS; cpio->extract_flags |= ARCHIVE_EXTRACT_PERM; cpio->extract_flags |= ARCHIVE_EXTRACT_FFLAGS; cpio->extract_flags |= ARCHIVE_EXTRACT_ACL; #if !defined(_WIN32) && !defined(__CYGWIN__) if (geteuid() == 0) cpio->extract_flags |= ARCHIVE_EXTRACT_OWNER; #endif cpio->bytes_per_block = 512; cpio->filename = NULL; cpio->matching = archive_match_new(); if (cpio->matching == NULL) lafe_errc(1, 0, "Out of memory"); while ((opt = cpio_getopt(cpio)) != -1) { switch (opt) { case '0': /* GNU convention: --null, -0 */ cpio->option_null = 1; break; case 'A': /* NetBSD/OpenBSD */ cpio->option_append = 1; break; case 'a': /* POSIX 1997 */ cpio->option_atime_restore = 1; break; case 'B': /* POSIX 1997 */ cpio->bytes_per_block = 5120; break; case OPTION_B64ENCODE: cpio->add_filter = opt; break; case 'C': /* NetBSD/OpenBSD */ cpio->bytes_per_block = atoi(cpio->argument); if (cpio->bytes_per_block <= 0) lafe_errc(1, 0, "Invalid blocksize %s", cpio->argument); break; case 'c': /* POSIX 1997 */ cpio->format = "odc"; break; case 'd': /* POSIX 1997 */ cpio->extract_flags &= ~ARCHIVE_EXTRACT_NO_AUTODIR; break; case 'E': /* NetBSD/OpenBSD */ if (archive_match_include_pattern_from_file( cpio->matching, cpio->argument, cpio->option_null) != ARCHIVE_OK) lafe_errc(1, 0, "Error : %s", archive_error_string(cpio->matching)); break; case 'F': /* NetBSD/OpenBSD/GNU cpio */ cpio->filename = cpio->argument; break; case 'f': /* POSIX 1997 */ if (archive_match_exclude_pattern(cpio->matching, cpio->argument) != ARCHIVE_OK) lafe_errc(1, 0, "Error : %s", archive_error_string(cpio->matching)); break; case OPTION_GRZIP: cpio->compress = opt; break; case 'H': /* GNU cpio (also --format) */ cpio->format = cpio->argument; break; case 'h': long_help(); break; case 'I': /* NetBSD/OpenBSD */ cpio->filename = cpio->argument; break; case 'i': /* POSIX 1997 */ if (cpio->mode != '\0') lafe_errc(1, 0, "Cannot use both -i and -%c", cpio->mode); cpio->mode = opt; break; case 'J': /* GNU tar, others */ cpio->compress = opt; break; case 'j': /* GNU tar, others */ cpio->compress = opt; break; case OPTION_INSECURE: cpio->extract_flags &= ~ARCHIVE_EXTRACT_SECURE_SYMLINKS; cpio->extract_flags &= ~ARCHIVE_EXTRACT_SECURE_NODOTDOT; cpio->extract_flags &= ~ARCHIVE_EXTRACT_SECURE_NOABSOLUTEPATHS; break; case 'L': /* GNU cpio */ cpio->option_follow_links = 1; break; case 'l': /* POSIX 1997 */ cpio->option_link = 1; break; case OPTION_LRZIP: case OPTION_LZ4: case OPTION_LZMA: /* GNU tar, others */ case OPTION_LZOP: /* GNU tar, others */ + case OPTION_ZSTD: cpio->compress = opt; break; case 'm': /* POSIX 1997 */ cpio->extract_flags |= ARCHIVE_EXTRACT_TIME; break; case 'n': /* GNU cpio */ cpio->option_numeric_uid_gid = 1; break; case OPTION_NO_PRESERVE_OWNER: /* GNU cpio */ cpio->extract_flags &= ~ARCHIVE_EXTRACT_OWNER; break; case 'O': /* GNU cpio */ cpio->filename = cpio->argument; break; case 'o': /* POSIX 1997 */ if (cpio->mode != '\0') lafe_errc(1, 0, "Cannot use both -o and -%c", cpio->mode); cpio->mode = opt; break; case 'p': /* POSIX 1997 */ if (cpio->mode != '\0') lafe_errc(1, 0, "Cannot use both -p and -%c", cpio->mode); cpio->mode = opt; cpio->extract_flags &= ~ARCHIVE_EXTRACT_SECURE_NODOTDOT; cpio->extract_flags &= ~ARCHIVE_EXTRACT_SECURE_NOABSOLUTEPATHS; break; case OPTION_PASSPHRASE: cpio->passphrase = cpio->argument; break; case OPTION_PRESERVE_OWNER: cpio->extract_flags |= ARCHIVE_EXTRACT_OWNER; break; case OPTION_QUIET: /* GNU cpio */ cpio->quiet = 1; break; case 'R': /* GNU cpio, also --owner */ /* TODO: owner_parse should return uname/gname * also; use that to set [ug]name_override. */ errmsg = owner_parse(cpio->argument, &uid, &gid); if (errmsg) { lafe_warnc(-1, "%s", errmsg); usage(); } if (uid != -1) { cpio->uid_override = uid; cpio->uname_override = NULL; } if (gid != -1) { cpio->gid_override = gid; cpio->gname_override = NULL; } break; case 'r': /* POSIX 1997 */ cpio->option_rename = 1; break; case 't': /* POSIX 1997 */ cpio->option_list = 1; break; case 'u': /* POSIX 1997 */ cpio->extract_flags &= ~ARCHIVE_EXTRACT_NO_OVERWRITE_NEWER; break; case OPTION_UUENCODE: cpio->add_filter = opt; break; case 'v': /* POSIX 1997 */ cpio->verbose++; break; case 'V': /* GNU cpio */ cpio->dot++; break; case OPTION_VERSION: /* GNU convention */ version(); break; #if 0 /* * cpio_getopt() handles -W specially, so it's not * available here. */ case 'W': /* Obscure, but useful GNU convention. */ break; #endif case 'y': /* tar convention */ cpio->compress = opt; break; case 'Z': /* tar convention */ cpio->compress = opt; break; case 'z': /* tar convention */ cpio->compress = opt; break; default: usage(); } } /* * Sanity-check args, error out on nonsensical combinations. */ /* -t implies -i if no mode was specified. */ if (cpio->option_list && cpio->mode == '\0') cpio->mode = 'i'; /* -t requires -i */ if (cpio->option_list && cpio->mode != 'i') lafe_errc(1, 0, "Option -t requires -i"); /* -n requires -it */ if (cpio->option_numeric_uid_gid && !cpio->option_list) lafe_errc(1, 0, "Option -n requires -it"); /* Can only specify format when writing */ if (cpio->format != NULL && cpio->mode != 'o') lafe_errc(1, 0, "Option --format requires -o"); /* -l requires -p */ if (cpio->option_link && cpio->mode != 'p') lafe_errc(1, 0, "Option -l requires -p"); /* -v overrides -V */ if (cpio->dot && cpio->verbose) cpio->dot = 0; /* TODO: Flag other nonsensical combinations. */ switch (cpio->mode) { case 'o': /* TODO: Implement old binary format in libarchive, use that here. */ if (cpio->format == NULL) cpio->format = "odc"; /* Default format */ mode_out(cpio); break; case 'i': while (*cpio->argv != NULL) { if (archive_match_include_pattern(cpio->matching, *cpio->argv) != ARCHIVE_OK) lafe_errc(1, 0, "Error : %s", archive_error_string(cpio->matching)); --cpio->argc; ++cpio->argv; } if (cpio->option_list) mode_list(cpio); else mode_in(cpio); break; case 'p': if (*cpio->argv == NULL || **cpio->argv == '\0') lafe_errc(1, 0, "-p mode requires a target directory"); mode_pass(cpio, *cpio->argv); break; default: lafe_errc(1, 0, "Must specify at least one of -i, -o, or -p"); } archive_match_free(cpio->matching); free_cache(cpio->gname_cache); free_cache(cpio->uname_cache); free(cpio->destdir); passphrase_free(cpio->ppbuff); return (cpio->return_value); } static void usage(void) { const char *p; p = lafe_getprogname(); fprintf(stderr, "Brief Usage:\n"); fprintf(stderr, " List: %s -it < archive\n", p); fprintf(stderr, " Extract: %s -i < archive\n", p); fprintf(stderr, " Create: %s -o < filenames > archive\n", p); fprintf(stderr, " Help: %s --help\n", p); exit(1); } static const char *long_help_msg = "First option must be a mode specifier:\n" " -i Input -o Output -p Pass\n" "Common Options:\n" " -v Verbose filenames -V one dot per file\n" "Create: %p -o [options] < [list of files] > [archive]\n" " -J,-y,-z,--lzma Compress archive with xz/bzip2/gzip/lzma\n" " --format {odc|newc|ustar} Select archive format\n" "List: %p -it < [archive]\n" "Extract: %p -i [options] < [archive]\n"; /* * Note that the word 'bsdcpio' will always appear in the first line * of output. * * In particular, /bin/sh scripts that need to test for the presence * of bsdcpio can use the following template: * * if (cpio --help 2>&1 | grep bsdcpio >/dev/null 2>&1 ) then \ * echo bsdcpio; else echo not bsdcpio; fi */ static void long_help(void) { const char *prog; const char *p; prog = lafe_getprogname(); fflush(stderr); p = (strcmp(prog,"bsdcpio") != 0) ? "(bsdcpio)" : ""; printf("%s%s: manipulate archive files\n", prog, p); for (p = long_help_msg; *p != '\0'; p++) { if (*p == '%') { if (p[1] == 'p') { fputs(prog, stdout); p++; } else putchar('%'); } else putchar(*p); } version(); } static void version(void) { fprintf(stdout,"bsdcpio %s - %s\n", BSDCPIO_VERSION_STRING, archive_version_details()); exit(0); } static void mode_out(struct cpio *cpio) { struct archive_entry *entry, *spare; struct lafe_line_reader *lr; const char *p; int r; if (cpio->option_append) lafe_errc(1, 0, "Append mode not yet supported."); cpio->archive_read_disk = archive_read_disk_new(); if (cpio->archive_read_disk == NULL) lafe_errc(1, 0, "Failed to allocate archive object"); if (cpio->option_follow_links) archive_read_disk_set_symlink_logical(cpio->archive_read_disk); else archive_read_disk_set_symlink_physical(cpio->archive_read_disk); archive_read_disk_set_standard_lookup(cpio->archive_read_disk); cpio->archive = archive_write_new(); if (cpio->archive == NULL) lafe_errc(1, 0, "Failed to allocate archive object"); switch (cpio->compress) { case OPTION_GRZIP: r = archive_write_add_filter_grzip(cpio->archive); break; case 'J': r = archive_write_add_filter_xz(cpio->archive); break; case OPTION_LRZIP: r = archive_write_add_filter_lrzip(cpio->archive); break; case OPTION_LZ4: r = archive_write_add_filter_lz4(cpio->archive); break; case OPTION_LZMA: r = archive_write_add_filter_lzma(cpio->archive); break; case OPTION_LZOP: r = archive_write_add_filter_lzop(cpio->archive); + break; + case OPTION_ZSTD: + r = archive_write_add_filter_zstd(cpio->archive); break; case 'j': case 'y': r = archive_write_add_filter_bzip2(cpio->archive); break; case 'z': r = archive_write_add_filter_gzip(cpio->archive); break; case 'Z': r = archive_write_add_filter_compress(cpio->archive); break; default: r = archive_write_add_filter_none(cpio->archive); break; } if (r < ARCHIVE_WARN) lafe_errc(1, 0, "Requested compression not available"); switch (cpio->add_filter) { case 0: r = ARCHIVE_OK; break; case OPTION_B64ENCODE: r = archive_write_add_filter_b64encode(cpio->archive); break; case OPTION_UUENCODE: r = archive_write_add_filter_uuencode(cpio->archive); break; } if (r < ARCHIVE_WARN) lafe_errc(1, 0, "Requested filter not available"); r = archive_write_set_format_by_name(cpio->archive, cpio->format); if (r != ARCHIVE_OK) lafe_errc(1, 0, "%s", archive_error_string(cpio->archive)); archive_write_set_bytes_per_block(cpio->archive, cpio->bytes_per_block); cpio->linkresolver = archive_entry_linkresolver_new(); archive_entry_linkresolver_set_strategy(cpio->linkresolver, archive_format(cpio->archive)); if (cpio->passphrase != NULL) r = archive_write_set_passphrase(cpio->archive, cpio->passphrase); else r = archive_write_set_passphrase_callback(cpio->archive, cpio, &passphrase_callback); if (r != ARCHIVE_OK) lafe_errc(1, 0, "%s", archive_error_string(cpio->archive)); /* * The main loop: Copy each file into the output archive. */ r = archive_write_open_filename(cpio->archive, cpio->filename); if (r != ARCHIVE_OK) lafe_errc(1, 0, "%s", archive_error_string(cpio->archive)); lr = lafe_line_reader("-", cpio->option_null); while ((p = lafe_line_reader_next(lr)) != NULL) file_to_archive(cpio, p); lafe_line_reader_free(lr); /* * The hardlink detection may have queued up a couple of entries * that can now be flushed. */ entry = NULL; archive_entry_linkify(cpio->linkresolver, &entry, &spare); while (entry != NULL) { entry_to_archive(cpio, entry); archive_entry_free(entry); entry = NULL; archive_entry_linkify(cpio->linkresolver, &entry, &spare); } r = archive_write_close(cpio->archive); if (cpio->dot) fprintf(stderr, "\n"); if (r != ARCHIVE_OK) lafe_errc(1, 0, "%s", archive_error_string(cpio->archive)); if (!cpio->quiet) { int64_t blocks = (archive_filter_bytes(cpio->archive, 0) + 511) / 512; fprintf(stderr, "%lu %s\n", (unsigned long)blocks, blocks == 1 ? "block" : "blocks"); } archive_write_free(cpio->archive); archive_entry_linkresolver_free(cpio->linkresolver); } static const char * remove_leading_slash(const char *p) { const char *rp; /* Remove leading "//./" or "//?/" or "//?/UNC/" * (absolute path prefixes used by Windows API) */ if ((p[0] == '/' || p[0] == '\\') && (p[1] == '/' || p[1] == '\\') && (p[2] == '.' || p[2] == '?') && (p[3] == '/' || p[3] == '\\')) { if (p[2] == '?' && (p[4] == 'U' || p[4] == 'u') && (p[5] == 'N' || p[5] == 'n') && (p[6] == 'C' || p[6] == 'c') && (p[7] == '/' || p[7] == '\\')) p += 8; else p += 4; } do { rp = p; /* Remove leading drive letter from archives created * on Windows. */ if (((p[0] >= 'a' && p[0] <= 'z') || (p[0] >= 'A' && p[0] <= 'Z')) && p[1] == ':') { p += 2; } /* Remove leading "/../", "//", etc. */ while (p[0] == '/' || p[0] == '\\') { if (p[1] == '.' && p[2] == '.' && (p[3] == '/' || p[3] == '\\')) { p += 3; /* Remove "/..", leave "/" * for next pass. */ } else p += 1; /* Remove "/". */ } } while (rp != p); return (p); } /* * This is used by both out mode (to copy objects from disk into * an archive) and pass mode (to copy objects from disk to * an archive_write_disk "archive"). */ static int file_to_archive(struct cpio *cpio, const char *srcpath) { const char *destpath; struct archive_entry *entry, *spare; size_t len; int r; /* * Create an archive_entry describing the source file. * */ entry = archive_entry_new(); if (entry == NULL) lafe_errc(1, 0, "Couldn't allocate entry"); archive_entry_copy_sourcepath(entry, srcpath); r = archive_read_disk_entry_from_file(cpio->archive_read_disk, entry, -1, NULL); if (r < ARCHIVE_FAILED) lafe_errc(1, 0, "%s", archive_error_string(cpio->archive_read_disk)); if (r < ARCHIVE_OK) lafe_warnc(0, "%s", archive_error_string(cpio->archive_read_disk)); if (r <= ARCHIVE_FAILED) { archive_entry_free(entry); cpio->return_value = 1; return (r); } if (cpio->uid_override >= 0) { archive_entry_set_uid(entry, cpio->uid_override); archive_entry_set_uname(entry, cpio->uname_override); } if (cpio->gid_override >= 0) { archive_entry_set_gid(entry, cpio->gid_override); archive_entry_set_gname(entry, cpio->gname_override); } /* * Generate a destination path for this entry. * "destination path" is the name to which it will be copied in * pass mode or the name that will go into the archive in * output mode. */ destpath = srcpath; if (cpio->destdir) { len = strlen(cpio->destdir) + strlen(srcpath) + 8; if (len >= cpio->pass_destpath_alloc) { while (len >= cpio->pass_destpath_alloc) { cpio->pass_destpath_alloc += 512; cpio->pass_destpath_alloc *= 2; } free(cpio->pass_destpath); cpio->pass_destpath = malloc(cpio->pass_destpath_alloc); if (cpio->pass_destpath == NULL) lafe_errc(1, ENOMEM, "Can't allocate path buffer"); } strcpy(cpio->pass_destpath, cpio->destdir); strcat(cpio->pass_destpath, remove_leading_slash(srcpath)); destpath = cpio->pass_destpath; } if (cpio->option_rename) destpath = cpio_rename(destpath); if (destpath == NULL) return (0); archive_entry_copy_pathname(entry, destpath); /* * If we're trying to preserve hardlinks, match them here. */ spare = NULL; if (cpio->linkresolver != NULL && archive_entry_filetype(entry) != AE_IFDIR) { archive_entry_linkify(cpio->linkresolver, &entry, &spare); } if (entry != NULL) { r = entry_to_archive(cpio, entry); archive_entry_free(entry); if (spare != NULL) { if (r == 0) r = entry_to_archive(cpio, spare); archive_entry_free(spare); } } return (r); } static int entry_to_archive(struct cpio *cpio, struct archive_entry *entry) { const char *destpath = archive_entry_pathname(entry); const char *srcpath = archive_entry_sourcepath(entry); int fd = -1; ssize_t bytes_read; int r; /* Print out the destination name to the user. */ if (cpio->verbose) fprintf(stderr,"%s", destpath); if (cpio->dot) fprintf(stderr, "."); /* * Option_link only makes sense in pass mode and for * regular files. Also note: if a link operation fails * because of cross-device restrictions, we'll fall back * to copy mode for that entry. * * TODO: Test other cpio implementations to see if they * hard-link anything other than regular files here. */ if (cpio->option_link && archive_entry_filetype(entry) == AE_IFREG) { struct archive_entry *t; /* Save the original entry in case we need it later. */ t = archive_entry_clone(entry); if (t == NULL) lafe_errc(1, ENOMEM, "Can't create link"); /* Note: link(2) doesn't create parent directories, * so we use archive_write_header() instead as a * convenience. */ archive_entry_set_hardlink(t, srcpath); /* This is a straight link that carries no data. */ archive_entry_set_size(t, 0); r = archive_write_header(cpio->archive, t); archive_entry_free(t); if (r != ARCHIVE_OK) lafe_warnc(archive_errno(cpio->archive), "%s", archive_error_string(cpio->archive)); if (r == ARCHIVE_FATAL) exit(1); #ifdef EXDEV if (r != ARCHIVE_OK && archive_errno(cpio->archive) == EXDEV) { /* Cross-device link: Just fall through and use * the original entry to copy the file over. */ lafe_warnc(0, "Copying file instead"); } else #endif return (0); } /* * Make sure we can open the file (if necessary) before * trying to write the header. */ if (archive_entry_filetype(entry) == AE_IFREG) { if (archive_entry_size(entry) > 0) { fd = open(srcpath, O_RDONLY | O_BINARY); if (fd < 0) { lafe_warnc(errno, "%s: could not open file", srcpath); goto cleanup; } } } else { archive_entry_set_size(entry, 0); } r = archive_write_header(cpio->archive, entry); if (r != ARCHIVE_OK) lafe_warnc(archive_errno(cpio->archive), "%s: %s", srcpath, archive_error_string(cpio->archive)); if (r == ARCHIVE_FATAL) exit(1); if (r >= ARCHIVE_WARN && archive_entry_size(entry) > 0 && fd >= 0) { bytes_read = read(fd, cpio->buff, (unsigned)cpio->buff_size); while (bytes_read > 0) { ssize_t bytes_write; bytes_write = archive_write_data(cpio->archive, cpio->buff, bytes_read); if (bytes_write < 0) lafe_errc(1, archive_errno(cpio->archive), "%s", archive_error_string(cpio->archive)); if (bytes_write < bytes_read) { lafe_warnc(0, "Truncated write; file may have " "grown while being archived."); } bytes_read = read(fd, cpio->buff, (unsigned)cpio->buff_size); } } fd = restore_time(cpio, entry, srcpath, fd); cleanup: if (cpio->verbose) fprintf(stderr,"\n"); if (fd >= 0) close(fd); return (0); } static int restore_time(struct cpio *cpio, struct archive_entry *entry, const char *name, int fd) { #ifndef HAVE_UTIMES static int warned = 0; (void)cpio; /* UNUSED */ (void)entry; /* UNUSED */ (void)name; /* UNUSED */ if (!warned) lafe_warnc(0, "Can't restore access times on this platform"); warned = 1; return (fd); #else #if defined(_WIN32) && !defined(__CYGWIN__) struct __timeval times[2]; #else struct timeval times[2]; #endif if (!cpio->option_atime_restore) return (fd); times[1].tv_sec = archive_entry_mtime(entry); times[1].tv_usec = archive_entry_mtime_nsec(entry) / 1000; times[0].tv_sec = archive_entry_atime(entry); times[0].tv_usec = archive_entry_atime_nsec(entry) / 1000; #if defined(HAVE_FUTIMES) && !defined(__CYGWIN__) if (fd >= 0 && futimes(fd, times) == 0) return (fd); #endif /* * Some platform cannot restore access times if the file descriptor * is still opened. */ if (fd >= 0) { close(fd); fd = -1; } #ifdef HAVE_LUTIMES if (lutimes(name, times) != 0) #else if ((AE_IFLNK != archive_entry_filetype(entry)) && utimes(name, times) != 0) #endif lafe_warnc(errno, "Can't update time for %s", name); #endif return (fd); } static void mode_in(struct cpio *cpio) { struct archive *a; struct archive_entry *entry; struct archive *ext; const char *destpath; int r; ext = archive_write_disk_new(); if (ext == NULL) lafe_errc(1, 0, "Couldn't allocate restore object"); r = archive_write_disk_set_options(ext, cpio->extract_flags); if (r != ARCHIVE_OK) lafe_errc(1, 0, "%s", archive_error_string(ext)); a = archive_read_new(); if (a == NULL) lafe_errc(1, 0, "Couldn't allocate archive object"); archive_read_support_filter_all(a); archive_read_support_format_all(a); if (cpio->passphrase != NULL) r = archive_read_add_passphrase(a, cpio->passphrase); else r = archive_read_set_passphrase_callback(a, cpio, &passphrase_callback); if (r != ARCHIVE_OK) lafe_errc(1, 0, "%s", archive_error_string(a)); if (archive_read_open_filename(a, cpio->filename, cpio->bytes_per_block)) lafe_errc(1, archive_errno(a), "%s", archive_error_string(a)); for (;;) { r = archive_read_next_header(a, &entry); if (r == ARCHIVE_EOF) break; if (r != ARCHIVE_OK) { lafe_errc(1, archive_errno(a), "%s", archive_error_string(a)); } if (archive_match_path_excluded(cpio->matching, entry)) continue; if (cpio->option_rename) { destpath = cpio_rename(archive_entry_pathname(entry)); archive_entry_set_pathname(entry, destpath); } else destpath = archive_entry_pathname(entry); if (destpath == NULL) continue; if (cpio->verbose) fprintf(stderr, "%s\n", destpath); if (cpio->dot) fprintf(stderr, "."); if (cpio->uid_override >= 0) archive_entry_set_uid(entry, cpio->uid_override); if (cpio->gid_override >= 0) archive_entry_set_gid(entry, cpio->gid_override); r = archive_write_header(ext, entry); if (r != ARCHIVE_OK) { fprintf(stderr, "%s: %s\n", archive_entry_pathname(entry), archive_error_string(ext)); } else if (!archive_entry_size_is_set(entry) || archive_entry_size(entry) > 0) { r = extract_data(a, ext); if (r != ARCHIVE_OK) cpio->return_value = 1; } } r = archive_read_close(a); if (cpio->dot) fprintf(stderr, "\n"); if (r != ARCHIVE_OK) lafe_errc(1, 0, "%s", archive_error_string(a)); r = archive_write_close(ext); if (r != ARCHIVE_OK) lafe_errc(1, 0, "%s", archive_error_string(ext)); if (!cpio->quiet) { int64_t blocks = (archive_filter_bytes(a, 0) + 511) / 512; fprintf(stderr, "%lu %s\n", (unsigned long)blocks, blocks == 1 ? "block" : "blocks"); } archive_read_free(a); archive_write_free(ext); exit(cpio->return_value); } /* * Exits if there's a fatal error. Returns ARCHIVE_OK * if everything is kosher. */ static int extract_data(struct archive *ar, struct archive *aw) { int r; size_t size; const void *block; int64_t offset; for (;;) { r = archive_read_data_block(ar, &block, &size, &offset); if (r == ARCHIVE_EOF) return (ARCHIVE_OK); if (r != ARCHIVE_OK) { lafe_warnc(archive_errno(ar), "%s", archive_error_string(ar)); exit(1); } r = (int)archive_write_data_block(aw, block, size, offset); if (r != ARCHIVE_OK) { lafe_warnc(archive_errno(aw), "%s", archive_error_string(aw)); return (r); } } } static void mode_list(struct cpio *cpio) { struct archive *a; struct archive_entry *entry; int r; a = archive_read_new(); if (a == NULL) lafe_errc(1, 0, "Couldn't allocate archive object"); archive_read_support_filter_all(a); archive_read_support_format_all(a); if (cpio->passphrase != NULL) r = archive_read_add_passphrase(a, cpio->passphrase); else r = archive_read_set_passphrase_callback(a, cpio, &passphrase_callback); if (r != ARCHIVE_OK) lafe_errc(1, 0, "%s", archive_error_string(a)); if (archive_read_open_filename(a, cpio->filename, cpio->bytes_per_block)) lafe_errc(1, archive_errno(a), "%s", archive_error_string(a)); for (;;) { r = archive_read_next_header(a, &entry); if (r == ARCHIVE_EOF) break; if (r != ARCHIVE_OK) { lafe_errc(1, archive_errno(a), "%s", archive_error_string(a)); } if (archive_match_path_excluded(cpio->matching, entry)) continue; if (cpio->verbose) list_item_verbose(cpio, entry); else fprintf(stdout, "%s\n", archive_entry_pathname(entry)); } r = archive_read_close(a); if (r != ARCHIVE_OK) lafe_errc(1, 0, "%s", archive_error_string(a)); if (!cpio->quiet) { int64_t blocks = (archive_filter_bytes(a, 0) + 511) / 512; fprintf(stderr, "%lu %s\n", (unsigned long)blocks, blocks == 1 ? "block" : "blocks"); } archive_read_free(a); exit(0); } /* * Display information about the current file. * * The format here roughly duplicates the output of 'ls -l'. * This is based on SUSv2, where 'tar tv' is documented as * listing additional information in an "unspecified format," * and 'pax -l' is documented as using the same format as 'ls -l'. */ static void list_item_verbose(struct cpio *cpio, struct archive_entry *entry) { char size[32]; char date[32]; char uids[16], gids[16]; const char *uname, *gname; FILE *out = stdout; const char *fmt; time_t mtime; static time_t now; if (!now) time(&now); if (cpio->option_numeric_uid_gid) { /* Format numeric uid/gid for display. */ strcpy(uids, cpio_i64toa(archive_entry_uid(entry))); uname = uids; strcpy(gids, cpio_i64toa(archive_entry_gid(entry))); gname = gids; } else { /* Use uname if it's present, else lookup name from uid. */ uname = archive_entry_uname(entry); if (uname == NULL) uname = lookup_uname(cpio, (uid_t)archive_entry_uid(entry)); /* Use gname if it's present, else lookup name from gid. */ gname = archive_entry_gname(entry); if (gname == NULL) gname = lookup_gname(cpio, (uid_t)archive_entry_gid(entry)); } /* Print device number or file size. */ if (archive_entry_filetype(entry) == AE_IFCHR || archive_entry_filetype(entry) == AE_IFBLK) { snprintf(size, sizeof(size), "%lu,%lu", (unsigned long)archive_entry_rdevmajor(entry), (unsigned long)archive_entry_rdevminor(entry)); } else { strcpy(size, cpio_i64toa(archive_entry_size(entry))); } /* Format the time using 'ls -l' conventions. */ mtime = archive_entry_mtime(entry); #if defined(_WIN32) && !defined(__CYGWIN__) /* Windows' strftime function does not support %e format. */ if (mtime - now > 365*86400/2 || mtime - now < -365*86400/2) fmt = cpio->day_first ? "%d %b %Y" : "%b %d %Y"; else fmt = cpio->day_first ? "%d %b %H:%M" : "%b %d %H:%M"; #else if (mtime - now > 365*86400/2 || mtime - now < -365*86400/2) fmt = cpio->day_first ? "%e %b %Y" : "%b %e %Y"; else fmt = cpio->day_first ? "%e %b %H:%M" : "%b %e %H:%M"; #endif strftime(date, sizeof(date), fmt, localtime(&mtime)); fprintf(out, "%s%3d %-8s %-8s %8s %12s %s", archive_entry_strmode(entry), archive_entry_nlink(entry), uname, gname, size, date, archive_entry_pathname(entry)); /* Extra information for links. */ if (archive_entry_hardlink(entry)) /* Hard link */ fprintf(out, " link to %s", archive_entry_hardlink(entry)); else if (archive_entry_symlink(entry)) /* Symbolic link */ fprintf(out, " -> %s", archive_entry_symlink(entry)); fprintf(out, "\n"); } static void mode_pass(struct cpio *cpio, const char *destdir) { struct lafe_line_reader *lr; const char *p; int r; size_t destdir_len; /* Ensure target dir has a trailing '/' to simplify path surgery. */ destdir_len = strlen(destdir); cpio->destdir = malloc(destdir_len + 8); memcpy(cpio->destdir, destdir, destdir_len); if (destdir_len == 0 || destdir[destdir_len - 1] != '/') cpio->destdir[destdir_len++] = '/'; cpio->destdir[destdir_len++] = '\0'; cpio->archive = archive_write_disk_new(); if (cpio->archive == NULL) lafe_errc(1, 0, "Failed to allocate archive object"); r = archive_write_disk_set_options(cpio->archive, cpio->extract_flags); if (r != ARCHIVE_OK) lafe_errc(1, 0, "%s", archive_error_string(cpio->archive)); cpio->linkresolver = archive_entry_linkresolver_new(); archive_write_disk_set_standard_lookup(cpio->archive); cpio->archive_read_disk = archive_read_disk_new(); if (cpio->archive_read_disk == NULL) lafe_errc(1, 0, "Failed to allocate archive object"); if (cpio->option_follow_links) archive_read_disk_set_symlink_logical(cpio->archive_read_disk); else archive_read_disk_set_symlink_physical(cpio->archive_read_disk); archive_read_disk_set_standard_lookup(cpio->archive_read_disk); lr = lafe_line_reader("-", cpio->option_null); while ((p = lafe_line_reader_next(lr)) != NULL) file_to_archive(cpio, p); lafe_line_reader_free(lr); archive_entry_linkresolver_free(cpio->linkresolver); r = archive_write_close(cpio->archive); if (cpio->dot) fprintf(stderr, "\n"); if (r != ARCHIVE_OK) lafe_errc(1, 0, "%s", archive_error_string(cpio->archive)); if (!cpio->quiet) { int64_t blocks = (archive_filter_bytes(cpio->archive, 0) + 511) / 512; fprintf(stderr, "%lu %s\n", (unsigned long)blocks, blocks == 1 ? "block" : "blocks"); } archive_write_free(cpio->archive); free(cpio->pass_destpath); } /* * Prompt for a new name for this entry. Returns a pointer to the * new name or NULL if the entry should not be copied. This * implements the semantics defined in POSIX.1-1996, which specifies * that an input of '.' means the name should be unchanged. GNU cpio * treats '.' as a literal new name. */ static const char * cpio_rename(const char *name) { static char buff[1024]; FILE *t; char *p, *ret; #if defined(_WIN32) && !defined(__CYGWIN__) FILE *to; t = fopen("CONIN$", "r"); if (t == NULL) return (name); to = fopen("CONOUT$", "w"); if (to == NULL) { fclose(t); return (name); } fprintf(to, "%s (Enter/./(new name))? ", name); fclose(to); #else t = fopen("/dev/tty", "r+"); if (t == NULL) return (name); fprintf(t, "%s (Enter/./(new name))? ", name); fflush(t); #endif p = fgets(buff, sizeof(buff), t); fclose(t); if (p == NULL) /* End-of-file is a blank line. */ return (NULL); while (*p == ' ' || *p == '\t') ++p; if (*p == '\n' || *p == '\0') /* Empty line. */ return (NULL); if (*p == '.' && p[1] == '\n') /* Single period preserves original name. */ return (name); ret = p; /* Trim the final newline. */ while (*p != '\0' && *p != '\n') ++p; /* Overwrite the final \n with a null character. */ *p = '\0'; return (ret); } static void free_cache(struct name_cache *cache) { size_t i; if (cache != NULL) { for (i = 0; i < cache->size; i++) free(cache->cache[i].name); free(cache); } } /* * Lookup uname/gname from uid/gid, return NULL if no match. */ static const char * lookup_name(struct cpio *cpio, struct name_cache **name_cache_variable, int (*lookup_fn)(struct cpio *, const char **, id_t), id_t id) { char asnum[16]; struct name_cache *cache; const char *name; int slot; if (*name_cache_variable == NULL) { *name_cache_variable = calloc(1, sizeof(struct name_cache)); if (*name_cache_variable == NULL) lafe_errc(1, ENOMEM, "No more memory"); (*name_cache_variable)->size = name_cache_size; } cache = *name_cache_variable; cache->probes++; slot = id % cache->size; if (cache->cache[slot].name != NULL) { if (cache->cache[slot].id == id) { cache->hits++; return (cache->cache[slot].name); } free(cache->cache[slot].name); cache->cache[slot].name = NULL; } if (lookup_fn(cpio, &name, id)) { /* If lookup failed, format it as a number. */ snprintf(asnum, sizeof(asnum), "%u", (unsigned)id); name = asnum; } cache->cache[slot].name = strdup(name); if (cache->cache[slot].name != NULL) { cache->cache[slot].id = id; return (cache->cache[slot].name); } /* * Conveniently, NULL marks an empty slot, so * if the strdup() fails, we've just failed to * cache it. No recovery necessary. */ return (NULL); } static const char * lookup_uname(struct cpio *cpio, uid_t uid) { return (lookup_name(cpio, &cpio->uname_cache, &lookup_uname_helper, (id_t)uid)); } static int lookup_uname_helper(struct cpio *cpio, const char **name, id_t id) { struct passwd *pwent; (void)cpio; /* UNUSED */ errno = 0; pwent = getpwuid((uid_t)id); if (pwent == NULL) { if (errno && errno != ENOENT) lafe_warnc(errno, "getpwuid(%s) failed", cpio_i64toa((int64_t)id)); return 1; } *name = pwent->pw_name; return 0; } static const char * lookup_gname(struct cpio *cpio, gid_t gid) { return (lookup_name(cpio, &cpio->gname_cache, &lookup_gname_helper, (id_t)gid)); } static int lookup_gname_helper(struct cpio *cpio, const char **name, id_t id) { struct group *grent; (void)cpio; /* UNUSED */ errno = 0; grent = getgrgid((gid_t)id); if (grent == NULL) { if (errno && errno != ENOENT) lafe_warnc(errno, "getgrgid(%s) failed", cpio_i64toa((int64_t)id)); return 1; } *name = grent->gr_name; return 0; } /* * It would be nice to just use printf() for formatting large numbers, * but the compatibility problems are a big headache. Hence the * following simple utility function. */ const char * cpio_i64toa(int64_t n0) { /* 2^64 =~ 1.8 * 10^19, so 20 decimal digits suffice. * We also need 1 byte for '-' and 1 for '\0'. */ static char buff[22]; int64_t n = n0 < 0 ? -n0 : n0; char *p = buff + sizeof(buff); *--p = '\0'; do { *--p = '0' + (int)(n % 10); n /= 10; } while (n > 0); if (n0 < 0) *--p = '-'; return p; } #define PPBUFF_SIZE 1024 static const char * passphrase_callback(struct archive *a, void *_client_data) { struct cpio *cpio = (struct cpio *)_client_data; (void)a; /* UNUSED */ if (cpio->ppbuff == NULL) { cpio->ppbuff = malloc(PPBUFF_SIZE); if (cpio->ppbuff == NULL) lafe_errc(1, errno, "Out of memory"); } return lafe_readpassphrase("Enter passphrase:", cpio->ppbuff, PPBUFF_SIZE); } static void passphrase_free(char *ppbuff) { if (ppbuff != NULL) { memset(ppbuff, 0, PPBUFF_SIZE); free(ppbuff); } } Index: stable/10/contrib/libarchive/cpio/cpio.h =================================================================== --- stable/10/contrib/libarchive/cpio/cpio.h (revision 324417) +++ stable/10/contrib/libarchive/cpio/cpio.h (revision 324418) @@ -1,119 +1,120 @@ /*- * Copyright (c) 2003-2007 Tim Kientzle * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. * * $FreeBSD$ */ #ifndef CPIO_H_INCLUDED #define CPIO_H_INCLUDED #include "cpio_platform.h" #include /* * The internal state for the "cpio" program. * * Keeping all of the state in a structure like this simplifies memory * leak testing (at exit, anything left on the heap is suspect). A * pointer to this structure is passed to most cpio internal * functions. */ struct cpio { /* Option parsing */ const char *argument; /* Options */ int add_filter; /* --uuencode */ const char *filename; int mode; /* -i -o -p */ int compress; /* -j, -y, or -z */ const char *format; /* -H format */ int bytes_per_block; /* -b block_size */ int verbose; /* -v */ int dot; /* -V */ int quiet; /* --quiet */ int extract_flags; /* Flags for extract operation */ const char *compress_program; int option_append; /* -A, only relevant for -o */ int option_atime_restore; /* -a */ int option_follow_links; /* -L */ int option_link; /* -l */ int option_list; /* -t */ char option_null; /* --null */ int option_numeric_uid_gid; /* -n */ int option_rename; /* -r */ char *destdir; size_t pass_destpath_alloc; char *pass_destpath; int uid_override; char *uname_override; int gid_override; char *gname_override; int day_first; /* true if locale prefers day/mon */ const char *passphrase; /* If >= 0, then close this when done. */ int fd; /* Miscellaneous state information */ struct archive *archive; struct archive *archive_read_disk; int argc; char **argv; int return_value; /* Value returned by main() */ struct archive_entry_linkresolver *linkresolver; struct name_cache *uname_cache; struct name_cache *gname_cache; /* Work data. */ struct archive *matching; char *buff; size_t buff_size; char *ppbuff; }; const char *owner_parse(const char *, int *, int *); /* Fake short equivalents for long options that otherwise lack them. */ enum { OPTION_B64ENCODE = 1, OPTION_GRZIP, OPTION_INSECURE, OPTION_LRZIP, OPTION_LZ4, OPTION_LZMA, OPTION_LZOP, OPTION_PASSPHRASE, OPTION_NO_PRESERVE_OWNER, OPTION_PRESERVE_OWNER, OPTION_QUIET, OPTION_UUENCODE, - OPTION_VERSION + OPTION_VERSION, + OPTION_ZSTD, }; int cpio_getopt(struct cpio *cpio); #endif Index: stable/10/contrib/libarchive/cpio/test/test_extract.cpio.zst.uu =================================================================== --- stable/10/contrib/libarchive/cpio/test/test_extract.cpio.zst.uu (nonexistent) +++ stable/10/contrib/libarchive/cpio/test/test_extract.cpio.zst.uu (revision 324418) @@ -0,0 +1,6 @@ +begin 644 test_extract.cpio.zst +M*+4O_01090,`,@41%X")&@#'G6T\K16_MR)#=DK)5:.1,2J0HY2"!(1!`!7R +M$(UB`2"*D41;J2UF&)<0!Y7X'TU<%W.\W^R]GO-WW^OO^QX0`%P<]30-!#U` +?!KD!`#XP,_`U4`HT3+RF:#!7Y\V@R)5"7P"^;WEUK@`` +` +end Index: stable/10/contrib/libarchive/cpio/test/test_extract_cpio_zstd.c =================================================================== --- stable/10/contrib/libarchive/cpio/test/test_extract_cpio_zstd.c (nonexistent) +++ stable/10/contrib/libarchive/cpio/test/test_extract_cpio_zstd.c (revision 324418) @@ -0,0 +1,48 @@ +/*- + * Copyright (c) 2017 Sean Purcell + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES + * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. + * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, + * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT + * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF + * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ +#include "test.h" +__FBSDID("$FreeBSD$"); + +DEFINE_TEST(test_extract_cpio_zstd) +{ + const char *reffile = "test_extract.cpio.zst"; + int f; + + extract_reference_file(reffile); + f = systemf("%s -it < %s >test.out 2>test.err", testprog, reffile); + if (f == 0 || canZstd()) { + assertEqualInt(0, systemf("%s -i < %s >test.out 2>test.err", + testprog, reffile)); + + assertFileExists("file1"); + assertTextFileContents("contents of file1.\n", "file1"); + assertFileExists("file2"); + assertTextFileContents("contents of file2.\n", "file2"); + assertEmptyFile("test.out"); + assertTextFileContents("1 block\n", "test.err"); + } else { + skipping("It seems zstd is not supported on this platform"); + } +} Property changes on: stable/10/contrib/libarchive/cpio/test/test_extract_cpio_zstd.c ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: stable/10/contrib/libarchive/cpio/test/test_option_zstd.c =================================================================== --- stable/10/contrib/libarchive/cpio/test/test_option_zstd.c (nonexistent) +++ stable/10/contrib/libarchive/cpio/test/test_option_zstd.c (revision 324418) @@ -0,0 +1,85 @@ +/*- + * Copyright (c) 2017 Sean Purcell + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES + * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. + * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, + * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT + * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF + * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ +#include "test.h" +__FBSDID("$FreeBSD$"); + +DEFINE_TEST(test_option_zstd) +{ + char *p; + int r; + size_t s; + + /* Create a file. */ + assertMakeFile("f", 0644, "a"); + + /* Archive it with zstd compression. */ + r = systemf("echo f | %s -o --zstd >archive.out 2>archive.err", + testprog); + p = slurpfile(&s, "archive.err"); + p[s] = '\0'; + if (r != 0) { + if (strstr(p, "Unsupported compression") != NULL) { + skipping("This version of bsdcpio was compiled " + "without zstd support"); + goto done; + } + /* POSIX permits different handling of the spawnp + * system call used to launch the subsidiary + * program: */ + /* Some systems fail immediately to spawn the new process. */ + if (strstr(p, "Can't launch") != NULL && !canZstd()) { + skipping("This version of bsdcpio uses an external zstd program " + "but no such program is available on this system."); + goto done; + } + /* Some systems successfully spawn the new process, + * but fail to exec a program within that process. + * This results in failure at the first attempt to + * write. */ + if (strstr(p, "Can't write") != NULL && !canZstd()) { + skipping("This version of bsdcpio uses an external zstd program " + "but no such program is available on this system."); + goto done; + } + /* On some systems the error won't be detected until closing + time, by a 127 exit error returned by waitpid. */ + if (strstr(p, "Error closing") != NULL && !canZstd()) { + skipping("This version of bsdcpio uses an external zstd program " + "but no such program is available on this system."); + return; + } + failure("--zstd option is broken: %s", p); + assertEqualInt(r, 0); + goto done; + } + free(p); + /* Check that the archive file has an zstd signature. */ + p = slurpfile(&s, "archive.out"); + assert(s > 2); + assertEqualMem(p, "\x28\xb5\x2f\xfd", 4); + +done: + free(p); +} Property changes on: stable/10/contrib/libarchive/cpio/test/test_option_zstd.c ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: stable/10/contrib/libarchive/libarchive/archive.h =================================================================== --- stable/10/contrib/libarchive/libarchive/archive.h (revision 324417) +++ stable/10/contrib/libarchive/libarchive/archive.h (revision 324418) @@ -1,1187 +1,1191 @@ /*- * Copyright (c) 2003-2010 Tim Kientzle * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. * * $FreeBSD$ */ #ifndef ARCHIVE_H_INCLUDED #define ARCHIVE_H_INCLUDED /* * The version number is expressed as a single integer that makes it * easy to compare versions at build time: for version a.b.c, the * version number is printf("%d%03d%03d",a,b,c). For example, if you * know your application requires version 2.12.108 or later, you can * assert that ARCHIVE_VERSION_NUMBER >= 2012108. */ /* Note: Compiler will complain if this does not match archive_entry.h! */ #define ARCHIVE_VERSION_NUMBER 3003002 #include #include /* for wchar_t */ #include /* For FILE * */ #include /* For time_t */ /* * Note: archive.h is for use outside of libarchive; the configuration * headers (config.h, archive_platform.h, etc.) are purely internal. * Do NOT use HAVE_XXX configuration macros to control the behavior of * this header! If you must conditionalize, use predefined compiler and/or * platform macros. */ #if defined(__BORLANDC__) && __BORLANDC__ >= 0x560 # include #elif !defined(__WATCOMC__) && !defined(_MSC_VER) && !defined(__INTERIX) && !defined(__BORLANDC__) && !defined(_SCO_DS) && !defined(__osf__) # include #endif /* Get appropriate definitions of 64-bit integer */ #if !defined(__LA_INT64_T_DEFINED) /* Older code relied on the __LA_INT64_T macro; after 4.0 we'll switch to the typedef exclusively. */ # if ARCHIVE_VERSION_NUMBER < 4000000 #define __LA_INT64_T la_int64_t # endif #define __LA_INT64_T_DEFINED # if defined(_WIN32) && !defined(__CYGWIN__) && !defined(__WATCOMC__) typedef __int64 la_int64_t; # else # include /* ssize_t */ # if defined(_SCO_DS) || defined(__osf__) typedef long long la_int64_t; # else typedef int64_t la_int64_t; # endif # endif #endif /* The la_ssize_t should match the type used in 'struct stat' */ #if !defined(__LA_SSIZE_T_DEFINED) /* Older code relied on the __LA_SSIZE_T macro; after 4.0 we'll switch to the typedef exclusively. */ # if ARCHIVE_VERSION_NUMBER < 4000000 #define __LA_SSIZE_T la_ssize_t # endif #define __LA_SSIZE_T_DEFINED # if defined(_WIN32) && !defined(__CYGWIN__) && !defined(__WATCOMC__) # if defined(_SSIZE_T_DEFINED) || defined(_SSIZE_T_) typedef ssize_t la_ssize_t; # elif defined(_WIN64) typedef __int64 la_ssize_t; # else typedef long la_ssize_t; # endif # else # include /* ssize_t */ typedef ssize_t la_ssize_t; # endif #endif /* Large file support for Android */ #ifdef __ANDROID__ #include "android_lf.h" #endif /* * On Windows, define LIBARCHIVE_STATIC if you're building or using a * .lib. The default here assumes you're building a DLL. Only * libarchive source should ever define __LIBARCHIVE_BUILD. */ #if ((defined __WIN32__) || (defined _WIN32) || defined(__CYGWIN__)) && (!defined LIBARCHIVE_STATIC) # ifdef __LIBARCHIVE_BUILD # ifdef __GNUC__ # define __LA_DECL __attribute__((dllexport)) extern # else # define __LA_DECL __declspec(dllexport) # endif # else # ifdef __GNUC__ # define __LA_DECL # else # define __LA_DECL __declspec(dllimport) # endif # endif #else /* Static libraries or non-Windows needs no special declaration. */ # define __LA_DECL #endif #if defined(__GNUC__) && __GNUC__ >= 3 && !defined(__MINGW32__) #define __LA_PRINTF(fmtarg, firstvararg) \ __attribute__((__format__ (__printf__, fmtarg, firstvararg))) #else #define __LA_PRINTF(fmtarg, firstvararg) /* nothing */ #endif #if defined(__GNUC__) && __GNUC__ >= 3 && __GNUC_MINOR__ >= 1 # define __LA_DEPRECATED __attribute__((deprecated)) #else # define __LA_DEPRECATED #endif #ifdef __cplusplus extern "C" { #endif /* * The version number is provided as both a macro and a function. * The macro identifies the installed header; the function identifies * the library version (which may not be the same if you're using a * dynamically-linked version of the library). Of course, if the * header and library are very different, you should expect some * strangeness. Don't do that. */ __LA_DECL int archive_version_number(void); /* * Textual name/version of the library, useful for version displays. */ #define ARCHIVE_VERSION_ONLY_STRING "3.3.2" #define ARCHIVE_VERSION_STRING "libarchive " ARCHIVE_VERSION_ONLY_STRING __LA_DECL const char * archive_version_string(void); /* * Detailed textual name/version of the library and its dependencies. * This has the form: * "libarchive x.y.z zlib/a.b.c liblzma/d.e.f ... etc ..." * the list of libraries described here will vary depending on how * libarchive was compiled. */ __LA_DECL const char * archive_version_details(void); /* * Returns NULL if libarchive was compiled without the associated library. * Otherwise, returns the version number that libarchive was compiled * against. */ __LA_DECL const char * archive_zlib_version(void); __LA_DECL const char * archive_liblzma_version(void); __LA_DECL const char * archive_bzlib_version(void); __LA_DECL const char * archive_liblz4_version(void); +__LA_DECL const char * archive_libzstd_version(void); /* Declare our basic types. */ struct archive; struct archive_entry; /* * Error codes: Use archive_errno() and archive_error_string() * to retrieve details. Unless specified otherwise, all functions * that return 'int' use these codes. */ #define ARCHIVE_EOF 1 /* Found end of archive. */ #define ARCHIVE_OK 0 /* Operation was successful. */ #define ARCHIVE_RETRY (-10) /* Retry might succeed. */ #define ARCHIVE_WARN (-20) /* Partial success. */ /* For example, if write_header "fails", then you can't push data. */ #define ARCHIVE_FAILED (-25) /* Current operation cannot complete. */ /* But if write_header is "fatal," then this archive is dead and useless. */ #define ARCHIVE_FATAL (-30) /* No more operations are possible. */ /* * As far as possible, archive_errno returns standard platform errno codes. * Of course, the details vary by platform, so the actual definitions * here are stored in "archive_platform.h". The symbols are listed here * for reference; as a rule, clients should not need to know the exact * platform-dependent error code. */ /* Unrecognized or invalid file format. */ /* #define ARCHIVE_ERRNO_FILE_FORMAT */ /* Illegal usage of the library. */ /* #define ARCHIVE_ERRNO_PROGRAMMER_ERROR */ /* Unknown or unclassified error. */ /* #define ARCHIVE_ERRNO_MISC */ /* * Callbacks are invoked to automatically read/skip/write/open/close the * archive. You can provide your own for complex tasks (like breaking * archives across multiple tapes) or use standard ones built into the * library. */ /* Returns pointer and size of next block of data from archive. */ typedef la_ssize_t archive_read_callback(struct archive *, void *_client_data, const void **_buffer); /* Skips at most request bytes from archive and returns the skipped amount. * This may skip fewer bytes than requested; it may even skip zero bytes. * If you do skip fewer bytes than requested, libarchive will invoke your * read callback and discard data as necessary to make up the full skip. */ typedef la_int64_t archive_skip_callback(struct archive *, void *_client_data, la_int64_t request); /* Seeks to specified location in the file and returns the position. * Whence values are SEEK_SET, SEEK_CUR, SEEK_END from stdio.h. * Return ARCHIVE_FATAL if the seek fails for any reason. */ typedef la_int64_t archive_seek_callback(struct archive *, void *_client_data, la_int64_t offset, int whence); /* Returns size actually written, zero on EOF, -1 on error. */ typedef la_ssize_t archive_write_callback(struct archive *, void *_client_data, const void *_buffer, size_t _length); typedef int archive_open_callback(struct archive *, void *_client_data); typedef int archive_close_callback(struct archive *, void *_client_data); /* Switches from one client data object to the next/prev client data object. * This is useful for reading from different data blocks such as a set of files * that make up one large file. */ typedef int archive_switch_callback(struct archive *, void *_client_data1, void *_client_data2); /* * Returns a passphrase used for encryption or decryption, NULL on nothing * to do and give it up. */ typedef const char *archive_passphrase_callback(struct archive *, void *_client_data); /* * Codes to identify various stream filters. */ #define ARCHIVE_FILTER_NONE 0 #define ARCHIVE_FILTER_GZIP 1 #define ARCHIVE_FILTER_BZIP2 2 #define ARCHIVE_FILTER_COMPRESS 3 #define ARCHIVE_FILTER_PROGRAM 4 #define ARCHIVE_FILTER_LZMA 5 #define ARCHIVE_FILTER_XZ 6 #define ARCHIVE_FILTER_UU 7 #define ARCHIVE_FILTER_RPM 8 #define ARCHIVE_FILTER_LZIP 9 #define ARCHIVE_FILTER_LRZIP 10 #define ARCHIVE_FILTER_LZOP 11 #define ARCHIVE_FILTER_GRZIP 12 #define ARCHIVE_FILTER_LZ4 13 +#define ARCHIVE_FILTER_ZSTD 14 #if ARCHIVE_VERSION_NUMBER < 4000000 #define ARCHIVE_COMPRESSION_NONE ARCHIVE_FILTER_NONE #define ARCHIVE_COMPRESSION_GZIP ARCHIVE_FILTER_GZIP #define ARCHIVE_COMPRESSION_BZIP2 ARCHIVE_FILTER_BZIP2 #define ARCHIVE_COMPRESSION_COMPRESS ARCHIVE_FILTER_COMPRESS #define ARCHIVE_COMPRESSION_PROGRAM ARCHIVE_FILTER_PROGRAM #define ARCHIVE_COMPRESSION_LZMA ARCHIVE_FILTER_LZMA #define ARCHIVE_COMPRESSION_XZ ARCHIVE_FILTER_XZ #define ARCHIVE_COMPRESSION_UU ARCHIVE_FILTER_UU #define ARCHIVE_COMPRESSION_RPM ARCHIVE_FILTER_RPM #define ARCHIVE_COMPRESSION_LZIP ARCHIVE_FILTER_LZIP #define ARCHIVE_COMPRESSION_LRZIP ARCHIVE_FILTER_LRZIP #endif /* * Codes returned by archive_format. * * Top 16 bits identifies the format family (e.g., "tar"); lower * 16 bits indicate the variant. This is updated by read_next_header. * Note that the lower 16 bits will often vary from entry to entry. * In some cases, this variation occurs as libarchive learns more about * the archive (for example, later entries might utilize extensions that * weren't necessary earlier in the archive; in this case, libarchive * will change the format code to indicate the extended format that * was used). In other cases, it's because different tools have * modified the archive and so different parts of the archive * actually have slightly different formats. (Both tar and cpio store * format codes in each entry, so it is quite possible for each * entry to be in a different format.) */ #define ARCHIVE_FORMAT_BASE_MASK 0xff0000 #define ARCHIVE_FORMAT_CPIO 0x10000 #define ARCHIVE_FORMAT_CPIO_POSIX (ARCHIVE_FORMAT_CPIO | 1) #define ARCHIVE_FORMAT_CPIO_BIN_LE (ARCHIVE_FORMAT_CPIO | 2) #define ARCHIVE_FORMAT_CPIO_BIN_BE (ARCHIVE_FORMAT_CPIO | 3) #define ARCHIVE_FORMAT_CPIO_SVR4_NOCRC (ARCHIVE_FORMAT_CPIO | 4) #define ARCHIVE_FORMAT_CPIO_SVR4_CRC (ARCHIVE_FORMAT_CPIO | 5) #define ARCHIVE_FORMAT_CPIO_AFIO_LARGE (ARCHIVE_FORMAT_CPIO | 6) #define ARCHIVE_FORMAT_SHAR 0x20000 #define ARCHIVE_FORMAT_SHAR_BASE (ARCHIVE_FORMAT_SHAR | 1) #define ARCHIVE_FORMAT_SHAR_DUMP (ARCHIVE_FORMAT_SHAR | 2) #define ARCHIVE_FORMAT_TAR 0x30000 #define ARCHIVE_FORMAT_TAR_USTAR (ARCHIVE_FORMAT_TAR | 1) #define ARCHIVE_FORMAT_TAR_PAX_INTERCHANGE (ARCHIVE_FORMAT_TAR | 2) #define ARCHIVE_FORMAT_TAR_PAX_RESTRICTED (ARCHIVE_FORMAT_TAR | 3) #define ARCHIVE_FORMAT_TAR_GNUTAR (ARCHIVE_FORMAT_TAR | 4) #define ARCHIVE_FORMAT_ISO9660 0x40000 #define ARCHIVE_FORMAT_ISO9660_ROCKRIDGE (ARCHIVE_FORMAT_ISO9660 | 1) #define ARCHIVE_FORMAT_ZIP 0x50000 #define ARCHIVE_FORMAT_EMPTY 0x60000 #define ARCHIVE_FORMAT_AR 0x70000 #define ARCHIVE_FORMAT_AR_GNU (ARCHIVE_FORMAT_AR | 1) #define ARCHIVE_FORMAT_AR_BSD (ARCHIVE_FORMAT_AR | 2) #define ARCHIVE_FORMAT_MTREE 0x80000 #define ARCHIVE_FORMAT_RAW 0x90000 #define ARCHIVE_FORMAT_XAR 0xA0000 #define ARCHIVE_FORMAT_LHA 0xB0000 #define ARCHIVE_FORMAT_CAB 0xC0000 #define ARCHIVE_FORMAT_RAR 0xD0000 #define ARCHIVE_FORMAT_7ZIP 0xE0000 #define ARCHIVE_FORMAT_WARC 0xF0000 /* * Codes returned by archive_read_format_capabilities(). * * This list can be extended with values between 0 and 0xffff. * The original purpose of this list was to let different archive * format readers expose their general capabilities in terms of * encryption. */ #define ARCHIVE_READ_FORMAT_CAPS_NONE (0) /* no special capabilities */ #define ARCHIVE_READ_FORMAT_CAPS_ENCRYPT_DATA (1<<0) /* reader can detect encrypted data */ #define ARCHIVE_READ_FORMAT_CAPS_ENCRYPT_METADATA (1<<1) /* reader can detect encryptable metadata (pathname, mtime, etc.) */ /* * Codes returned by archive_read_has_encrypted_entries(). * * In case the archive does not support encryption detection at all * ARCHIVE_READ_FORMAT_ENCRYPTION_UNSUPPORTED is returned. If the reader * for some other reason (e.g. not enough bytes read) cannot say if * there are encrypted entries, ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW * is returned. */ #define ARCHIVE_READ_FORMAT_ENCRYPTION_UNSUPPORTED -2 #define ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW -1 /*- * Basic outline for reading an archive: * 1) Ask archive_read_new for an archive reader object. * 2) Update any global properties as appropriate. * In particular, you'll certainly want to call appropriate * archive_read_support_XXX functions. * 3) Call archive_read_open_XXX to open the archive * 4) Repeatedly call archive_read_next_header to get information about * successive archive entries. Call archive_read_data to extract * data for entries of interest. * 5) Call archive_read_free to end processing. */ __LA_DECL struct archive *archive_read_new(void); /* * The archive_read_support_XXX calls enable auto-detect for this * archive handle. They also link in the necessary support code. * For example, if you don't want bzlib linked in, don't invoke * support_compression_bzip2(). The "all" functions provide the * obvious shorthand. */ #if ARCHIVE_VERSION_NUMBER < 4000000 __LA_DECL int archive_read_support_compression_all(struct archive *) __LA_DEPRECATED; __LA_DECL int archive_read_support_compression_bzip2(struct archive *) __LA_DEPRECATED; __LA_DECL int archive_read_support_compression_compress(struct archive *) __LA_DEPRECATED; __LA_DECL int archive_read_support_compression_gzip(struct archive *) __LA_DEPRECATED; __LA_DECL int archive_read_support_compression_lzip(struct archive *) __LA_DEPRECATED; __LA_DECL int archive_read_support_compression_lzma(struct archive *) __LA_DEPRECATED; __LA_DECL int archive_read_support_compression_none(struct archive *) __LA_DEPRECATED; __LA_DECL int archive_read_support_compression_program(struct archive *, const char *command) __LA_DEPRECATED; __LA_DECL int archive_read_support_compression_program_signature (struct archive *, const char *, const void * /* match */, size_t) __LA_DEPRECATED; __LA_DECL int archive_read_support_compression_rpm(struct archive *) __LA_DEPRECATED; __LA_DECL int archive_read_support_compression_uu(struct archive *) __LA_DEPRECATED; __LA_DECL int archive_read_support_compression_xz(struct archive *) __LA_DEPRECATED; #endif __LA_DECL int archive_read_support_filter_all(struct archive *); __LA_DECL int archive_read_support_filter_bzip2(struct archive *); __LA_DECL int archive_read_support_filter_compress(struct archive *); __LA_DECL int archive_read_support_filter_gzip(struct archive *); __LA_DECL int archive_read_support_filter_grzip(struct archive *); __LA_DECL int archive_read_support_filter_lrzip(struct archive *); __LA_DECL int archive_read_support_filter_lz4(struct archive *); __LA_DECL int archive_read_support_filter_lzip(struct archive *); __LA_DECL int archive_read_support_filter_lzma(struct archive *); __LA_DECL int archive_read_support_filter_lzop(struct archive *); __LA_DECL int archive_read_support_filter_none(struct archive *); __LA_DECL int archive_read_support_filter_program(struct archive *, const char *command); __LA_DECL int archive_read_support_filter_program_signature (struct archive *, const char * /* cmd */, const void * /* match */, size_t); __LA_DECL int archive_read_support_filter_rpm(struct archive *); __LA_DECL int archive_read_support_filter_uu(struct archive *); __LA_DECL int archive_read_support_filter_xz(struct archive *); +__LA_DECL int archive_read_support_filter_zstd(struct archive *); __LA_DECL int archive_read_support_format_7zip(struct archive *); __LA_DECL int archive_read_support_format_all(struct archive *); __LA_DECL int archive_read_support_format_ar(struct archive *); __LA_DECL int archive_read_support_format_by_code(struct archive *, int); __LA_DECL int archive_read_support_format_cab(struct archive *); __LA_DECL int archive_read_support_format_cpio(struct archive *); __LA_DECL int archive_read_support_format_empty(struct archive *); __LA_DECL int archive_read_support_format_gnutar(struct archive *); __LA_DECL int archive_read_support_format_iso9660(struct archive *); __LA_DECL int archive_read_support_format_lha(struct archive *); __LA_DECL int archive_read_support_format_mtree(struct archive *); __LA_DECL int archive_read_support_format_rar(struct archive *); __LA_DECL int archive_read_support_format_raw(struct archive *); __LA_DECL int archive_read_support_format_tar(struct archive *); __LA_DECL int archive_read_support_format_warc(struct archive *); __LA_DECL int archive_read_support_format_xar(struct archive *); /* archive_read_support_format_zip() enables both streamable and seekable * zip readers. */ __LA_DECL int archive_read_support_format_zip(struct archive *); /* Reads Zip archives as stream from beginning to end. Doesn't * correctly handle SFX ZIP files or ZIP archives that have been modified * in-place. */ __LA_DECL int archive_read_support_format_zip_streamable(struct archive *); /* Reads starting from central directory; requires seekable input. */ __LA_DECL int archive_read_support_format_zip_seekable(struct archive *); /* Functions to manually set the format and filters to be used. This is * useful to bypass the bidding process when the format and filters to use * is known in advance. */ __LA_DECL int archive_read_set_format(struct archive *, int); __LA_DECL int archive_read_append_filter(struct archive *, int); __LA_DECL int archive_read_append_filter_program(struct archive *, const char *); __LA_DECL int archive_read_append_filter_program_signature (struct archive *, const char *, const void * /* match */, size_t); /* Set various callbacks. */ __LA_DECL int archive_read_set_open_callback(struct archive *, archive_open_callback *); __LA_DECL int archive_read_set_read_callback(struct archive *, archive_read_callback *); __LA_DECL int archive_read_set_seek_callback(struct archive *, archive_seek_callback *); __LA_DECL int archive_read_set_skip_callback(struct archive *, archive_skip_callback *); __LA_DECL int archive_read_set_close_callback(struct archive *, archive_close_callback *); /* Callback used to switch between one data object to the next */ __LA_DECL int archive_read_set_switch_callback(struct archive *, archive_switch_callback *); /* This sets the first data object. */ __LA_DECL int archive_read_set_callback_data(struct archive *, void *); /* This sets data object at specified index */ __LA_DECL int archive_read_set_callback_data2(struct archive *, void *, unsigned int); /* This adds a data object at the specified index. */ __LA_DECL int archive_read_add_callback_data(struct archive *, void *, unsigned int); /* This appends a data object to the end of list */ __LA_DECL int archive_read_append_callback_data(struct archive *, void *); /* This prepends a data object to the beginning of list */ __LA_DECL int archive_read_prepend_callback_data(struct archive *, void *); /* Opening freezes the callbacks. */ __LA_DECL int archive_read_open1(struct archive *); /* Convenience wrappers around the above. */ __LA_DECL int archive_read_open(struct archive *, void *_client_data, archive_open_callback *, archive_read_callback *, archive_close_callback *); __LA_DECL int archive_read_open2(struct archive *, void *_client_data, archive_open_callback *, archive_read_callback *, archive_skip_callback *, archive_close_callback *); /* * A variety of shortcuts that invoke archive_read_open() with * canned callbacks suitable for common situations. The ones that * accept a block size handle tape blocking correctly. */ /* Use this if you know the filename. Note: NULL indicates stdin. */ __LA_DECL int archive_read_open_filename(struct archive *, const char *_filename, size_t _block_size); /* Use this for reading multivolume files by filenames. * NOTE: Must be NULL terminated. Sorting is NOT done. */ __LA_DECL int archive_read_open_filenames(struct archive *, const char **_filenames, size_t _block_size); __LA_DECL int archive_read_open_filename_w(struct archive *, const wchar_t *_filename, size_t _block_size); /* archive_read_open_file() is a deprecated synonym for ..._open_filename(). */ __LA_DECL int archive_read_open_file(struct archive *, const char *_filename, size_t _block_size) __LA_DEPRECATED; /* Read an archive that's stored in memory. */ __LA_DECL int archive_read_open_memory(struct archive *, const void * buff, size_t size); /* A more involved version that is only used for internal testing. */ __LA_DECL int archive_read_open_memory2(struct archive *a, const void *buff, size_t size, size_t read_size); /* Read an archive that's already open, using the file descriptor. */ __LA_DECL int archive_read_open_fd(struct archive *, int _fd, size_t _block_size); /* Read an archive that's already open, using a FILE *. */ /* Note: DO NOT use this with tape drives. */ __LA_DECL int archive_read_open_FILE(struct archive *, FILE *_file); /* Parses and returns next entry header. */ __LA_DECL int archive_read_next_header(struct archive *, struct archive_entry **); /* Parses and returns next entry header using the archive_entry passed in */ __LA_DECL int archive_read_next_header2(struct archive *, struct archive_entry *); /* * Retrieve the byte offset in UNCOMPRESSED data where last-read * header started. */ __LA_DECL la_int64_t archive_read_header_position(struct archive *); /* * Returns 1 if the archive contains at least one encrypted entry. * If the archive format not support encryption at all * ARCHIVE_READ_FORMAT_ENCRYPTION_UNSUPPORTED is returned. * If for any other reason (e.g. not enough data read so far) * we cannot say whether there are encrypted entries, then * ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW is returned. * In general, this function will return values below zero when the * reader is uncertain or totally incapable of encryption support. * When this function returns 0 you can be sure that the reader * supports encryption detection but no encrypted entries have * been found yet. * * NOTE: If the metadata/header of an archive is also encrypted, you * cannot rely on the number of encrypted entries. That is why this * function does not return the number of encrypted entries but# * just shows that there are some. */ __LA_DECL int archive_read_has_encrypted_entries(struct archive *); /* * Returns a bitmask of capabilities that are supported by the archive format reader. * If the reader has no special capabilities, ARCHIVE_READ_FORMAT_CAPS_NONE is returned. */ __LA_DECL int archive_read_format_capabilities(struct archive *); /* Read data from the body of an entry. Similar to read(2). */ __LA_DECL la_ssize_t archive_read_data(struct archive *, void *, size_t); /* Seek within the body of an entry. Similar to lseek(2). */ __LA_DECL la_int64_t archive_seek_data(struct archive *, la_int64_t, int); /* * A zero-copy version of archive_read_data that also exposes the file offset * of each returned block. Note that the client has no way to specify * the desired size of the block. The API does guarantee that offsets will * be strictly increasing and that returned blocks will not overlap. */ __LA_DECL int archive_read_data_block(struct archive *a, const void **buff, size_t *size, la_int64_t *offset); /*- * Some convenience functions that are built on archive_read_data: * 'skip': skips entire entry * 'into_buffer': writes data into memory buffer that you provide * 'into_fd': writes data to specified filedes */ __LA_DECL int archive_read_data_skip(struct archive *); __LA_DECL int archive_read_data_into_fd(struct archive *, int fd); /* * Set read options. */ /* Apply option to the format only. */ __LA_DECL int archive_read_set_format_option(struct archive *_a, const char *m, const char *o, const char *v); /* Apply option to the filter only. */ __LA_DECL int archive_read_set_filter_option(struct archive *_a, const char *m, const char *o, const char *v); /* Apply option to both the format and the filter. */ __LA_DECL int archive_read_set_option(struct archive *_a, const char *m, const char *o, const char *v); /* Apply option string to both the format and the filter. */ __LA_DECL int archive_read_set_options(struct archive *_a, const char *opts); /* * Add a decryption passphrase. */ __LA_DECL int archive_read_add_passphrase(struct archive *, const char *); __LA_DECL int archive_read_set_passphrase_callback(struct archive *, void *client_data, archive_passphrase_callback *); /*- * Convenience function to recreate the current entry (whose header * has just been read) on disk. * * This does quite a bit more than just copy data to disk. It also: * - Creates intermediate directories as required. * - Manages directory permissions: non-writable directories will * be initially created with write permission enabled; when the * archive is closed, dir permissions are edited to the values specified * in the archive. * - Checks hardlinks: hardlinks will not be extracted unless the * linked-to file was also extracted within the same session. (TODO) */ /* The "flags" argument selects optional behavior, 'OR' the flags you want. */ /* Default: Do not try to set owner/group. */ #define ARCHIVE_EXTRACT_OWNER (0x0001) /* Default: Do obey umask, do not restore SUID/SGID/SVTX bits. */ #define ARCHIVE_EXTRACT_PERM (0x0002) /* Default: Do not restore mtime/atime. */ #define ARCHIVE_EXTRACT_TIME (0x0004) /* Default: Replace existing files. */ #define ARCHIVE_EXTRACT_NO_OVERWRITE (0x0008) /* Default: Try create first, unlink only if create fails with EEXIST. */ #define ARCHIVE_EXTRACT_UNLINK (0x0010) /* Default: Do not restore ACLs. */ #define ARCHIVE_EXTRACT_ACL (0x0020) /* Default: Do not restore fflags. */ #define ARCHIVE_EXTRACT_FFLAGS (0x0040) /* Default: Do not restore xattrs. */ #define ARCHIVE_EXTRACT_XATTR (0x0080) /* Default: Do not try to guard against extracts redirected by symlinks. */ /* Note: With ARCHIVE_EXTRACT_UNLINK, will remove any intermediate symlink. */ #define ARCHIVE_EXTRACT_SECURE_SYMLINKS (0x0100) /* Default: Do not reject entries with '..' as path elements. */ #define ARCHIVE_EXTRACT_SECURE_NODOTDOT (0x0200) /* Default: Create parent directories as needed. */ #define ARCHIVE_EXTRACT_NO_AUTODIR (0x0400) /* Default: Overwrite files, even if one on disk is newer. */ #define ARCHIVE_EXTRACT_NO_OVERWRITE_NEWER (0x0800) /* Detect blocks of 0 and write holes instead. */ #define ARCHIVE_EXTRACT_SPARSE (0x1000) /* Default: Do not restore Mac extended metadata. */ /* This has no effect except on Mac OS. */ #define ARCHIVE_EXTRACT_MAC_METADATA (0x2000) /* Default: Use HFS+ compression if it was compressed. */ /* This has no effect except on Mac OS v10.6 or later. */ #define ARCHIVE_EXTRACT_NO_HFS_COMPRESSION (0x4000) /* Default: Do not use HFS+ compression if it was not compressed. */ /* This has no effect except on Mac OS v10.6 or later. */ #define ARCHIVE_EXTRACT_HFS_COMPRESSION_FORCED (0x8000) /* Default: Do not reject entries with absolute paths */ #define ARCHIVE_EXTRACT_SECURE_NOABSOLUTEPATHS (0x10000) /* Default: Do not clear no-change flags when unlinking object */ #define ARCHIVE_EXTRACT_CLEAR_NOCHANGE_FFLAGS (0x20000) __LA_DECL int archive_read_extract(struct archive *, struct archive_entry *, int flags); __LA_DECL int archive_read_extract2(struct archive *, struct archive_entry *, struct archive * /* dest */); __LA_DECL void archive_read_extract_set_progress_callback(struct archive *, void (*_progress_func)(void *), void *_user_data); /* Record the dev/ino of a file that will not be written. This is * generally set to the dev/ino of the archive being read. */ __LA_DECL void archive_read_extract_set_skip_file(struct archive *, la_int64_t, la_int64_t); /* Close the file and release most resources. */ __LA_DECL int archive_read_close(struct archive *); /* Release all resources and destroy the object. */ /* Note that archive_read_free will call archive_read_close for you. */ __LA_DECL int archive_read_free(struct archive *); #if ARCHIVE_VERSION_NUMBER < 4000000 /* Synonym for archive_read_free() for backwards compatibility. */ __LA_DECL int archive_read_finish(struct archive *) __LA_DEPRECATED; #endif /*- * To create an archive: * 1) Ask archive_write_new for an archive writer object. * 2) Set any global properties. In particular, you should set * the compression and format to use. * 3) Call archive_write_open to open the file (most people * will use archive_write_open_file or archive_write_open_fd, * which provide convenient canned I/O callbacks for you). * 4) For each entry: * - construct an appropriate struct archive_entry structure * - archive_write_header to write the header * - archive_write_data to write the entry data * 5) archive_write_close to close the output * 6) archive_write_free to cleanup the writer and release resources */ __LA_DECL struct archive *archive_write_new(void); __LA_DECL int archive_write_set_bytes_per_block(struct archive *, int bytes_per_block); __LA_DECL int archive_write_get_bytes_per_block(struct archive *); /* XXX This is badly misnamed; suggestions appreciated. XXX */ __LA_DECL int archive_write_set_bytes_in_last_block(struct archive *, int bytes_in_last_block); __LA_DECL int archive_write_get_bytes_in_last_block(struct archive *); /* The dev/ino of a file that won't be archived. This is used * to avoid recursively adding an archive to itself. */ __LA_DECL int archive_write_set_skip_file(struct archive *, la_int64_t, la_int64_t); #if ARCHIVE_VERSION_NUMBER < 4000000 __LA_DECL int archive_write_set_compression_bzip2(struct archive *) __LA_DEPRECATED; __LA_DECL int archive_write_set_compression_compress(struct archive *) __LA_DEPRECATED; __LA_DECL int archive_write_set_compression_gzip(struct archive *) __LA_DEPRECATED; __LA_DECL int archive_write_set_compression_lzip(struct archive *) __LA_DEPRECATED; __LA_DECL int archive_write_set_compression_lzma(struct archive *) __LA_DEPRECATED; __LA_DECL int archive_write_set_compression_none(struct archive *) __LA_DEPRECATED; __LA_DECL int archive_write_set_compression_program(struct archive *, const char *cmd) __LA_DEPRECATED; __LA_DECL int archive_write_set_compression_xz(struct archive *) __LA_DEPRECATED; #endif /* A convenience function to set the filter based on the code. */ __LA_DECL int archive_write_add_filter(struct archive *, int filter_code); __LA_DECL int archive_write_add_filter_by_name(struct archive *, const char *name); __LA_DECL int archive_write_add_filter_b64encode(struct archive *); __LA_DECL int archive_write_add_filter_bzip2(struct archive *); __LA_DECL int archive_write_add_filter_compress(struct archive *); __LA_DECL int archive_write_add_filter_grzip(struct archive *); __LA_DECL int archive_write_add_filter_gzip(struct archive *); __LA_DECL int archive_write_add_filter_lrzip(struct archive *); __LA_DECL int archive_write_add_filter_lz4(struct archive *); __LA_DECL int archive_write_add_filter_lzip(struct archive *); __LA_DECL int archive_write_add_filter_lzma(struct archive *); __LA_DECL int archive_write_add_filter_lzop(struct archive *); __LA_DECL int archive_write_add_filter_none(struct archive *); __LA_DECL int archive_write_add_filter_program(struct archive *, const char *cmd); __LA_DECL int archive_write_add_filter_uuencode(struct archive *); __LA_DECL int archive_write_add_filter_xz(struct archive *); +__LA_DECL int archive_write_add_filter_zstd(struct archive *); /* A convenience function to set the format based on the code or name. */ __LA_DECL int archive_write_set_format(struct archive *, int format_code); __LA_DECL int archive_write_set_format_by_name(struct archive *, const char *name); /* To minimize link pollution, use one or more of the following. */ __LA_DECL int archive_write_set_format_7zip(struct archive *); __LA_DECL int archive_write_set_format_ar_bsd(struct archive *); __LA_DECL int archive_write_set_format_ar_svr4(struct archive *); __LA_DECL int archive_write_set_format_cpio(struct archive *); __LA_DECL int archive_write_set_format_cpio_newc(struct archive *); __LA_DECL int archive_write_set_format_gnutar(struct archive *); __LA_DECL int archive_write_set_format_iso9660(struct archive *); __LA_DECL int archive_write_set_format_mtree(struct archive *); __LA_DECL int archive_write_set_format_mtree_classic(struct archive *); /* TODO: int archive_write_set_format_old_tar(struct archive *); */ __LA_DECL int archive_write_set_format_pax(struct archive *); __LA_DECL int archive_write_set_format_pax_restricted(struct archive *); __LA_DECL int archive_write_set_format_raw(struct archive *); __LA_DECL int archive_write_set_format_shar(struct archive *); __LA_DECL int archive_write_set_format_shar_dump(struct archive *); __LA_DECL int archive_write_set_format_ustar(struct archive *); __LA_DECL int archive_write_set_format_v7tar(struct archive *); __LA_DECL int archive_write_set_format_warc(struct archive *); __LA_DECL int archive_write_set_format_xar(struct archive *); __LA_DECL int archive_write_set_format_zip(struct archive *); __LA_DECL int archive_write_set_format_filter_by_ext(struct archive *a, const char *filename); __LA_DECL int archive_write_set_format_filter_by_ext_def(struct archive *a, const char *filename, const char * def_ext); __LA_DECL int archive_write_zip_set_compression_deflate(struct archive *); __LA_DECL int archive_write_zip_set_compression_store(struct archive *); __LA_DECL int archive_write_open(struct archive *, void *, archive_open_callback *, archive_write_callback *, archive_close_callback *); __LA_DECL int archive_write_open_fd(struct archive *, int _fd); __LA_DECL int archive_write_open_filename(struct archive *, const char *_file); __LA_DECL int archive_write_open_filename_w(struct archive *, const wchar_t *_file); /* A deprecated synonym for archive_write_open_filename() */ __LA_DECL int archive_write_open_file(struct archive *, const char *_file) __LA_DEPRECATED; __LA_DECL int archive_write_open_FILE(struct archive *, FILE *); /* _buffSize is the size of the buffer, _used refers to a variable that * will be updated after each write into the buffer. */ __LA_DECL int archive_write_open_memory(struct archive *, void *_buffer, size_t _buffSize, size_t *_used); /* * Note that the library will truncate writes beyond the size provided * to archive_write_header or pad if the provided data is short. */ __LA_DECL int archive_write_header(struct archive *, struct archive_entry *); __LA_DECL la_ssize_t archive_write_data(struct archive *, const void *, size_t); /* This interface is currently only available for archive_write_disk handles. */ __LA_DECL la_ssize_t archive_write_data_block(struct archive *, const void *, size_t, la_int64_t); __LA_DECL int archive_write_finish_entry(struct archive *); __LA_DECL int archive_write_close(struct archive *); /* Marks the archive as FATAL so that a subsequent free() operation * won't try to close() cleanly. Provides a fast abort capability * when the client discovers that things have gone wrong. */ __LA_DECL int archive_write_fail(struct archive *); /* This can fail if the archive wasn't already closed, in which case * archive_write_free() will implicitly call archive_write_close(). */ __LA_DECL int archive_write_free(struct archive *); #if ARCHIVE_VERSION_NUMBER < 4000000 /* Synonym for archive_write_free() for backwards compatibility. */ __LA_DECL int archive_write_finish(struct archive *) __LA_DEPRECATED; #endif /* * Set write options. */ /* Apply option to the format only. */ __LA_DECL int archive_write_set_format_option(struct archive *_a, const char *m, const char *o, const char *v); /* Apply option to the filter only. */ __LA_DECL int archive_write_set_filter_option(struct archive *_a, const char *m, const char *o, const char *v); /* Apply option to both the format and the filter. */ __LA_DECL int archive_write_set_option(struct archive *_a, const char *m, const char *o, const char *v); /* Apply option string to both the format and the filter. */ __LA_DECL int archive_write_set_options(struct archive *_a, const char *opts); /* * Set a encryption passphrase. */ __LA_DECL int archive_write_set_passphrase(struct archive *_a, const char *p); __LA_DECL int archive_write_set_passphrase_callback(struct archive *, void *client_data, archive_passphrase_callback *); /*- * ARCHIVE_WRITE_DISK API * * To create objects on disk: * 1) Ask archive_write_disk_new for a new archive_write_disk object. * 2) Set any global properties. In particular, you probably * want to set the options. * 3) For each entry: * - construct an appropriate struct archive_entry structure * - archive_write_header to create the file/dir/etc on disk * - archive_write_data to write the entry data * 4) archive_write_free to cleanup the writer and release resources * * In particular, you can use this in conjunction with archive_read() * to pull entries out of an archive and create them on disk. */ __LA_DECL struct archive *archive_write_disk_new(void); /* This file will not be overwritten. */ __LA_DECL int archive_write_disk_set_skip_file(struct archive *, la_int64_t, la_int64_t); /* Set flags to control how the next item gets created. * This accepts a bitmask of ARCHIVE_EXTRACT_XXX flags defined above. */ __LA_DECL int archive_write_disk_set_options(struct archive *, int flags); /* * The lookup functions are given uname/uid (or gname/gid) pairs and * return a uid (gid) suitable for this system. These are used for * restoring ownership and for setting ACLs. The default functions * are naive, they just return the uid/gid. These are small, so reasonable * for applications that don't need to preserve ownership; they * are probably also appropriate for applications that are doing * same-system backup and restore. */ /* * The "standard" lookup functions use common system calls to lookup * the uname/gname, falling back to the uid/gid if the names can't be * found. They cache lookups and are reasonably fast, but can be very * large, so they are not used unless you ask for them. In * particular, these match the specifications of POSIX "pax" and old * POSIX "tar". */ __LA_DECL int archive_write_disk_set_standard_lookup(struct archive *); /* * If neither the default (naive) nor the standard (big) functions suit * your needs, you can write your own and register them. Be sure to * include a cleanup function if you have allocated private data. */ __LA_DECL int archive_write_disk_set_group_lookup(struct archive *, void * /* private_data */, la_int64_t (*)(void *, const char *, la_int64_t), void (* /* cleanup */)(void *)); __LA_DECL int archive_write_disk_set_user_lookup(struct archive *, void * /* private_data */, la_int64_t (*)(void *, const char *, la_int64_t), void (* /* cleanup */)(void *)); __LA_DECL la_int64_t archive_write_disk_gid(struct archive *, const char *, la_int64_t); __LA_DECL la_int64_t archive_write_disk_uid(struct archive *, const char *, la_int64_t); /* * ARCHIVE_READ_DISK API * * This is still evolving and somewhat experimental. */ __LA_DECL struct archive *archive_read_disk_new(void); /* The names for symlink modes here correspond to an old BSD * command-line argument convention: -L, -P, -H */ /* Follow all symlinks. */ __LA_DECL int archive_read_disk_set_symlink_logical(struct archive *); /* Follow no symlinks. */ __LA_DECL int archive_read_disk_set_symlink_physical(struct archive *); /* Follow symlink initially, then not. */ __LA_DECL int archive_read_disk_set_symlink_hybrid(struct archive *); /* TODO: Handle Linux stat32/stat64 ugliness. */ __LA_DECL int archive_read_disk_entry_from_file(struct archive *, struct archive_entry *, int /* fd */, const struct stat *); /* Look up gname for gid or uname for uid. */ /* Default implementations are very, very stupid. */ __LA_DECL const char *archive_read_disk_gname(struct archive *, la_int64_t); __LA_DECL const char *archive_read_disk_uname(struct archive *, la_int64_t); /* "Standard" implementation uses getpwuid_r, getgrgid_r and caches the * results for performance. */ __LA_DECL int archive_read_disk_set_standard_lookup(struct archive *); /* You can install your own lookups if you like. */ __LA_DECL int archive_read_disk_set_gname_lookup(struct archive *, void * /* private_data */, const char *(* /* lookup_fn */)(void *, la_int64_t), void (* /* cleanup_fn */)(void *)); __LA_DECL int archive_read_disk_set_uname_lookup(struct archive *, void * /* private_data */, const char *(* /* lookup_fn */)(void *, la_int64_t), void (* /* cleanup_fn */)(void *)); /* Start traversal. */ __LA_DECL int archive_read_disk_open(struct archive *, const char *); __LA_DECL int archive_read_disk_open_w(struct archive *, const wchar_t *); /* * Request that current entry be visited. If you invoke it on every * directory, you'll get a physical traversal. This is ignored if the * current entry isn't a directory or a link to a directory. So, if * you invoke this on every returned path, you'll get a full logical * traversal. */ __LA_DECL int archive_read_disk_descend(struct archive *); __LA_DECL int archive_read_disk_can_descend(struct archive *); __LA_DECL int archive_read_disk_current_filesystem(struct archive *); __LA_DECL int archive_read_disk_current_filesystem_is_synthetic(struct archive *); __LA_DECL int archive_read_disk_current_filesystem_is_remote(struct archive *); /* Request that the access time of the entry visited by traversal be restored. */ __LA_DECL int archive_read_disk_set_atime_restored(struct archive *); /* * Set behavior. The "flags" argument selects optional behavior. */ /* Request that the access time of the entry visited by traversal be restored. * This is the same as archive_read_disk_set_atime_restored. */ #define ARCHIVE_READDISK_RESTORE_ATIME (0x0001) /* Default: Do not skip an entry which has nodump flags. */ #define ARCHIVE_READDISK_HONOR_NODUMP (0x0002) /* Default: Skip a mac resource fork file whose prefix is "._" because of * using copyfile. */ #define ARCHIVE_READDISK_MAC_COPYFILE (0x0004) /* Default: Traverse mount points. */ #define ARCHIVE_READDISK_NO_TRAVERSE_MOUNTS (0x0008) /* Default: Xattrs are read from disk. */ #define ARCHIVE_READDISK_NO_XATTR (0x0010) /* Default: ACLs are read from disk. */ #define ARCHIVE_READDISK_NO_ACL (0x0020) /* Default: File flags are read from disk. */ #define ARCHIVE_READDISK_NO_FFLAGS (0x0040) __LA_DECL int archive_read_disk_set_behavior(struct archive *, int flags); /* * Set archive_match object that will be used in archive_read_disk to * know whether an entry should be skipped. The callback function * _excluded_func will be invoked when an entry is skipped by the result * of archive_match. */ __LA_DECL int archive_read_disk_set_matching(struct archive *, struct archive *_matching, void (*_excluded_func) (struct archive *, void *, struct archive_entry *), void *_client_data); __LA_DECL int archive_read_disk_set_metadata_filter_callback(struct archive *, int (*_metadata_filter_func)(struct archive *, void *, struct archive_entry *), void *_client_data); /* Simplified cleanup interface; * This calls archive_read_free() or archive_write_free() as needed. */ __LA_DECL int archive_free(struct archive *); /* * Accessor functions to read/set various information in * the struct archive object: */ /* Number of filters in the current filter pipeline. */ /* Filter #0 is the one closest to the format, -1 is a synonym for the * last filter, which is always the pseudo-filter that wraps the * client callbacks. */ __LA_DECL int archive_filter_count(struct archive *); __LA_DECL la_int64_t archive_filter_bytes(struct archive *, int); __LA_DECL int archive_filter_code(struct archive *, int); __LA_DECL const char * archive_filter_name(struct archive *, int); #if ARCHIVE_VERSION_NUMBER < 4000000 /* These don't properly handle multiple filters, so are deprecated and * will eventually be removed. */ /* As of libarchive 3.0, this is an alias for archive_filter_bytes(a, -1); */ __LA_DECL la_int64_t archive_position_compressed(struct archive *) __LA_DEPRECATED; /* As of libarchive 3.0, this is an alias for archive_filter_bytes(a, 0); */ __LA_DECL la_int64_t archive_position_uncompressed(struct archive *) __LA_DEPRECATED; /* As of libarchive 3.0, this is an alias for archive_filter_name(a, 0); */ __LA_DECL const char *archive_compression_name(struct archive *) __LA_DEPRECATED; /* As of libarchive 3.0, this is an alias for archive_filter_code(a, 0); */ __LA_DECL int archive_compression(struct archive *) __LA_DEPRECATED; #endif __LA_DECL int archive_errno(struct archive *); __LA_DECL const char *archive_error_string(struct archive *); __LA_DECL const char *archive_format_name(struct archive *); __LA_DECL int archive_format(struct archive *); __LA_DECL void archive_clear_error(struct archive *); __LA_DECL void archive_set_error(struct archive *, int _err, const char *fmt, ...) __LA_PRINTF(3, 4); __LA_DECL void archive_copy_error(struct archive *dest, struct archive *src); __LA_DECL int archive_file_count(struct archive *); /* * ARCHIVE_MATCH API */ __LA_DECL struct archive *archive_match_new(void); __LA_DECL int archive_match_free(struct archive *); /* * Test if archive_entry is excluded. * This is a convenience function. This is the same as calling all * archive_match_path_excluded, archive_match_time_excluded * and archive_match_owner_excluded. */ __LA_DECL int archive_match_excluded(struct archive *, struct archive_entry *); /* * Test if pathname is excluded. The conditions are set by following functions. */ __LA_DECL int archive_match_path_excluded(struct archive *, struct archive_entry *); /* Add exclusion pathname pattern. */ __LA_DECL int archive_match_exclude_pattern(struct archive *, const char *); __LA_DECL int archive_match_exclude_pattern_w(struct archive *, const wchar_t *); /* Add exclusion pathname pattern from file. */ __LA_DECL int archive_match_exclude_pattern_from_file(struct archive *, const char *, int _nullSeparator); __LA_DECL int archive_match_exclude_pattern_from_file_w(struct archive *, const wchar_t *, int _nullSeparator); /* Add inclusion pathname pattern. */ __LA_DECL int archive_match_include_pattern(struct archive *, const char *); __LA_DECL int archive_match_include_pattern_w(struct archive *, const wchar_t *); /* Add inclusion pathname pattern from file. */ __LA_DECL int archive_match_include_pattern_from_file(struct archive *, const char *, int _nullSeparator); __LA_DECL int archive_match_include_pattern_from_file_w(struct archive *, const wchar_t *, int _nullSeparator); /* * How to get statistic information for inclusion patterns. */ /* Return the amount number of unmatched inclusion patterns. */ __LA_DECL int archive_match_path_unmatched_inclusions(struct archive *); /* Return the pattern of unmatched inclusion with ARCHIVE_OK. * Return ARCHIVE_EOF if there is no inclusion pattern. */ __LA_DECL int archive_match_path_unmatched_inclusions_next( struct archive *, const char **); __LA_DECL int archive_match_path_unmatched_inclusions_next_w( struct archive *, const wchar_t **); /* * Test if a file is excluded by its time stamp. * The conditions are set by following functions. */ __LA_DECL int archive_match_time_excluded(struct archive *, struct archive_entry *); /* * Flags to tell a matching type of time stamps. These are used for * following functions. */ /* Time flag: mtime to be tested. */ #define ARCHIVE_MATCH_MTIME (0x0100) /* Time flag: ctime to be tested. */ #define ARCHIVE_MATCH_CTIME (0x0200) /* Comparison flag: Match the time if it is newer than. */ #define ARCHIVE_MATCH_NEWER (0x0001) /* Comparison flag: Match the time if it is older than. */ #define ARCHIVE_MATCH_OLDER (0x0002) /* Comparison flag: Match the time if it is equal to. */ #define ARCHIVE_MATCH_EQUAL (0x0010) /* Set inclusion time. */ __LA_DECL int archive_match_include_time(struct archive *, int _flag, time_t _sec, long _nsec); /* Set inclusion time by a date string. */ __LA_DECL int archive_match_include_date(struct archive *, int _flag, const char *_datestr); __LA_DECL int archive_match_include_date_w(struct archive *, int _flag, const wchar_t *_datestr); /* Set inclusion time by a particular file. */ __LA_DECL int archive_match_include_file_time(struct archive *, int _flag, const char *_pathname); __LA_DECL int archive_match_include_file_time_w(struct archive *, int _flag, const wchar_t *_pathname); /* Add exclusion entry. */ __LA_DECL int archive_match_exclude_entry(struct archive *, int _flag, struct archive_entry *); /* * Test if a file is excluded by its uid ,gid, uname or gname. * The conditions are set by following functions. */ __LA_DECL int archive_match_owner_excluded(struct archive *, struct archive_entry *); /* Add inclusion uid, gid, uname and gname. */ __LA_DECL int archive_match_include_uid(struct archive *, la_int64_t); __LA_DECL int archive_match_include_gid(struct archive *, la_int64_t); __LA_DECL int archive_match_include_uname(struct archive *, const char *); __LA_DECL int archive_match_include_uname_w(struct archive *, const wchar_t *); __LA_DECL int archive_match_include_gname(struct archive *, const char *); __LA_DECL int archive_match_include_gname_w(struct archive *, const wchar_t *); /* Utility functions */ /* Convenience function to sort a NULL terminated list of strings */ __LA_DECL int archive_utility_string_sort(char **); #ifdef __cplusplus } #endif /* These are meaningless outside of this header. */ #undef __LA_DECL #endif /* !ARCHIVE_H_INCLUDED */ Index: stable/10/contrib/libarchive/libarchive/archive_cmdline.c =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_cmdline.c (revision 324417) +++ stable/10/contrib/libarchive/libarchive/archive_cmdline.c (revision 324418) @@ -1,227 +1,227 @@ /*- * Copyright (c) 2012 Michihiro NAKAJIMA * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "archive_platform.h" __FBSDID("$FreeBSD$"); #ifdef HAVE_STRING_H # include #endif #ifdef HAVE_STDLIB_H # include #endif #include "archive.h" #include "archive_cmdline_private.h" #include "archive_string.h" static int cmdline_set_path(struct archive_cmdline *, const char *); static int cmdline_add_arg(struct archive_cmdline *, const char *); static ssize_t extract_quotation(struct archive_string *as, const char *p) { const char *s; for (s = p + 1; *s;) { if (*s == '\\') { if (s[1] != '\0') { archive_strappend_char(as, s[1]); s += 2; } else s++; } else if (*s == '"') break; else { archive_strappend_char(as, s[0]); s++; } } if (*s != '"') return (ARCHIVE_FAILED);/* Invalid sequence. */ return ((ssize_t)(s + 1 - p)); } static ssize_t get_argument(struct archive_string *as, const char *p) { const char *s = p; archive_string_empty(as); /* Skip beginning space characters. */ while (*s != '\0' && *s == ' ') s++; /* Copy non-space characters. */ while (*s != '\0' && *s != ' ') { if (*s == '\\') { if (s[1] != '\0') { archive_strappend_char(as, s[1]); s += 2; } else { s++;/* Ignore this character.*/ break; } } else if (*s == '"') { ssize_t q = extract_quotation(as, s); if (q < 0) return (ARCHIVE_FAILED);/* Invalid sequence. */ s += q; } else { archive_strappend_char(as, s[0]); s++; } } return ((ssize_t)(s - p)); } /* * Set up command line arguments. - * Returns ARChIVE_OK if everything okey. - * Returns ARChIVE_FAILED if there is a lack of the `"' terminator or an + * Returns ARCHIVE_OK if everything okey. + * Returns ARCHIVE_FAILED if there is a lack of the `"' terminator or an * empty command line. - * Returns ARChIVE_FATAL if no memory. + * Returns ARCHIVE_FATAL if no memory. */ int __archive_cmdline_parse(struct archive_cmdline *data, const char *cmd) { struct archive_string as; const char *p; ssize_t al; int r; archive_string_init(&as); /* Get first argument as a command path. */ al = get_argument(&as, cmd); if (al < 0) { r = ARCHIVE_FAILED;/* Invalid sequence. */ goto exit_function; } if (archive_strlen(&as) == 0) { r = ARCHIVE_FAILED;/* An empty command path. */ goto exit_function; } r = cmdline_set_path(data, as.s); if (r != ARCHIVE_OK) goto exit_function; p = strrchr(as.s, '/'); if (p == NULL) p = as.s; else p++; r = cmdline_add_arg(data, p); if (r != ARCHIVE_OK) goto exit_function; cmd += al; for (;;) { al = get_argument(&as, cmd); if (al < 0) { r = ARCHIVE_FAILED;/* Invalid sequence. */ goto exit_function; } if (al == 0) break; cmd += al; if (archive_strlen(&as) == 0 && *cmd == '\0') break; r = cmdline_add_arg(data, as.s); if (r != ARCHIVE_OK) goto exit_function; } r = ARCHIVE_OK; exit_function: archive_string_free(&as); return (r); } /* * Set the program path. */ static int cmdline_set_path(struct archive_cmdline *data, const char *path) { char *newptr; newptr = realloc(data->path, strlen(path) + 1); if (newptr == NULL) return (ARCHIVE_FATAL); data->path = newptr; strcpy(data->path, path); return (ARCHIVE_OK); } /* * Add a argument for the program. */ static int cmdline_add_arg(struct archive_cmdline *data, const char *arg) { char **newargv; if (data->path == NULL) return (ARCHIVE_FAILED); newargv = realloc(data->argv, (data->argc + 2) * sizeof(char *)); if (newargv == NULL) return (ARCHIVE_FATAL); data->argv = newargv; data->argv[data->argc] = strdup(arg); if (data->argv[data->argc] == NULL) return (ARCHIVE_FATAL); /* Set the terminator of argv. */ data->argv[++data->argc] = NULL; return (ARCHIVE_OK); } struct archive_cmdline * __archive_cmdline_allocate(void) { return (struct archive_cmdline *) calloc(1, sizeof(struct archive_cmdline)); } /* * Release the resources. */ int __archive_cmdline_free(struct archive_cmdline *data) { if (data) { free(data->path); if (data->argv != NULL) { int i; for (i = 0; data->argv[i] != NULL; i++) free(data->argv[i]); free(data->argv); } free(data); } return (ARCHIVE_OK); } Index: stable/10/contrib/libarchive/libarchive/archive_platform.h =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_platform.h (revision 324417) +++ stable/10/contrib/libarchive/libarchive/archive_platform.h (revision 324418) @@ -1,183 +1,194 @@ /*- * Copyright (c) 2003-2007 Tim Kientzle * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. * * $FreeBSD$ */ /* !!ONLY FOR USE INTERNALLY TO LIBARCHIVE!! */ /* * This header is the first thing included in any of the libarchive * source files. As far as possible, platform-specific issues should * be dealt with here and not within individual source files. I'm * actively trying to minimize #if blocks within the main source, * since they obfuscate the code. */ #ifndef ARCHIVE_PLATFORM_H_INCLUDED #define ARCHIVE_PLATFORM_H_INCLUDED /* archive.h and archive_entry.h require this. */ #define __LIBARCHIVE_BUILD 1 #if defined(PLATFORM_CONFIG_H) /* Use hand-built config.h in environments that need it. */ #include PLATFORM_CONFIG_H #elif defined(HAVE_CONFIG_H) /* Most POSIX platforms use the 'configure' script to build config.h */ #include "config.h" #else /* Warn if the library hasn't been (automatically or manually) configured. */ #error Oops: No config.h and no pre-built configuration in archive_platform.h. #endif +/* On macOS check for some symbols based on the deployment target version. */ +#if defined(__APPLE__) +# undef HAVE_FUTIMENS +# undef HAVE_UTIMENSAT +# include +# if MAC_OS_X_VERSION_MIN_REQUIRED >= 101300 +# define HAVE_FUTIMENS 1 +# define HAVE_UTIMENSAT 1 +# endif +#endif + /* It should be possible to get rid of this by extending the feature-test * macros to cover Windows API functions, probably along with non-trivial * refactoring of code to find structures that sit more cleanly on top of * either Windows or Posix APIs. */ #if (defined(__WIN32__) || defined(_WIN32) || defined(__WIN32)) && !defined(__CYGWIN__) #include "archive_windows.h" #endif /* * The config files define a lot of feature macros. The following * uses those macros to select/define replacements and include key * headers as required. */ /* Get a real definition for __FBSDID or __RCSID if we can */ #if HAVE_SYS_CDEFS_H #include #endif /* If not, define them so as to avoid dangling semicolons. */ #ifndef __FBSDID #define __FBSDID(a) struct _undefined_hack #endif #ifndef __RCSID #define __RCSID(a) struct _undefined_hack #endif /* Try to get standard C99-style integer type definitions. */ #if HAVE_INTTYPES_H #include #endif #if HAVE_STDINT_H #include #endif /* Borland warns about its own constants! */ #if defined(__BORLANDC__) # if HAVE_DECL_UINT64_MAX # undef UINT64_MAX # undef HAVE_DECL_UINT64_MAX # endif # if HAVE_DECL_UINT64_MIN # undef UINT64_MIN # undef HAVE_DECL_UINT64_MIN # endif # if HAVE_DECL_INT64_MAX # undef INT64_MAX # undef HAVE_DECL_INT64_MAX # endif # if HAVE_DECL_INT64_MIN # undef INT64_MIN # undef HAVE_DECL_INT64_MIN # endif #endif /* Some platforms lack the standard *_MAX definitions. */ #if !HAVE_DECL_SIZE_MAX #define SIZE_MAX (~(size_t)0) #endif #if !HAVE_DECL_SSIZE_MAX #define SSIZE_MAX ((ssize_t)(SIZE_MAX >> 1)) #endif #if !HAVE_DECL_UINT32_MAX #define UINT32_MAX (~(uint32_t)0) #endif #if !HAVE_DECL_INT32_MAX #define INT32_MAX ((int32_t)(UINT32_MAX >> 1)) #endif #if !HAVE_DECL_INT32_MIN #define INT32_MIN ((int32_t)(~INT32_MAX)) #endif #if !HAVE_DECL_UINT64_MAX #define UINT64_MAX (~(uint64_t)0) #endif #if !HAVE_DECL_INT64_MAX #define INT64_MAX ((int64_t)(UINT64_MAX >> 1)) #endif #if !HAVE_DECL_INT64_MIN #define INT64_MIN ((int64_t)(~INT64_MAX)) #endif #if !HAVE_DECL_UINTMAX_MAX #define UINTMAX_MAX (~(uintmax_t)0) #endif #if !HAVE_DECL_INTMAX_MAX #define INTMAX_MAX ((intmax_t)(UINTMAX_MAX >> 1)) #endif #if !HAVE_DECL_INTMAX_MIN #define INTMAX_MIN ((intmax_t)(~INTMAX_MAX)) #endif /* * If we can't restore metadata using a file descriptor, then * for compatibility's sake, close files before trying to restore metadata. */ #if defined(HAVE_FCHMOD) || defined(HAVE_FUTIMES) || defined(HAVE_ACL_SET_FD) || defined(HAVE_ACL_SET_FD_NP) || defined(HAVE_FCHOWN) #define CAN_RESTORE_METADATA_FD #endif /* * glibc 2.24 deprecates readdir_r */ #if defined(HAVE_READDIR_R) && (!defined(__GLIBC__) || !defined(__GLIBC_MINOR__) || __GLIBC__ < 2 || (__GLIBC__ == 2 && __GLIBC_MINOR__ < 24)) #define USE_READDIR_R 1 #else #undef USE_READDIR_R #endif /* Set up defaults for internal error codes. */ #ifndef ARCHIVE_ERRNO_FILE_FORMAT #if HAVE_EFTYPE #define ARCHIVE_ERRNO_FILE_FORMAT EFTYPE #else #if HAVE_EILSEQ #define ARCHIVE_ERRNO_FILE_FORMAT EILSEQ #else #define ARCHIVE_ERRNO_FILE_FORMAT EINVAL #endif #endif #endif #ifndef ARCHIVE_ERRNO_PROGRAMMER #define ARCHIVE_ERRNO_PROGRAMMER EINVAL #endif #ifndef ARCHIVE_ERRNO_MISC #define ARCHIVE_ERRNO_MISC (-1) #endif #endif /* !ARCHIVE_PLATFORM_H_INCLUDED */ Index: stable/10/contrib/libarchive/libarchive/archive_read_append_filter.c =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_read_append_filter.c (revision 324417) +++ stable/10/contrib/libarchive/libarchive/archive_read_append_filter.c (revision 324418) @@ -1,200 +1,204 @@ /*- * Copyright (c) 2003-2012 Tim Kientzle * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "archive_platform.h" __FBSDID("$FreeBSD$"); #ifdef HAVE_ERRNO_H #include #endif #include "archive.h" #include "archive_private.h" #include "archive_read_private.h" int archive_read_append_filter(struct archive *_a, int code) { int r1, r2, number_bidders, i; char str[20]; struct archive_read_filter_bidder *bidder; struct archive_read_filter *filter; struct archive_read *a = (struct archive_read *)_a; r2 = (ARCHIVE_OK); switch (code) { case ARCHIVE_FILTER_NONE: /* No filter to add, so do nothing. * NOTE: An initial "NONE" type filter is always set at the end of the * filter chain. */ r1 = (ARCHIVE_OK); break; case ARCHIVE_FILTER_GZIP: strcpy(str, "gzip"); r1 = archive_read_support_filter_gzip(_a); break; case ARCHIVE_FILTER_BZIP2: strcpy(str, "bzip2"); r1 = archive_read_support_filter_bzip2(_a); break; case ARCHIVE_FILTER_COMPRESS: strcpy(str, "compress (.Z)"); r1 = archive_read_support_filter_compress(_a); break; case ARCHIVE_FILTER_PROGRAM: archive_set_error(&a->archive, ARCHIVE_ERRNO_PROGRAMMER, "Cannot append program filter using archive_read_append_filter"); return (ARCHIVE_FATAL); case ARCHIVE_FILTER_LZMA: strcpy(str, "lzma"); r1 = archive_read_support_filter_lzma(_a); break; case ARCHIVE_FILTER_XZ: strcpy(str, "xz"); r1 = archive_read_support_filter_xz(_a); break; case ARCHIVE_FILTER_UU: strcpy(str, "uu"); r1 = archive_read_support_filter_uu(_a); break; case ARCHIVE_FILTER_RPM: strcpy(str, "rpm"); r1 = archive_read_support_filter_rpm(_a); break; case ARCHIVE_FILTER_LZ4: strcpy(str, "lz4"); r1 = archive_read_support_filter_lz4(_a); break; + case ARCHIVE_FILTER_ZSTD: + strcpy(str, "zstd"); + r1 = archive_read_support_filter_zstd(_a); + break; case ARCHIVE_FILTER_LZIP: strcpy(str, "lzip"); r1 = archive_read_support_filter_lzip(_a); break; case ARCHIVE_FILTER_LRZIP: strcpy(str, "lrzip"); r1 = archive_read_support_filter_lrzip(_a); break; default: archive_set_error(&a->archive, ARCHIVE_ERRNO_PROGRAMMER, "Invalid filter code specified"); return (ARCHIVE_FATAL); } if (code != ARCHIVE_FILTER_NONE) { number_bidders = sizeof(a->bidders) / sizeof(a->bidders[0]); bidder = a->bidders; for (i = 0; i < number_bidders; i++, bidder++) { if (!bidder->name || !strcmp(bidder->name, str)) break; } if (!bidder->name || strcmp(bidder->name, str)) { archive_set_error(&a->archive, ARCHIVE_ERRNO_PROGRAMMER, "Internal error: Unable to append filter"); return (ARCHIVE_FATAL); } filter = (struct archive_read_filter *)calloc(1, sizeof(*filter)); if (filter == NULL) { archive_set_error(&a->archive, ENOMEM, "Out of memory"); return (ARCHIVE_FATAL); } filter->bidder = bidder; filter->archive = a; filter->upstream = a->filter; a->filter = filter; r2 = (bidder->init)(a->filter); if (r2 != ARCHIVE_OK) { __archive_read_free_filters(a); return (ARCHIVE_FATAL); } } a->bypass_filter_bidding = 1; return (r1 < r2) ? r1 : r2; } int archive_read_append_filter_program(struct archive *_a, const char *cmd) { return (archive_read_append_filter_program_signature(_a, cmd, NULL, 0)); } int archive_read_append_filter_program_signature(struct archive *_a, const char *cmd, const void *signature, size_t signature_len) { int r, number_bidders, i; struct archive_read_filter_bidder *bidder; struct archive_read_filter *filter; struct archive_read *a = (struct archive_read *)_a; if (archive_read_support_filter_program_signature(_a, cmd, signature, signature_len) != (ARCHIVE_OK)) return (ARCHIVE_FATAL); number_bidders = sizeof(a->bidders) / sizeof(a->bidders[0]); bidder = a->bidders; for (i = 0; i < number_bidders; i++, bidder++) { /* Program bidder name set to filter name after initialization */ if (bidder->data && !bidder->name) break; } if (!bidder->data) { archive_set_error(&a->archive, ARCHIVE_ERRNO_PROGRAMMER, "Internal error: Unable to append program filter"); return (ARCHIVE_FATAL); } filter = (struct archive_read_filter *)calloc(1, sizeof(*filter)); if (filter == NULL) { archive_set_error(&a->archive, ENOMEM, "Out of memory"); return (ARCHIVE_FATAL); } filter->bidder = bidder; filter->archive = a; filter->upstream = a->filter; a->filter = filter; r = (bidder->init)(a->filter); if (r != ARCHIVE_OK) { __archive_read_free_filters(a); return (ARCHIVE_FATAL); } bidder->name = a->filter->name; a->bypass_filter_bidding = 1; return r; } Index: stable/10/contrib/libarchive/libarchive/archive_read_disk_entry_from_file.c =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_read_disk_entry_from_file.c (revision 324417) +++ stable/10/contrib/libarchive/libarchive/archive_read_disk_entry_from_file.c (revision 324418) @@ -1,1029 +1,1029 @@ /*- * Copyright (c) 2003-2009 Tim Kientzle * Copyright (c) 2010-2012 Michihiro NAKAJIMA * Copyright (c) 2016 Martin Matuska * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "archive_platform.h" __FBSDID("$FreeBSD"); /* This is the tree-walking code for POSIX systems. */ #if !defined(_WIN32) || defined(__CYGWIN__) #ifdef HAVE_SYS_TYPES_H #include #endif #ifdef HAVE_SYS_EXTATTR_H #include #endif #ifdef HAVE_SYS_IOCTL_H #include #endif #ifdef HAVE_SYS_PARAM_H #include #endif #ifdef HAVE_SYS_STAT_H #include #endif #if defined(HAVE_SYS_XATTR_H) #include #elif defined(HAVE_ATTR_XATTR_H) #include #endif #ifdef HAVE_SYS_EA_H #include #endif #ifdef HAVE_COPYFILE_H #include #endif #ifdef HAVE_ERRNO_H #include #endif #ifdef HAVE_FCNTL_H #include #endif #ifdef HAVE_LIMITS_H #include #endif #ifdef HAVE_LINUX_TYPES_H #include #endif #ifdef HAVE_LINUX_FIEMAP_H #include #endif #ifdef HAVE_LINUX_FS_H #include #endif /* * Some Linux distributions have both linux/ext2_fs.h and ext2fs/ext2_fs.h. * As the include guards don't agree, the order of include is important. */ #ifdef HAVE_LINUX_EXT2_FS_H #include /* for Linux file flags */ #endif #if defined(HAVE_EXT2FS_EXT2_FS_H) && !defined(__CYGWIN__) #include /* Linux file flags, broken on Cygwin */ #endif #ifdef HAVE_PATHS_H #include #endif #ifdef HAVE_UNISTD_H #include #endif #include "archive.h" #include "archive_entry.h" #include "archive_private.h" #include "archive_read_disk_private.h" #ifndef O_CLOEXEC #define O_CLOEXEC 0 #endif static int setup_mac_metadata(struct archive_read_disk *, struct archive_entry *, int *fd); static int setup_xattrs(struct archive_read_disk *, struct archive_entry *, int *fd); static int setup_sparse(struct archive_read_disk *, struct archive_entry *, int *fd); #if defined(HAVE_LINUX_FIEMAP_H) static int setup_sparse_fiemap(struct archive_read_disk *, struct archive_entry *, int *fd); #endif #if !ARCHIVE_ACL_SUPPORT int archive_read_disk_entry_setup_acls(struct archive_read_disk *a, struct archive_entry *entry, int *fd) { (void)a; /* UNUSED */ (void)entry; /* UNUSED */ (void)fd; /* UNUSED */ return (ARCHIVE_OK); } #endif /* * Enter working directory and return working pathname of archive_entry. * If a pointer to an integer is provided and its value is below zero - * open a file descriptor on this pahtname. + * open a file descriptor on this pathname. */ const char * archive_read_disk_entry_setup_path(struct archive_read_disk *a, struct archive_entry *entry, int *fd) { const char *path; path = archive_entry_sourcepath(entry); if (path == NULL || (a->tree != NULL && a->tree_enter_working_dir(a->tree) != 0)) path = archive_entry_pathname(entry); if (path == NULL) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Couldn't determine path"); } else if (fd != NULL && *fd < 0 && a->tree != NULL && (a->follow_symlinks || archive_entry_filetype(entry) != AE_IFLNK)) { *fd = a->open_on_current_dir(a->tree, path, O_RDONLY | O_NONBLOCK); } return (path); } int archive_read_disk_entry_from_file(struct archive *_a, struct archive_entry *entry, int fd, const struct stat *st) { struct archive_read_disk *a = (struct archive_read_disk *)_a; const char *path, *name; struct stat s; int initial_fd = fd; int r, r1; archive_clear_error(_a); path = archive_entry_sourcepath(entry); if (path == NULL) path = archive_entry_pathname(entry); if (a->tree == NULL) { if (st == NULL) { #if HAVE_FSTAT if (fd >= 0) { if (fstat(fd, &s) != 0) { archive_set_error(&a->archive, errno, "Can't fstat"); return (ARCHIVE_FAILED); } } else #endif #if HAVE_LSTAT if (!a->follow_symlinks) { if (lstat(path, &s) != 0) { archive_set_error(&a->archive, errno, "Can't lstat %s", path); return (ARCHIVE_FAILED); } } else #endif if (stat(path, &s) != 0) { archive_set_error(&a->archive, errno, "Can't stat %s", path); return (ARCHIVE_FAILED); } st = &s; } archive_entry_copy_stat(entry, st); } /* Lookup uname/gname */ name = archive_read_disk_uname(_a, archive_entry_uid(entry)); if (name != NULL) archive_entry_copy_uname(entry, name); name = archive_read_disk_gname(_a, archive_entry_gid(entry)); if (name != NULL) archive_entry_copy_gname(entry, name); #ifdef HAVE_STRUCT_STAT_ST_FLAGS /* On FreeBSD, we get flags for free with the stat. */ /* TODO: Does this belong in copy_stat()? */ if ((a->flags & ARCHIVE_READDISK_NO_FFLAGS) == 0 && st->st_flags != 0) archive_entry_set_fflags(entry, st->st_flags, 0); #endif #if (defined(FS_IOC_GETFLAGS) && defined(HAVE_WORKING_FS_IOC_GETFLAGS)) || \ (defined(EXT2_IOC_GETFLAGS) && defined(HAVE_WORKING_EXT2_IOC_GETFLAGS)) /* Linux requires an extra ioctl to pull the flags. Although * this is an extra step, it has a nice side-effect: We get an * open file descriptor which we can use in the subsequent lookups. */ if ((a->flags & ARCHIVE_READDISK_NO_FFLAGS) == 0 && (S_ISREG(st->st_mode) || S_ISDIR(st->st_mode))) { if (fd < 0) { if (a->tree != NULL) fd = a->open_on_current_dir(a->tree, path, O_RDONLY | O_NONBLOCK | O_CLOEXEC); else fd = open(path, O_RDONLY | O_NONBLOCK | O_CLOEXEC); __archive_ensure_cloexec_flag(fd); } if (fd >= 0) { int stflags; r = ioctl(fd, #if defined(FS_IOC_GETFLAGS) FS_IOC_GETFLAGS, #else EXT2_IOC_GETFLAGS, #endif &stflags); if (r == 0 && stflags != 0) archive_entry_set_fflags(entry, stflags, 0); } } #endif #if defined(HAVE_READLINK) || defined(HAVE_READLINKAT) if (S_ISLNK(st->st_mode)) { size_t linkbuffer_len = st->st_size + 1; char *linkbuffer; int lnklen; linkbuffer = malloc(linkbuffer_len); if (linkbuffer == NULL) { archive_set_error(&a->archive, ENOMEM, "Couldn't read link data"); return (ARCHIVE_FAILED); } if (a->tree != NULL) { #ifdef HAVE_READLINKAT lnklen = readlinkat(a->tree_current_dir_fd(a->tree), path, linkbuffer, linkbuffer_len); #else if (a->tree_enter_working_dir(a->tree) != 0) { archive_set_error(&a->archive, errno, "Couldn't read link data"); free(linkbuffer); return (ARCHIVE_FAILED); } lnklen = readlink(path, linkbuffer, linkbuffer_len); #endif /* HAVE_READLINKAT */ } else lnklen = readlink(path, linkbuffer, linkbuffer_len); if (lnklen < 0) { archive_set_error(&a->archive, errno, "Couldn't read link data"); free(linkbuffer); return (ARCHIVE_FAILED); } linkbuffer[lnklen] = 0; archive_entry_set_symlink(entry, linkbuffer); free(linkbuffer); } #endif /* HAVE_READLINK || HAVE_READLINKAT */ r = 0; if ((a->flags & ARCHIVE_READDISK_NO_ACL) == 0) r = archive_read_disk_entry_setup_acls(a, entry, &fd); if ((a->flags & ARCHIVE_READDISK_NO_XATTR) == 0) { r1 = setup_xattrs(a, entry, &fd); if (r1 < r) r = r1; } if (a->flags & ARCHIVE_READDISK_MAC_COPYFILE) { r1 = setup_mac_metadata(a, entry, &fd); if (r1 < r) r = r1; } r1 = setup_sparse(a, entry, &fd); if (r1 < r) r = r1; /* If we opened the file earlier in this function, close it. */ if (initial_fd != fd) close(fd); return (r); } #if defined(__APPLE__) && defined(HAVE_COPYFILE_H) /* * The Mac OS "copyfile()" API copies the extended metadata for a * file into a separate file in AppleDouble format (see RFC 1740). * * Mac OS tar and cpio implementations store this extended * metadata as a separate entry just before the regular entry * with a "._" prefix added to the filename. * * Note that this is currently done unconditionally; the tar program has * an option to discard this information before the archive is written. * * TODO: If there's a failure, report it and return ARCHIVE_WARN. */ static int setup_mac_metadata(struct archive_read_disk *a, struct archive_entry *entry, int *fd) { int tempfd = -1; int copyfile_flags = COPYFILE_NOFOLLOW | COPYFILE_ACL | COPYFILE_XATTR; struct stat copyfile_stat; int ret = ARCHIVE_OK; void *buff = NULL; int have_attrs; const char *name, *tempdir; struct archive_string tempfile; (void)fd; /* UNUSED */ name = archive_read_disk_entry_setup_path(a, entry, NULL); if (name == NULL) return (ARCHIVE_WARN); /* Short-circuit if there's nothing to do. */ have_attrs = copyfile(name, NULL, 0, copyfile_flags | COPYFILE_CHECK); if (have_attrs == -1) { archive_set_error(&a->archive, errno, "Could not check extended attributes"); return (ARCHIVE_WARN); } if (have_attrs == 0) return (ARCHIVE_OK); tempdir = NULL; if (issetugid() == 0) tempdir = getenv("TMPDIR"); if (tempdir == NULL) tempdir = _PATH_TMP; archive_string_init(&tempfile); archive_strcpy(&tempfile, tempdir); archive_strcat(&tempfile, "tar.md.XXXXXX"); tempfd = mkstemp(tempfile.s); if (tempfd < 0) { archive_set_error(&a->archive, errno, "Could not open extended attribute file"); ret = ARCHIVE_WARN; goto cleanup; } __archive_ensure_cloexec_flag(tempfd); /* XXX I wish copyfile() could pack directly to a memory * buffer; that would avoid the temp file here. For that * matter, it would be nice if fcopyfile() actually worked, * that would reduce the many open/close races here. */ if (copyfile(name, tempfile.s, 0, copyfile_flags | COPYFILE_PACK)) { archive_set_error(&a->archive, errno, "Could not pack extended attributes"); ret = ARCHIVE_WARN; goto cleanup; } if (fstat(tempfd, ©file_stat)) { archive_set_error(&a->archive, errno, "Could not check size of extended attributes"); ret = ARCHIVE_WARN; goto cleanup; } buff = malloc(copyfile_stat.st_size); if (buff == NULL) { archive_set_error(&a->archive, errno, "Could not allocate memory for extended attributes"); ret = ARCHIVE_WARN; goto cleanup; } if (copyfile_stat.st_size != read(tempfd, buff, copyfile_stat.st_size)) { archive_set_error(&a->archive, errno, "Could not read extended attributes into memory"); ret = ARCHIVE_WARN; goto cleanup; } archive_entry_copy_mac_metadata(entry, buff, copyfile_stat.st_size); cleanup: if (tempfd >= 0) { close(tempfd); unlink(tempfile.s); } archive_string_free(&tempfile); free(buff); return (ret); } #else /* * Stub implementation for non-Mac systems. */ static int setup_mac_metadata(struct archive_read_disk *a, struct archive_entry *entry, int *fd) { (void)a; /* UNUSED */ (void)entry; /* UNUSED */ (void)fd; /* UNUSED */ return (ARCHIVE_OK); } #endif #if ARCHIVE_XATTR_LINUX || ARCHIVE_XATTR_DARWIN || ARCHIVE_XATTR_AIX /* * Linux, Darwin and AIX extended attribute support. * * TODO: By using a stack-allocated buffer for the first * call to getxattr(), we might be able to avoid the second * call entirely. We only need the second call if the * stack-allocated buffer is too small. But a modest buffer * of 1024 bytes or so will often be big enough. Same applies * to listxattr(). */ static int setup_xattr(struct archive_read_disk *a, struct archive_entry *entry, const char *name, int fd, const char *accpath) { ssize_t size; void *value = NULL; if (fd >= 0) { #if ARCHIVE_XATTR_LINUX size = fgetxattr(fd, name, NULL, 0); #elif ARCHIVE_XATTR_DARWIN size = fgetxattr(fd, name, NULL, 0, 0, 0); #elif ARCHIVE_XATTR_AIX size = fgetea(fd, name, NULL, 0); #endif } else if (!a->follow_symlinks) { #if ARCHIVE_XATTR_LINUX size = lgetxattr(accpath, name, NULL, 0); #elif ARCHIVE_XATTR_DARWIN size = getxattr(accpath, name, NULL, 0, 0, XATTR_NOFOLLOW); #elif ARCHIVE_XATTR_AIX size = lgetea(accpath, name, NULL, 0); #endif } else { #if ARCHIVE_XATTR_LINUX size = getxattr(accpath, name, NULL, 0); #elif ARCHIVE_XATTR_DARWIN size = getxattr(accpath, name, NULL, 0, 0, 0); #elif ARCHIVE_XATTR_AIX size = getea(accpath, name, NULL, 0); #endif } if (size == -1) { archive_set_error(&a->archive, errno, "Couldn't query extended attribute"); return (ARCHIVE_WARN); } if (size > 0 && (value = malloc(size)) == NULL) { archive_set_error(&a->archive, errno, "Out of memory"); return (ARCHIVE_FATAL); } if (fd >= 0) { #if ARCHIVE_XATTR_LINUX size = fgetxattr(fd, name, value, size); #elif ARCHIVE_XATTR_DARWIN size = fgetxattr(fd, name, value, size, 0, 0); #elif ARCHIVE_XATTR_AIX size = fgetea(fd, name, value, size); #endif } else if (!a->follow_symlinks) { #if ARCHIVE_XATTR_LINUX size = lgetxattr(accpath, name, value, size); #elif ARCHIVE_XATTR_DARWIN size = getxattr(accpath, name, value, size, 0, XATTR_NOFOLLOW); #elif ARCHIVE_XATTR_AIX size = lgetea(accpath, name, value, size); #endif } else { #if ARCHIVE_XATTR_LINUX size = getxattr(accpath, name, value, size); #elif ARCHIVE_XATTR_DARWIN size = getxattr(accpath, name, value, size, 0, 0); #elif ARCHIVE_XATTR_AIX size = getea(accpath, name, value, size); #endif } if (size == -1) { archive_set_error(&a->archive, errno, "Couldn't read extended attribute"); return (ARCHIVE_WARN); } archive_entry_xattr_add_entry(entry, name, value, size); free(value); return (ARCHIVE_OK); } static int setup_xattrs(struct archive_read_disk *a, struct archive_entry *entry, int *fd) { char *list, *p; const char *path; ssize_t list_size; path = NULL; if (*fd < 0) { path = archive_read_disk_entry_setup_path(a, entry, fd); if (path == NULL) return (ARCHIVE_WARN); } if (*fd >= 0) { #if ARCHIVE_XATTR_LINUX list_size = flistxattr(*fd, NULL, 0); #elif ARCHIVE_XATTR_DARWIN list_size = flistxattr(*fd, NULL, 0, 0); #elif ARCHIVE_XATTR_AIX list_size = flistea(*fd, NULL, 0); #endif } else if (!a->follow_symlinks) { #if ARCHIVE_XATTR_LINUX list_size = llistxattr(path, NULL, 0); #elif ARCHIVE_XATTR_DARWIN list_size = listxattr(path, NULL, 0, XATTR_NOFOLLOW); #elif ARCHIVE_XATTR_AIX list_size = llistea(path, NULL, 0); #endif } else { #if ARCHIVE_XATTR_LINUX list_size = listxattr(path, NULL, 0); #elif ARCHIVE_XATTR_DARWIN list_size = listxattr(path, NULL, 0, 0); #elif ARCHIVE_XATTR_AIX list_size = listea(path, NULL, 0); #endif } if (list_size == -1) { if (errno == ENOTSUP || errno == ENOSYS) return (ARCHIVE_OK); archive_set_error(&a->archive, errno, "Couldn't list extended attributes"); return (ARCHIVE_WARN); } if (list_size == 0) return (ARCHIVE_OK); if ((list = malloc(list_size)) == NULL) { archive_set_error(&a->archive, errno, "Out of memory"); return (ARCHIVE_FATAL); } if (*fd >= 0) { #if ARCHIVE_XATTR_LINUX list_size = flistxattr(*fd, list, list_size); #elif ARCHIVE_XATTR_DARWIN list_size = flistxattr(*fd, list, list_size, 0); #elif ARCHIVE_XATTR_AIX list_size = flistea(*fd, list, list_size); #endif } else if (!a->follow_symlinks) { #if ARCHIVE_XATTR_LINUX list_size = llistxattr(path, list, list_size); #elif ARCHIVE_XATTR_DARWIN list_size = listxattr(path, list, list_size, XATTR_NOFOLLOW); #elif ARCHIVE_XATTR_AIX list_size = llistea(path, list, list_size); #endif } else { #if ARCHIVE_XATTR_LINUX list_size = listxattr(path, list, list_size); #elif ARCHIVE_XATTR_DARWIN list_size = listxattr(path, list, list_size, 0); #elif ARCHIVE_XATTR_AIX list_size = listea(path, list, list_size); #endif } if (list_size == -1) { archive_set_error(&a->archive, errno, "Couldn't retrieve extended attributes"); free(list); return (ARCHIVE_WARN); } for (p = list; (p - list) < list_size; p += strlen(p) + 1) { if (strncmp(p, "system.", 7) == 0 || strncmp(p, "xfsroot.", 8) == 0) continue; setup_xattr(a, entry, p, *fd, path); } free(list); return (ARCHIVE_OK); } #elif ARCHIVE_XATTR_FREEBSD /* * FreeBSD extattr interface. */ /* TODO: Implement this. Follow the Linux model above, but * with FreeBSD-specific system calls, of course. Be careful * to not include the system extattrs that hold ACLs; we handle * those separately. */ static int setup_xattr(struct archive_read_disk *a, struct archive_entry *entry, int namespace, const char *name, const char *fullname, int fd, const char *path); static int setup_xattr(struct archive_read_disk *a, struct archive_entry *entry, int namespace, const char *name, const char *fullname, int fd, const char *accpath) { ssize_t size; void *value = NULL; if (fd >= 0) size = extattr_get_fd(fd, namespace, name, NULL, 0); else if (!a->follow_symlinks) size = extattr_get_link(accpath, namespace, name, NULL, 0); else size = extattr_get_file(accpath, namespace, name, NULL, 0); if (size == -1) { archive_set_error(&a->archive, errno, "Couldn't query extended attribute"); return (ARCHIVE_WARN); } if (size > 0 && (value = malloc(size)) == NULL) { archive_set_error(&a->archive, errno, "Out of memory"); return (ARCHIVE_FATAL); } if (fd >= 0) size = extattr_get_fd(fd, namespace, name, value, size); else if (!a->follow_symlinks) size = extattr_get_link(accpath, namespace, name, value, size); else size = extattr_get_file(accpath, namespace, name, value, size); if (size == -1) { free(value); archive_set_error(&a->archive, errno, "Couldn't read extended attribute"); return (ARCHIVE_WARN); } archive_entry_xattr_add_entry(entry, fullname, value, size); free(value); return (ARCHIVE_OK); } static int setup_xattrs(struct archive_read_disk *a, struct archive_entry *entry, int *fd) { char buff[512]; char *list, *p; ssize_t list_size; const char *path; int namespace = EXTATTR_NAMESPACE_USER; path = NULL; if (*fd < 0) { path = archive_read_disk_entry_setup_path(a, entry, fd); if (path == NULL) return (ARCHIVE_WARN); } if (*fd >= 0) list_size = extattr_list_fd(*fd, namespace, NULL, 0); else if (!a->follow_symlinks) list_size = extattr_list_link(path, namespace, NULL, 0); else list_size = extattr_list_file(path, namespace, NULL, 0); if (list_size == -1 && errno == EOPNOTSUPP) return (ARCHIVE_OK); if (list_size == -1) { archive_set_error(&a->archive, errno, "Couldn't list extended attributes"); return (ARCHIVE_WARN); } if (list_size == 0) return (ARCHIVE_OK); if ((list = malloc(list_size)) == NULL) { archive_set_error(&a->archive, errno, "Out of memory"); return (ARCHIVE_FATAL); } if (*fd >= 0) list_size = extattr_list_fd(*fd, namespace, list, list_size); else if (!a->follow_symlinks) list_size = extattr_list_link(path, namespace, list, list_size); else list_size = extattr_list_file(path, namespace, list, list_size); if (list_size == -1) { archive_set_error(&a->archive, errno, "Couldn't retrieve extended attributes"); free(list); return (ARCHIVE_WARN); } p = list; while ((p - list) < list_size) { size_t len = 255 & (int)*p; char *name; strcpy(buff, "user."); name = buff + strlen(buff); memcpy(name, p + 1, len); name[len] = '\0'; setup_xattr(a, entry, namespace, name, buff, *fd, path); p += 1 + len; } free(list); return (ARCHIVE_OK); } #else /* * Generic (stub) extended attribute support. */ static int setup_xattrs(struct archive_read_disk *a, struct archive_entry *entry, int *fd) { (void)a; /* UNUSED */ (void)entry; /* UNUSED */ (void)fd; /* UNUSED */ return (ARCHIVE_OK); } #endif #if defined(HAVE_LINUX_FIEMAP_H) /* * Linux FIEMAP sparse interface. * * The FIEMAP ioctl returns an "extent" for each physical allocation * on disk. We need to process those to generate a more compact list * of logical file blocks. We also need to be very careful to use * FIEMAP_FLAG_SYNC here, since there are reports that Linux sometimes * does not report allocations for newly-written data that hasn't * been synced to disk. * * It's important to return a minimal sparse file list because we want * to not trigger sparse file extensions if we don't have to, since * not all readers support them. */ static int setup_sparse_fiemap(struct archive_read_disk *a, struct archive_entry *entry, int *fd) { char buff[4096]; struct fiemap *fm; struct fiemap_extent *fe; int64_t size; int count, do_fiemap, iters; int exit_sts = ARCHIVE_OK; const char *path; if (archive_entry_filetype(entry) != AE_IFREG || archive_entry_size(entry) <= 0 || archive_entry_hardlink(entry) != NULL) return (ARCHIVE_OK); if (*fd < 0) { path = archive_read_disk_entry_setup_path(a, entry, NULL); if (path == NULL) return (ARCHIVE_FAILED); if (a->tree != NULL) *fd = a->open_on_current_dir(a->tree, path, O_RDONLY | O_NONBLOCK | O_CLOEXEC); else *fd = open(path, O_RDONLY | O_NONBLOCK | O_CLOEXEC); if (*fd < 0) { archive_set_error(&a->archive, errno, "Can't open `%s'", path); return (ARCHIVE_FAILED); } __archive_ensure_cloexec_flag(*fd); } /* Initialize buffer to avoid the error valgrind complains about. */ memset(buff, 0, sizeof(buff)); count = (sizeof(buff) - sizeof(*fm))/sizeof(*fe); fm = (struct fiemap *)buff; fm->fm_start = 0; fm->fm_length = ~0ULL;; fm->fm_flags = FIEMAP_FLAG_SYNC; fm->fm_extent_count = count; do_fiemap = 1; size = archive_entry_size(entry); for (iters = 0; ; ++iters) { int i, r; r = ioctl(*fd, FS_IOC_FIEMAP, fm); if (r < 0) { /* When something error happens, it is better we * should return ARCHIVE_OK because an earlier * version(<2.6.28) cannot perform FS_IOC_FIEMAP. */ goto exit_setup_sparse_fiemap; } if (fm->fm_mapped_extents == 0) { if (iters == 0) { /* Fully sparse file; insert a zero-length "data" entry */ archive_entry_sparse_add_entry(entry, 0, 0); } break; } fe = fm->fm_extents; for (i = 0; i < (int)fm->fm_mapped_extents; i++, fe++) { if (!(fe->fe_flags & FIEMAP_EXTENT_UNWRITTEN)) { /* The fe_length of the last block does not * adjust itself to its size files. */ int64_t length = fe->fe_length; if (fe->fe_logical + length > (uint64_t)size) length -= fe->fe_logical + length - size; if (fe->fe_logical == 0 && length == size) { /* This is not sparse. */ do_fiemap = 0; break; } if (length > 0) archive_entry_sparse_add_entry(entry, fe->fe_logical, length); } if (fe->fe_flags & FIEMAP_EXTENT_LAST) do_fiemap = 0; } if (do_fiemap) { fe = fm->fm_extents + fm->fm_mapped_extents -1; fm->fm_start = fe->fe_logical + fe->fe_length; } else break; } exit_setup_sparse_fiemap: return (exit_sts); } #if !defined(SEEK_HOLE) || !defined(SEEK_DATA) static int setup_sparse(struct archive_read_disk *a, struct archive_entry *entry, int *fd) { return setup_sparse_fiemap(a, entry, fd); } #endif #endif /* defined(HAVE_LINUX_FIEMAP_H) */ #if defined(SEEK_HOLE) && defined(SEEK_DATA) /* * SEEK_HOLE sparse interface (FreeBSD, Linux, Solaris) */ static int setup_sparse(struct archive_read_disk *a, struct archive_entry *entry, int *fd) { int64_t size; off_t initial_off; off_t off_s, off_e; int exit_sts = ARCHIVE_OK; int check_fully_sparse = 0; const char *path; if (archive_entry_filetype(entry) != AE_IFREG || archive_entry_size(entry) <= 0 || archive_entry_hardlink(entry) != NULL) return (ARCHIVE_OK); /* Does filesystem support the reporting of hole ? */ if (*fd < 0) path = archive_read_disk_entry_setup_path(a, entry, fd); else path = NULL; if (*fd >= 0) { #ifdef _PC_MIN_HOLE_SIZE if (fpathconf(*fd, _PC_MIN_HOLE_SIZE) <= 0) return (ARCHIVE_OK); #endif initial_off = lseek(*fd, 0, SEEK_CUR); if (initial_off != 0) lseek(*fd, 0, SEEK_SET); } else { if (path == NULL) return (ARCHIVE_FAILED); #ifdef _PC_MIN_HOLE_SIZE if (pathconf(path, _PC_MIN_HOLE_SIZE) <= 0) return (ARCHIVE_OK); #endif *fd = open(path, O_RDONLY | O_NONBLOCK | O_CLOEXEC); if (*fd < 0) { archive_set_error(&a->archive, errno, "Can't open `%s'", path); return (ARCHIVE_FAILED); } __archive_ensure_cloexec_flag(*fd); initial_off = 0; } #ifndef _PC_MIN_HOLE_SIZE /* Check if the underlying filesystem supports seek hole */ off_s = lseek(*fd, 0, SEEK_HOLE); if (off_s < 0) #if defined(HAVE_LINUX_FIEMAP_H) return setup_sparse_fiemap(a, entry, fd); #else goto exit_setup_sparse; #endif else if (off_s > 0) lseek(*fd, 0, SEEK_SET); #endif off_s = 0; size = archive_entry_size(entry); while (off_s < size) { off_s = lseek(*fd, off_s, SEEK_DATA); if (off_s == (off_t)-1) { if (errno == ENXIO) { /* no more hole */ if (archive_entry_sparse_count(entry) == 0) { /* Potentially a fully-sparse file. */ check_fully_sparse = 1; } break; } archive_set_error(&a->archive, errno, "lseek(SEEK_HOLE) failed"); exit_sts = ARCHIVE_FAILED; goto exit_setup_sparse; } off_e = lseek(*fd, off_s, SEEK_HOLE); if (off_e == (off_t)-1) { if (errno == ENXIO) { off_e = lseek(*fd, 0, SEEK_END); if (off_e != (off_t)-1) break;/* no more data */ } archive_set_error(&a->archive, errno, "lseek(SEEK_DATA) failed"); exit_sts = ARCHIVE_FAILED; goto exit_setup_sparse; } if (off_s == 0 && off_e == size) break;/* This is not sparse. */ archive_entry_sparse_add_entry(entry, off_s, off_e - off_s); off_s = off_e; } if (check_fully_sparse) { if (lseek(*fd, 0, SEEK_HOLE) == 0 && lseek(*fd, 0, SEEK_END) == size) { /* Fully sparse file; insert a zero-length "data" entry */ archive_entry_sparse_add_entry(entry, 0, 0); } } exit_setup_sparse: lseek(*fd, initial_off, SEEK_SET); return (exit_sts); } #elif !defined(HAVE_LINUX_FIEMAP_H) /* * Generic (stub) sparse support. */ static int setup_sparse(struct archive_read_disk *a, struct archive_entry *entry, int *fd) { (void)a; /* UNUSED */ (void)entry; /* UNUSED */ (void)fd; /* UNUSED */ return (ARCHIVE_OK); } #endif #endif /* !defined(_WIN32) || defined(__CYGWIN__) */ Index: stable/10/contrib/libarchive/libarchive/archive_read_filter.3 =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_read_filter.3 (revision 324417) +++ stable/10/contrib/libarchive/libarchive/archive_read_filter.3 (revision 324418) @@ -1,150 +1,154 @@ .\" Copyright (c) 2003-2011 Tim Kientzle .\" All rights reserved. .\" .\" Redistribution and use in source and binary forms, with or without .\" modification, are permitted provided that the following conditions .\" are met: .\" 1. Redistributions of source code must retain the above copyright .\" notice, this list of conditions and the following disclaimer. .\" 2. Redistributions in binary form must reproduce the above copyright .\" notice, this list of conditions and the following disclaimer in the .\" documentation and/or other materials provided with the distribution. .\" .\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE .\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF .\" SUCH DAMAGE. .\" .\" $FreeBSD$ .\" .Dd August 14, 2014 .Dt ARCHIVE_READ_FILTER 3 .Os .Sh NAME .Nm archive_read_support_filter_all , .Nm archive_read_support_filter_bzip2 , .Nm archive_read_support_filter_compress , .Nm archive_read_support_filter_gzip , .Nm archive_read_support_filter_lz4 , .Nm archive_read_support_filter_lzma , .Nm archive_read_support_filter_none , .Nm archive_read_support_filter_rpm , .Nm archive_read_support_filter_uu , .Nm archive_read_support_filter_xz , +.Nm archive_read_support_filter_zstd , .Nm archive_read_support_filter_program , .Nm archive_read_support_filter_program_signature .Nd functions for reading streaming archives .\" .Sh LIBRARY Streaming Archive Library (libarchive, -larchive) .Sh SYNOPSIS .In archive.h .Ft int .Fn archive_read_support_filter_all "struct archive *" .Ft int .Fn archive_read_support_filter_bzip2 "struct archive *" .Ft int .Fn archive_read_support_filter_compress "struct archive *" .Ft int .Fn archive_read_support_filter_grzip "struct archive *" .Ft int .Fn archive_read_support_filter_gzip "struct archive *" .Ft int .Fn archive_read_support_filter_lrzip "struct archive *" .Ft int .Fn archive_read_support_filter_lz4 "struct archive *" .Ft int .Fn archive_read_support_filter_lzma "struct archive *" .Ft int .Fn archive_read_support_filter_lzop "struct archive *" .Ft int .Fn archive_read_support_filter_none "struct archive *" .Ft int .Fn archive_read_support_filter_rpm "struct archive *" .Ft int .Fn archive_read_support_filter_uu "struct archive *" .Ft int .Fn archive_read_support_filter_xz "struct archive *" .Ft int +.Fn archive_read_support_filter_zstd "struct archive *" +.Ft int .Fo archive_read_support_filter_program .Fa "struct archive *" .Fa "const char *cmd" .Fc .Ft int .Fo archive_read_support_filter_program_signature .Fa "struct archive *" .Fa "const char *cmd" .Fa "const void *signature" .Fa "size_t signature_length" .Fc .\" .Sh DESCRIPTION .Bl -tag -compact -width indent .It Xo .Fn archive_read_support_filter_bzip2 , .Fn archive_read_support_filter_compress , .Fn archive_read_support_filter_grzip , .Fn archive_read_support_filter_gzip , .Fn archive_read_support_filter_lrzip , .Fn archive_read_support_filter_lz4 , .Fn archive_read_support_filter_lzma , .Fn archive_read_support_filter_lzop , .Fn archive_read_support_filter_none , .Fn archive_read_support_filter_rpm , .Fn archive_read_support_filter_uu , -.Fn archive_read_support_filter_xz +.Fn archive_read_support_filter_xz , +.Fn archive_read_support_filter_zstd , .Xc Enables auto-detection code and decompression support for the specified compression. These functions may fall back on external programs if an appropriate library was not available at build time. Decompression using an external program is usually slower than decompression through built-in libraries. Note that .Dq none is always enabled by default. .It Fn archive_read_support_filter_all Enables all available decompression filters. .It Fn archive_read_support_filter_program Data is fed through the specified external program before being dearchived. Note that this disables automatic detection of the compression format, so it makes no sense to specify this in conjunction with any other decompression option. .It Fn archive_read_support_filter_program_signature This feeds data through the specified external program but only if the initial bytes of the data match the specified signature value. .El .\" .\". Sh EXAMPLE .\" .Sh RETURN VALUES These functions return .Cm ARCHIVE_OK if the compression is fully supported, .Cm ARCHIVE_WARN if the compression is supported only through an external program. .Pp .Fn archive_read_support_filter_none always succeeds. .\" .Sh ERRORS Detailed error codes and textual descriptions are available from the .Fn archive_errno and .Fn archive_error_string functions. .\" .Sh SEE ALSO .Xr libarchive 3 , .Xr archive_read 3 , .Xr archive_read_data 3 , .Xr archive_read_format 3 , .Xr archive_read_format 3 Index: stable/10/contrib/libarchive/libarchive/archive_read_support_filter_all.c =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_read_support_filter_all.c (revision 324417) +++ stable/10/contrib/libarchive/libarchive/archive_read_support_filter_all.c (revision 324418) @@ -1,83 +1,85 @@ /*- * Copyright (c) 2003-2011 Tim Kientzle * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "archive_platform.h" __FBSDID("$FreeBSD$"); #include "archive.h" #include "archive_private.h" #if ARCHIVE_VERSION_NUMBER < 4000000 /* Deprecated; remove in libarchive 4.0 */ int archive_read_support_compression_all(struct archive *a) { return archive_read_support_filter_all(a); } #endif int archive_read_support_filter_all(struct archive *a) { archive_check_magic(a, ARCHIVE_READ_MAGIC, ARCHIVE_STATE_NEW, "archive_read_support_filter_all"); /* Bzip falls back to "bunzip2" command-line */ archive_read_support_filter_bzip2(a); /* The decompress code doesn't use an outside library. */ archive_read_support_filter_compress(a); /* Gzip decompress falls back to "gzip -d" command-line. */ archive_read_support_filter_gzip(a); /* Lzip falls back to "unlzip" command-line program. */ archive_read_support_filter_lzip(a); /* The LZMA file format has a very weak signature, so it * may not be feasible to keep this here, but we'll try. * This will come back out if there are problems. */ /* Lzma falls back to "unlzma" command-line program. */ archive_read_support_filter_lzma(a); /* Xz falls back to "unxz" command-line program. */ archive_read_support_filter_xz(a); /* The decode code doesn't use an outside library. */ archive_read_support_filter_uu(a); /* The decode code doesn't use an outside library. */ archive_read_support_filter_rpm(a); /* The decode code always uses "lrzip -q -d" command-line. */ archive_read_support_filter_lrzip(a); /* Lzop decompress falls back to "lzop -d" command-line. */ archive_read_support_filter_lzop(a); /* The decode code always uses "grzip -d" command-line. */ archive_read_support_filter_grzip(a); /* Lz4 falls back to "lz4 -d" command-line program. */ archive_read_support_filter_lz4(a); + /* Zstd falls back to "zstd -d" command-line program. */ + archive_read_support_filter_zstd(a); /* Note: We always return ARCHIVE_OK here, even if some of the * above return ARCHIVE_WARN. The intent here is to enable * "as much as possible." Clients who need specific * compression should enable those individually so they can * verify the level of support. */ /* Clear any warning messages set by the above functions. */ archive_clear_error(a); return (ARCHIVE_OK); } Index: stable/10/contrib/libarchive/libarchive/archive_read_support_filter_zstd.c =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_read_support_filter_zstd.c (nonexistent) +++ stable/10/contrib/libarchive/libarchive/archive_read_support_filter_zstd.c (revision 324418) @@ -0,0 +1,292 @@ +/*- + * Copyright (c) 2009-2011 Sean Purcell + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES + * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. + * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, + * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT + * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF + * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ + +#include "archive_platform.h" + +__FBSDID("$FreeBSD$"); + +#ifdef HAVE_ERRNO_H +#include +#endif + +#ifdef HAVE_ERRNO_H +#include +#endif +#include +#ifdef HAVE_STDLIB_H +#include +#endif +#ifdef HAVE_STRING_H +#include +#endif +#ifdef HAVE_UNISTD_H +#include +#endif +#if HAVE_ZSTD_H +#include +#endif + +#include "archive.h" +#include "archive_endian.h" +#include "archive_private.h" +#include "archive_read_private.h" + +#if HAVE_ZSTD_H && HAVE_LIBZSTD + +struct private_data { + ZSTD_DStream *dstream; + unsigned char *out_block; + size_t out_block_size; + int64_t total_out; + char in_frame; /* True = in the middle of a zstd frame. */ + char eof; /* True = found end of compressed data. */ +}; + +/* Zstd Filter. */ +static ssize_t zstd_filter_read(struct archive_read_filter *, const void**); +static int zstd_filter_close(struct archive_read_filter *); +#endif + +/* + * Note that we can detect zstd compressed files even if we can't decompress + * them. (In fact, we like detecting them because we can give better error + * messages.) So the bid framework here gets compiled even if no zstd library + * is available. + */ +static int zstd_bidder_bid(struct archive_read_filter_bidder *, + struct archive_read_filter *); +static int zstd_bidder_init(struct archive_read_filter *); + +int +archive_read_support_filter_zstd(struct archive *_a) +{ + struct archive_read *a = (struct archive_read *)_a; + struct archive_read_filter_bidder *bidder; + + archive_check_magic(_a, ARCHIVE_READ_MAGIC, + ARCHIVE_STATE_NEW, "archive_read_support_filter_zstd"); + + if (__archive_read_get_bidder(a, &bidder) != ARCHIVE_OK) + return (ARCHIVE_FATAL); + + bidder->data = NULL; + bidder->name = "zstd"; + bidder->bid = zstd_bidder_bid; + bidder->init = zstd_bidder_init; + bidder->options = NULL; + bidder->free = NULL; +#if HAVE_ZSTD_H && HAVE_LIBZSTD + return (ARCHIVE_OK); +#else + archive_set_error(_a, ARCHIVE_ERRNO_MISC, + "Using external zstd program for zstd decompression"); + return (ARCHIVE_WARN); +#endif +} + +/* + * Test whether we can handle this data. + */ +static int +zstd_bidder_bid(struct archive_read_filter_bidder *self, + struct archive_read_filter *filter) +{ + const unsigned char *buffer; + ssize_t avail; + unsigned prefix; + + /* Zstd frame magic values */ + const unsigned zstd_magic = 0xFD2FB528U; + + (void) self; /* UNUSED */ + + buffer = __archive_read_filter_ahead(filter, 4, &avail); + if (buffer == NULL) + return (0); + + prefix = archive_le32dec(buffer); + if (prefix == zstd_magic) + return (32); + + return (0); +} + +#if !(HAVE_ZSTD_H && HAVE_LIBZSTD) + +/* + * If we don't have the library on this system, we can't do the + * decompression directly. We can, however, try to run "zstd -d" + * in case that's available. + */ +static int +zstd_bidder_init(struct archive_read_filter *self) +{ + int r; + + r = __archive_read_program(self, "zstd -d -qq"); + /* Note: We set the format here even if __archive_read_program() + * above fails. We do, after all, know what the format is + * even if we weren't able to read it. */ + self->code = ARCHIVE_FILTER_ZSTD; + self->name = "zstd"; + return (r); +} + +#else + +/* + * Initialize the filter object + */ +static int +zstd_bidder_init(struct archive_read_filter *self) +{ + struct private_data *state; + const size_t out_block_size = ZSTD_DStreamOutSize(); + void *out_block; + ZSTD_DStream *dstream; + + self->code = ARCHIVE_FILTER_ZSTD; + self->name = "zstd"; + + state = (struct private_data *)calloc(sizeof(*state), 1); + out_block = (unsigned char *)malloc(out_block_size); + dstream = ZSTD_createDStream(); + + if (state == NULL || out_block == NULL || dstream == NULL) { + free(out_block); + free(state); + ZSTD_freeDStream(dstream); /* supports free on NULL */ + archive_set_error(&self->archive->archive, ENOMEM, + "Can't allocate data for zstd decompression"); + return (ARCHIVE_FATAL); + } + + self->data = state; + + state->out_block_size = out_block_size; + state->out_block = out_block; + state->dstream = dstream; + self->read = zstd_filter_read; + self->skip = NULL; /* not supported */ + self->close = zstd_filter_close; + + state->eof = 0; + state->in_frame = 0; + + return (ARCHIVE_OK); +} + +static ssize_t +zstd_filter_read(struct archive_read_filter *self, const void **p) +{ + struct private_data *state; + size_t decompressed; + ssize_t avail_in; + ZSTD_outBuffer out; + ZSTD_inBuffer in; + + state = (struct private_data *)self->data; + + out = (ZSTD_outBuffer) { state->out_block, state->out_block_size, 0 }; + + /* Try to fill the output buffer. */ + while (out.pos < out.size && !state->eof) { + if (!state->in_frame) { + const size_t ret = ZSTD_initDStream(state->dstream); + if (ZSTD_isError(ret)) { + archive_set_error(&self->archive->archive, + ARCHIVE_ERRNO_MISC, + "Error initializing zstd decompressor: %s", + ZSTD_getErrorName(ret)); + return (ARCHIVE_FATAL); + } + } + in.src = __archive_read_filter_ahead(self->upstream, 1, + &avail_in); + if (avail_in < 0) { + return avail_in; + } + if (in.src == NULL && avail_in == 0) { + if (!state->in_frame) { + /* end of stream */ + state->eof = 1; + break; + } else { + archive_set_error(&self->archive->archive, + ARCHIVE_ERRNO_MISC, + "Truncated zstd input"); + return (ARCHIVE_FATAL); + } + } + in.size = avail_in; + in.pos = 0; + + { + const size_t ret = + ZSTD_decompressStream(state->dstream, &out, &in); + + if (ZSTD_isError(ret)) { + archive_set_error(&self->archive->archive, + ARCHIVE_ERRNO_MISC, + "Zstd decompression failed: %s", + ZSTD_getErrorName(ret)); + return (ARCHIVE_FATAL); + } + + /* Decompressor made some progress */ + __archive_read_filter_consume(self->upstream, in.pos); + + /* ret guaranteed to be > 0 if frame isn't done yet */ + state->in_frame = (ret != 0); + } + } + + decompressed = out.pos; + state->total_out += decompressed; + if (decompressed == 0) + *p = NULL; + else + *p = state->out_block; + return (decompressed); +} + +/* + * Clean up the decompressor. + */ +static int +zstd_filter_close(struct archive_read_filter *self) +{ + struct private_data *state; + + state = (struct private_data *)self->data; + + ZSTD_freeDStream(state->dstream); + free(state->out_block); + free(state); + + return (ARCHIVE_OK); +} + +#endif /* HAVE_ZLIB_H && HAVE_LIBZSTD */ Property changes on: stable/10/contrib/libarchive/libarchive/archive_read_support_filter_zstd.c ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: stable/10/contrib/libarchive/libarchive/archive_read_support_format_cpio.c =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_read_support_format_cpio.c (revision 324417) +++ stable/10/contrib/libarchive/libarchive/archive_read_support_format_cpio.c (revision 324418) @@ -1,1080 +1,1087 @@ /*- * Copyright (c) 2003-2007 Tim Kientzle * Copyright (c) 2010-2012 Michihiro NAKAJIMA * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "archive_platform.h" __FBSDID("$FreeBSD$"); #ifdef HAVE_ERRNO_H #include #endif /* #include */ /* See archive_platform.h */ #ifdef HAVE_STDLIB_H #include #endif #ifdef HAVE_STRING_H #include #endif #include "archive.h" #include "archive_entry.h" #include "archive_entry_locale.h" #include "archive_private.h" #include "archive_read_private.h" #define bin_magic_offset 0 #define bin_magic_size 2 #define bin_dev_offset 2 #define bin_dev_size 2 #define bin_ino_offset 4 #define bin_ino_size 2 #define bin_mode_offset 6 #define bin_mode_size 2 #define bin_uid_offset 8 #define bin_uid_size 2 #define bin_gid_offset 10 #define bin_gid_size 2 #define bin_nlink_offset 12 #define bin_nlink_size 2 #define bin_rdev_offset 14 #define bin_rdev_size 2 #define bin_mtime_offset 16 #define bin_mtime_size 4 #define bin_namesize_offset 20 #define bin_namesize_size 2 #define bin_filesize_offset 22 #define bin_filesize_size 4 #define bin_header_size 26 #define odc_magic_offset 0 #define odc_magic_size 6 #define odc_dev_offset 6 #define odc_dev_size 6 #define odc_ino_offset 12 #define odc_ino_size 6 #define odc_mode_offset 18 #define odc_mode_size 6 #define odc_uid_offset 24 #define odc_uid_size 6 #define odc_gid_offset 30 #define odc_gid_size 6 #define odc_nlink_offset 36 #define odc_nlink_size 6 #define odc_rdev_offset 42 #define odc_rdev_size 6 #define odc_mtime_offset 48 #define odc_mtime_size 11 #define odc_namesize_offset 59 #define odc_namesize_size 6 #define odc_filesize_offset 65 #define odc_filesize_size 11 #define odc_header_size 76 #define newc_magic_offset 0 #define newc_magic_size 6 #define newc_ino_offset 6 #define newc_ino_size 8 #define newc_mode_offset 14 #define newc_mode_size 8 #define newc_uid_offset 22 #define newc_uid_size 8 #define newc_gid_offset 30 #define newc_gid_size 8 #define newc_nlink_offset 38 #define newc_nlink_size 8 #define newc_mtime_offset 46 #define newc_mtime_size 8 #define newc_filesize_offset 54 #define newc_filesize_size 8 #define newc_devmajor_offset 62 #define newc_devmajor_size 8 #define newc_devminor_offset 70 #define newc_devminor_size 8 #define newc_rdevmajor_offset 78 #define newc_rdevmajor_size 8 #define newc_rdevminor_offset 86 #define newc_rdevminor_size 8 #define newc_namesize_offset 94 #define newc_namesize_size 8 #define newc_checksum_offset 102 #define newc_checksum_size 8 #define newc_header_size 110 /* * An afio large ASCII header, which they named itself. * afio utility uses this header, if a file size is larger than 2G bytes * or inode/uid/gid is bigger than 65535(0xFFFF) or mtime is bigger than * 0x7fffffff, which we cannot record to odc header because of its limit. * If not, uses odc header. */ #define afiol_magic_offset 0 #define afiol_magic_size 6 #define afiol_dev_offset 6 #define afiol_dev_size 8 /* hex */ #define afiol_ino_offset 14 #define afiol_ino_size 16 /* hex */ #define afiol_ino_m_offset 30 /* 'm' */ #define afiol_mode_offset 31 #define afiol_mode_size 6 /* oct */ #define afiol_uid_offset 37 #define afiol_uid_size 8 /* hex */ #define afiol_gid_offset 45 #define afiol_gid_size 8 /* hex */ #define afiol_nlink_offset 53 #define afiol_nlink_size 8 /* hex */ #define afiol_rdev_offset 61 #define afiol_rdev_size 8 /* hex */ #define afiol_mtime_offset 69 #define afiol_mtime_size 16 /* hex */ #define afiol_mtime_n_offset 85 /* 'n' */ #define afiol_namesize_offset 86 #define afiol_namesize_size 4 /* hex */ #define afiol_flag_offset 90 #define afiol_flag_size 4 /* hex */ #define afiol_xsize_offset 94 #define afiol_xsize_size 4 /* hex */ #define afiol_xsize_s_offset 98 /* 's' */ #define afiol_filesize_offset 99 #define afiol_filesize_size 16 /* hex */ #define afiol_filesize_c_offset 115 /* ':' */ #define afiol_header_size 116 struct links_entry { struct links_entry *next; struct links_entry *previous; unsigned int links; dev_t dev; int64_t ino; char *name; }; #define CPIO_MAGIC 0x13141516 struct cpio { int magic; int (*read_header)(struct archive_read *, struct cpio *, struct archive_entry *, size_t *, size_t *); struct links_entry *links_head; int64_t entry_bytes_remaining; int64_t entry_bytes_unconsumed; int64_t entry_offset; int64_t entry_padding; struct archive_string_conv *opt_sconv; struct archive_string_conv *sconv_default; int init_default_conversion; }; static int64_t atol16(const char *, unsigned); static int64_t atol8(const char *, unsigned); static int archive_read_format_cpio_bid(struct archive_read *, int); static int archive_read_format_cpio_options(struct archive_read *, const char *, const char *); static int archive_read_format_cpio_cleanup(struct archive_read *); static int archive_read_format_cpio_read_data(struct archive_read *, const void **, size_t *, int64_t *); static int archive_read_format_cpio_read_header(struct archive_read *, struct archive_entry *); static int archive_read_format_cpio_skip(struct archive_read *); static int64_t be4(const unsigned char *); static int find_odc_header(struct archive_read *); static int find_newc_header(struct archive_read *); static int header_bin_be(struct archive_read *, struct cpio *, struct archive_entry *, size_t *, size_t *); static int header_bin_le(struct archive_read *, struct cpio *, struct archive_entry *, size_t *, size_t *); static int header_newc(struct archive_read *, struct cpio *, struct archive_entry *, size_t *, size_t *); static int header_odc(struct archive_read *, struct cpio *, struct archive_entry *, size_t *, size_t *); static int header_afiol(struct archive_read *, struct cpio *, struct archive_entry *, size_t *, size_t *); static int is_octal(const char *, size_t); static int is_hex(const char *, size_t); static int64_t le4(const unsigned char *); static int record_hardlink(struct archive_read *a, struct cpio *cpio, struct archive_entry *entry); int archive_read_support_format_cpio(struct archive *_a) { struct archive_read *a = (struct archive_read *)_a; struct cpio *cpio; int r; archive_check_magic(_a, ARCHIVE_READ_MAGIC, ARCHIVE_STATE_NEW, "archive_read_support_format_cpio"); cpio = (struct cpio *)calloc(1, sizeof(*cpio)); if (cpio == NULL) { archive_set_error(&a->archive, ENOMEM, "Can't allocate cpio data"); return (ARCHIVE_FATAL); } cpio->magic = CPIO_MAGIC; r = __archive_read_register_format(a, cpio, "cpio", archive_read_format_cpio_bid, archive_read_format_cpio_options, archive_read_format_cpio_read_header, archive_read_format_cpio_read_data, archive_read_format_cpio_skip, NULL, archive_read_format_cpio_cleanup, NULL, NULL); if (r != ARCHIVE_OK) free(cpio); return (ARCHIVE_OK); } static int archive_read_format_cpio_bid(struct archive_read *a, int best_bid) { const unsigned char *p; struct cpio *cpio; int bid; (void)best_bid; /* UNUSED */ cpio = (struct cpio *)(a->format->data); if ((p = __archive_read_ahead(a, 6, NULL)) == NULL) return (-1); bid = 0; if (memcmp(p, "070707", 6) == 0) { /* ASCII cpio archive (odc, POSIX.1) */ cpio->read_header = header_odc; bid += 48; /* * XXX TODO: More verification; Could check that only octal * digits appear in appropriate header locations. XXX */ } else if (memcmp(p, "070727", 6) == 0) { /* afio large ASCII cpio archive */ cpio->read_header = header_odc; bid += 48; /* * XXX TODO: More verification; Could check that almost hex * digits appear in appropriate header locations. XXX */ } else if (memcmp(p, "070701", 6) == 0) { /* ASCII cpio archive (SVR4 without CRC) */ cpio->read_header = header_newc; bid += 48; /* * XXX TODO: More verification; Could check that only hex * digits appear in appropriate header locations. XXX */ } else if (memcmp(p, "070702", 6) == 0) { /* ASCII cpio archive (SVR4 with CRC) */ /* XXX TODO: Flag that we should check the CRC. XXX */ cpio->read_header = header_newc; bid += 48; /* * XXX TODO: More verification; Could check that only hex * digits appear in appropriate header locations. XXX */ } else if (p[0] * 256 + p[1] == 070707) { /* big-endian binary cpio archives */ cpio->read_header = header_bin_be; bid += 16; /* Is more verification possible here? */ } else if (p[0] + p[1] * 256 == 070707) { /* little-endian binary cpio archives */ cpio->read_header = header_bin_le; bid += 16; /* Is more verification possible here? */ } else return (ARCHIVE_WARN); return (bid); } static int archive_read_format_cpio_options(struct archive_read *a, const char *key, const char *val) { struct cpio *cpio; int ret = ARCHIVE_FAILED; cpio = (struct cpio *)(a->format->data); if (strcmp(key, "compat-2x") == 0) { /* Handle filenames as libarchive 2.x */ cpio->init_default_conversion = (val != NULL)?1:0; return (ARCHIVE_OK); } else if (strcmp(key, "hdrcharset") == 0) { if (val == NULL || val[0] == 0) archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "cpio: hdrcharset option needs a character-set name"); else { cpio->opt_sconv = archive_string_conversion_from_charset( &a->archive, val, 0); if (cpio->opt_sconv != NULL) ret = ARCHIVE_OK; else ret = ARCHIVE_FATAL; } return (ret); } /* Note: The "warn" return is just to inform the options * supervisor that we didn't handle it. It will generate * a suitable error if no one used this option. */ return (ARCHIVE_WARN); } static int archive_read_format_cpio_read_header(struct archive_read *a, struct archive_entry *entry) { struct cpio *cpio; const void *h, *hl; struct archive_string_conv *sconv; size_t namelength; size_t name_pad; int r; cpio = (struct cpio *)(a->format->data); sconv = cpio->opt_sconv; if (sconv == NULL) { if (!cpio->init_default_conversion) { cpio->sconv_default = archive_string_default_conversion_for_read( &(a->archive)); cpio->init_default_conversion = 1; } sconv = cpio->sconv_default; } r = (cpio->read_header(a, cpio, entry, &namelength, &name_pad)); if (r < ARCHIVE_WARN) return (r); /* Read name from buffer. */ h = __archive_read_ahead(a, namelength + name_pad, NULL); if (h == NULL) return (ARCHIVE_FATAL); if (archive_entry_copy_pathname_l(entry, (const char *)h, namelength, sconv) != 0) { if (errno == ENOMEM) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory for Pathname"); return (ARCHIVE_FATAL); } archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Pathname can't be converted from %s to current locale.", archive_string_conversion_charset_name(sconv)); r = ARCHIVE_WARN; } cpio->entry_offset = 0; __archive_read_consume(a, namelength + name_pad); /* If this is a symlink, read the link contents. */ if (archive_entry_filetype(entry) == AE_IFLNK) { if (cpio->entry_bytes_remaining > 1024 * 1024) { archive_set_error(&a->archive, ENOMEM, "Rejecting malformed cpio archive: symlink contents exceed 1 megabyte"); return (ARCHIVE_FATAL); } hl = __archive_read_ahead(a, (size_t)cpio->entry_bytes_remaining, NULL); if (hl == NULL) return (ARCHIVE_FATAL); if (archive_entry_copy_symlink_l(entry, (const char *)hl, (size_t)cpio->entry_bytes_remaining, sconv) != 0) { if (errno == ENOMEM) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory for Linkname"); return (ARCHIVE_FATAL); } archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Linkname can't be converted from %s to " "current locale.", archive_string_conversion_charset_name(sconv)); r = ARCHIVE_WARN; } __archive_read_consume(a, cpio->entry_bytes_remaining); cpio->entry_bytes_remaining = 0; } /* XXX TODO: If the full mode is 0160200, then this is a Solaris * ACL description for the following entry. Read this body * and parse it as a Solaris-style ACL, then read the next * header. XXX */ /* Compare name to "TRAILER!!!" to test for end-of-archive. */ if (namelength == 11 && strncmp((const char *)h, "TRAILER!!!", 11) == 0) { /* TODO: Store file location of start of block. */ archive_clear_error(&a->archive); return (ARCHIVE_EOF); } /* Detect and record hardlinks to previously-extracted entries. */ if (record_hardlink(a, cpio, entry) != ARCHIVE_OK) { return (ARCHIVE_FATAL); } return (r); } static int archive_read_format_cpio_read_data(struct archive_read *a, const void **buff, size_t *size, int64_t *offset) { ssize_t bytes_read; struct cpio *cpio; cpio = (struct cpio *)(a->format->data); if (cpio->entry_bytes_unconsumed) { __archive_read_consume(a, cpio->entry_bytes_unconsumed); cpio->entry_bytes_unconsumed = 0; } if (cpio->entry_bytes_remaining > 0) { *buff = __archive_read_ahead(a, 1, &bytes_read); if (bytes_read <= 0) return (ARCHIVE_FATAL); if (bytes_read > cpio->entry_bytes_remaining) bytes_read = (ssize_t)cpio->entry_bytes_remaining; *size = bytes_read; cpio->entry_bytes_unconsumed = bytes_read; *offset = cpio->entry_offset; cpio->entry_offset += bytes_read; cpio->entry_bytes_remaining -= bytes_read; return (ARCHIVE_OK); } else { if (cpio->entry_padding != __archive_read_consume(a, cpio->entry_padding)) { return (ARCHIVE_FATAL); } cpio->entry_padding = 0; *buff = NULL; *size = 0; *offset = cpio->entry_offset; return (ARCHIVE_EOF); } } static int archive_read_format_cpio_skip(struct archive_read *a) { struct cpio *cpio = (struct cpio *)(a->format->data); int64_t to_skip = cpio->entry_bytes_remaining + cpio->entry_padding + cpio->entry_bytes_unconsumed; if (to_skip != __archive_read_consume(a, to_skip)) { return (ARCHIVE_FATAL); } cpio->entry_bytes_remaining = 0; cpio->entry_padding = 0; cpio->entry_bytes_unconsumed = 0; return (ARCHIVE_OK); } /* * Skip forward to the next cpio newc header by searching for the * 07070[12] string. This should be generalized and merged with * find_odc_header below. */ static int is_hex(const char *p, size_t len) { while (len-- > 0) { if ((*p >= '0' && *p <= '9') || (*p >= 'a' && *p <= 'f') || (*p >= 'A' && *p <= 'F')) ++p; else return (0); } return (1); } static int find_newc_header(struct archive_read *a) { const void *h; const char *p, *q; size_t skip, skipped = 0; ssize_t bytes; for (;;) { h = __archive_read_ahead(a, newc_header_size, &bytes); if (h == NULL) return (ARCHIVE_FATAL); p = h; q = p + bytes; /* Try the typical case first, then go into the slow search.*/ if (memcmp("07070", p, 5) == 0 && (p[5] == '1' || p[5] == '2') && is_hex(p, newc_header_size)) return (ARCHIVE_OK); /* * Scan ahead until we find something that looks * like a newc header. */ while (p + newc_header_size <= q) { switch (p[5]) { case '1': case '2': if (memcmp("07070", p, 5) == 0 && is_hex(p, newc_header_size)) { skip = p - (const char *)h; __archive_read_consume(a, skip); skipped += skip; if (skipped > 0) { archive_set_error(&a->archive, 0, "Skipped %d bytes before " "finding valid header", (int)skipped); return (ARCHIVE_WARN); } return (ARCHIVE_OK); } p += 2; break; case '0': p++; break; default: p += 6; break; } } skip = p - (const char *)h; __archive_read_consume(a, skip); skipped += skip; } } static int header_newc(struct archive_read *a, struct cpio *cpio, struct archive_entry *entry, size_t *namelength, size_t *name_pad) { const void *h; const char *header; int r; r = find_newc_header(a); if (r < ARCHIVE_WARN) return (r); /* Read fixed-size portion of header. */ h = __archive_read_ahead(a, newc_header_size, NULL); if (h == NULL) return (ARCHIVE_FATAL); /* Parse out hex fields. */ header = (const char *)h; if (memcmp(header + newc_magic_offset, "070701", 6) == 0) { a->archive.archive_format = ARCHIVE_FORMAT_CPIO_SVR4_NOCRC; a->archive.archive_format_name = "ASCII cpio (SVR4 with no CRC)"; } else if (memcmp(header + newc_magic_offset, "070702", 6) == 0) { a->archive.archive_format = ARCHIVE_FORMAT_CPIO_SVR4_CRC; a->archive.archive_format_name = "ASCII cpio (SVR4 with CRC)"; } else { /* TODO: Abort here? */ } archive_entry_set_devmajor(entry, (dev_t)atol16(header + newc_devmajor_offset, newc_devmajor_size)); archive_entry_set_devminor(entry, (dev_t)atol16(header + newc_devminor_offset, newc_devminor_size)); archive_entry_set_ino(entry, atol16(header + newc_ino_offset, newc_ino_size)); archive_entry_set_mode(entry, (mode_t)atol16(header + newc_mode_offset, newc_mode_size)); archive_entry_set_uid(entry, atol16(header + newc_uid_offset, newc_uid_size)); archive_entry_set_gid(entry, atol16(header + newc_gid_offset, newc_gid_size)); archive_entry_set_nlink(entry, (unsigned int)atol16(header + newc_nlink_offset, newc_nlink_size)); archive_entry_set_rdevmajor(entry, (dev_t)atol16(header + newc_rdevmajor_offset, newc_rdevmajor_size)); archive_entry_set_rdevminor(entry, (dev_t)atol16(header + newc_rdevminor_offset, newc_rdevminor_size)); archive_entry_set_mtime(entry, atol16(header + newc_mtime_offset, newc_mtime_size), 0); *namelength = (size_t)atol16(header + newc_namesize_offset, newc_namesize_size); /* Pad name to 2 more than a multiple of 4. */ *name_pad = (2 - *namelength) & 3; + /* Make sure that the padded name length fits into size_t. */ + if (*name_pad > SIZE_MAX - *namelength) { + archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, + "cpio archive has invalid namelength"); + return (ARCHIVE_FATAL); + } + /* * Note: entry_bytes_remaining is at least 64 bits and * therefore guaranteed to be big enough for a 33-bit file * size. */ cpio->entry_bytes_remaining = atol16(header + newc_filesize_offset, newc_filesize_size); archive_entry_set_size(entry, cpio->entry_bytes_remaining); /* Pad file contents to a multiple of 4. */ cpio->entry_padding = 3 & -cpio->entry_bytes_remaining; __archive_read_consume(a, newc_header_size); return (r); } /* * Skip forward to the next cpio odc header by searching for the * 070707 string. This is a hand-optimized search that could * probably be easily generalized to handle all character-based * cpio variants. */ static int is_octal(const char *p, size_t len) { while (len-- > 0) { if (*p < '0' || *p > '7') return (0); ++p; } return (1); } static int is_afio_large(const char *h, size_t len) { if (len < afiol_header_size) return (0); if (h[afiol_ino_m_offset] != 'm' || h[afiol_mtime_n_offset] != 'n' || h[afiol_xsize_s_offset] != 's' || h[afiol_filesize_c_offset] != ':') return (0); if (!is_hex(h + afiol_dev_offset, afiol_ino_m_offset - afiol_dev_offset)) return (0); if (!is_hex(h + afiol_mode_offset, afiol_mtime_n_offset - afiol_mode_offset)) return (0); if (!is_hex(h + afiol_namesize_offset, afiol_xsize_s_offset - afiol_namesize_offset)) return (0); if (!is_hex(h + afiol_filesize_offset, afiol_filesize_size)) return (0); return (1); } static int find_odc_header(struct archive_read *a) { const void *h; const char *p, *q; size_t skip, skipped = 0; ssize_t bytes; for (;;) { h = __archive_read_ahead(a, odc_header_size, &bytes); if (h == NULL) return (ARCHIVE_FATAL); p = h; q = p + bytes; /* Try the typical case first, then go into the slow search.*/ if (memcmp("070707", p, 6) == 0 && is_octal(p, odc_header_size)) return (ARCHIVE_OK); if (memcmp("070727", p, 6) == 0 && is_afio_large(p, bytes)) { a->archive.archive_format = ARCHIVE_FORMAT_CPIO_AFIO_LARGE; return (ARCHIVE_OK); } /* * Scan ahead until we find something that looks * like an odc header. */ while (p + odc_header_size <= q) { switch (p[5]) { case '7': if ((memcmp("070707", p, 6) == 0 && is_octal(p, odc_header_size)) || (memcmp("070727", p, 6) == 0 && is_afio_large(p, q - p))) { skip = p - (const char *)h; __archive_read_consume(a, skip); skipped += skip; if (p[4] == '2') a->archive.archive_format = ARCHIVE_FORMAT_CPIO_AFIO_LARGE; if (skipped > 0) { archive_set_error(&a->archive, 0, "Skipped %d bytes before " "finding valid header", (int)skipped); return (ARCHIVE_WARN); } return (ARCHIVE_OK); } p += 2; break; case '0': p++; break; default: p += 6; break; } } skip = p - (const char *)h; __archive_read_consume(a, skip); skipped += skip; } } static int header_odc(struct archive_read *a, struct cpio *cpio, struct archive_entry *entry, size_t *namelength, size_t *name_pad) { const void *h; int r; const char *header; a->archive.archive_format = ARCHIVE_FORMAT_CPIO_POSIX; a->archive.archive_format_name = "POSIX octet-oriented cpio"; /* Find the start of the next header. */ r = find_odc_header(a); if (r < ARCHIVE_WARN) return (r); if (a->archive.archive_format == ARCHIVE_FORMAT_CPIO_AFIO_LARGE) { int r2 = (header_afiol(a, cpio, entry, namelength, name_pad)); if (r2 == ARCHIVE_OK) return (r); else return (r2); } /* Read fixed-size portion of header. */ h = __archive_read_ahead(a, odc_header_size, NULL); if (h == NULL) return (ARCHIVE_FATAL); /* Parse out octal fields. */ header = (const char *)h; archive_entry_set_dev(entry, (dev_t)atol8(header + odc_dev_offset, odc_dev_size)); archive_entry_set_ino(entry, atol8(header + odc_ino_offset, odc_ino_size)); archive_entry_set_mode(entry, (mode_t)atol8(header + odc_mode_offset, odc_mode_size)); archive_entry_set_uid(entry, atol8(header + odc_uid_offset, odc_uid_size)); archive_entry_set_gid(entry, atol8(header + odc_gid_offset, odc_gid_size)); archive_entry_set_nlink(entry, (unsigned int)atol8(header + odc_nlink_offset, odc_nlink_size)); archive_entry_set_rdev(entry, (dev_t)atol8(header + odc_rdev_offset, odc_rdev_size)); archive_entry_set_mtime(entry, atol8(header + odc_mtime_offset, odc_mtime_size), 0); *namelength = (size_t)atol8(header + odc_namesize_offset, odc_namesize_size); *name_pad = 0; /* No padding of filename. */ /* * Note: entry_bytes_remaining is at least 64 bits and * therefore guaranteed to be big enough for a 33-bit file * size. */ cpio->entry_bytes_remaining = atol8(header + odc_filesize_offset, odc_filesize_size); archive_entry_set_size(entry, cpio->entry_bytes_remaining); cpio->entry_padding = 0; __archive_read_consume(a, odc_header_size); return (r); } /* * NOTE: if a filename suffix is ".z", it is the file gziped by afio. * it would be nice that we can show uncompressed file size and we can * uncompressed file contents automatically, unfortunately we have nothing * to get a uncompressed file size while reading each header. It means * we also cannot uncompress file contents under our framework. */ static int header_afiol(struct archive_read *a, struct cpio *cpio, struct archive_entry *entry, size_t *namelength, size_t *name_pad) { const void *h; const char *header; a->archive.archive_format = ARCHIVE_FORMAT_CPIO_AFIO_LARGE; a->archive.archive_format_name = "afio large ASCII"; /* Read fixed-size portion of header. */ h = __archive_read_ahead(a, afiol_header_size, NULL); if (h == NULL) return (ARCHIVE_FATAL); /* Parse out octal fields. */ header = (const char *)h; archive_entry_set_dev(entry, (dev_t)atol16(header + afiol_dev_offset, afiol_dev_size)); archive_entry_set_ino(entry, atol16(header + afiol_ino_offset, afiol_ino_size)); archive_entry_set_mode(entry, (mode_t)atol8(header + afiol_mode_offset, afiol_mode_size)); archive_entry_set_uid(entry, atol16(header + afiol_uid_offset, afiol_uid_size)); archive_entry_set_gid(entry, atol16(header + afiol_gid_offset, afiol_gid_size)); archive_entry_set_nlink(entry, (unsigned int)atol16(header + afiol_nlink_offset, afiol_nlink_size)); archive_entry_set_rdev(entry, (dev_t)atol16(header + afiol_rdev_offset, afiol_rdev_size)); archive_entry_set_mtime(entry, atol16(header + afiol_mtime_offset, afiol_mtime_size), 0); *namelength = (size_t)atol16(header + afiol_namesize_offset, afiol_namesize_size); *name_pad = 0; /* No padding of filename. */ cpio->entry_bytes_remaining = atol16(header + afiol_filesize_offset, afiol_filesize_size); archive_entry_set_size(entry, cpio->entry_bytes_remaining); cpio->entry_padding = 0; __archive_read_consume(a, afiol_header_size); return (ARCHIVE_OK); } static int header_bin_le(struct archive_read *a, struct cpio *cpio, struct archive_entry *entry, size_t *namelength, size_t *name_pad) { const void *h; const unsigned char *header; a->archive.archive_format = ARCHIVE_FORMAT_CPIO_BIN_LE; a->archive.archive_format_name = "cpio (little-endian binary)"; /* Read fixed-size portion of header. */ h = __archive_read_ahead(a, bin_header_size, NULL); if (h == NULL) { archive_set_error(&a->archive, 0, "End of file trying to read next cpio header"); return (ARCHIVE_FATAL); } /* Parse out binary fields. */ header = (const unsigned char *)h; archive_entry_set_dev(entry, header[bin_dev_offset] + header[bin_dev_offset + 1] * 256); archive_entry_set_ino(entry, header[bin_ino_offset] + header[bin_ino_offset + 1] * 256); archive_entry_set_mode(entry, header[bin_mode_offset] + header[bin_mode_offset + 1] * 256); archive_entry_set_uid(entry, header[bin_uid_offset] + header[bin_uid_offset + 1] * 256); archive_entry_set_gid(entry, header[bin_gid_offset] + header[bin_gid_offset + 1] * 256); archive_entry_set_nlink(entry, header[bin_nlink_offset] + header[bin_nlink_offset + 1] * 256); archive_entry_set_rdev(entry, header[bin_rdev_offset] + header[bin_rdev_offset + 1] * 256); archive_entry_set_mtime(entry, le4(header + bin_mtime_offset), 0); *namelength = header[bin_namesize_offset] + header[bin_namesize_offset + 1] * 256; *name_pad = *namelength & 1; /* Pad to even. */ cpio->entry_bytes_remaining = le4(header + bin_filesize_offset); archive_entry_set_size(entry, cpio->entry_bytes_remaining); cpio->entry_padding = cpio->entry_bytes_remaining & 1; /* Pad to even. */ __archive_read_consume(a, bin_header_size); return (ARCHIVE_OK); } static int header_bin_be(struct archive_read *a, struct cpio *cpio, struct archive_entry *entry, size_t *namelength, size_t *name_pad) { const void *h; const unsigned char *header; a->archive.archive_format = ARCHIVE_FORMAT_CPIO_BIN_BE; a->archive.archive_format_name = "cpio (big-endian binary)"; /* Read fixed-size portion of header. */ h = __archive_read_ahead(a, bin_header_size, NULL); if (h == NULL) { archive_set_error(&a->archive, 0, "End of file trying to read next cpio header"); return (ARCHIVE_FATAL); } /* Parse out binary fields. */ header = (const unsigned char *)h; archive_entry_set_dev(entry, header[bin_dev_offset] * 256 + header[bin_dev_offset + 1]); archive_entry_set_ino(entry, header[bin_ino_offset] * 256 + header[bin_ino_offset + 1]); archive_entry_set_mode(entry, header[bin_mode_offset] * 256 + header[bin_mode_offset + 1]); archive_entry_set_uid(entry, header[bin_uid_offset] * 256 + header[bin_uid_offset + 1]); archive_entry_set_gid(entry, header[bin_gid_offset] * 256 + header[bin_gid_offset + 1]); archive_entry_set_nlink(entry, header[bin_nlink_offset] * 256 + header[bin_nlink_offset + 1]); archive_entry_set_rdev(entry, header[bin_rdev_offset] * 256 + header[bin_rdev_offset + 1]); archive_entry_set_mtime(entry, be4(header + bin_mtime_offset), 0); *namelength = header[bin_namesize_offset] * 256 + header[bin_namesize_offset + 1]; *name_pad = *namelength & 1; /* Pad to even. */ cpio->entry_bytes_remaining = be4(header + bin_filesize_offset); archive_entry_set_size(entry, cpio->entry_bytes_remaining); cpio->entry_padding = cpio->entry_bytes_remaining & 1; /* Pad to even. */ __archive_read_consume(a, bin_header_size); return (ARCHIVE_OK); } static int archive_read_format_cpio_cleanup(struct archive_read *a) { struct cpio *cpio; cpio = (struct cpio *)(a->format->data); /* Free inode->name map */ while (cpio->links_head != NULL) { struct links_entry *lp = cpio->links_head->next; if (cpio->links_head->name) free(cpio->links_head->name); free(cpio->links_head); cpio->links_head = lp; } free(cpio); (a->format->data) = NULL; return (ARCHIVE_OK); } static int64_t le4(const unsigned char *p) { return ((p[0] << 16) + (((int64_t)p[1]) << 24) + (p[2] << 0) + (p[3] << 8)); } static int64_t be4(const unsigned char *p) { return ((((int64_t)p[0]) << 24) + (p[1] << 16) + (p[2] << 8) + (p[3])); } /* * Note that this implementation does not (and should not!) obey * locale settings; you cannot simply substitute strtol here, since * it does obey locale. */ static int64_t atol8(const char *p, unsigned char_cnt) { int64_t l; int digit; l = 0; while (char_cnt-- > 0) { if (*p >= '0' && *p <= '7') digit = *p - '0'; else return (l); p++; l <<= 3; l |= digit; } return (l); } static int64_t atol16(const char *p, unsigned char_cnt) { int64_t l; int digit; l = 0; while (char_cnt-- > 0) { if (*p >= 'a' && *p <= 'f') digit = *p - 'a' + 10; else if (*p >= 'A' && *p <= 'F') digit = *p - 'A' + 10; else if (*p >= '0' && *p <= '9') digit = *p - '0'; else return (l); p++; l <<= 4; l |= digit; } return (l); } static int record_hardlink(struct archive_read *a, struct cpio *cpio, struct archive_entry *entry) { struct links_entry *le; dev_t dev; int64_t ino; if (archive_entry_nlink(entry) <= 1) return (ARCHIVE_OK); dev = archive_entry_dev(entry); ino = archive_entry_ino64(entry); /* * First look in the list of multiply-linked files. If we've * already dumped it, convert this entry to a hard link entry. */ for (le = cpio->links_head; le; le = le->next) { if (le->dev == dev && le->ino == ino) { archive_entry_copy_hardlink(entry, le->name); if (--le->links <= 0) { if (le->previous != NULL) le->previous->next = le->next; if (le->next != NULL) le->next->previous = le->previous; if (cpio->links_head == le) cpio->links_head = le->next; free(le->name); free(le); } return (ARCHIVE_OK); } } le = (struct links_entry *)malloc(sizeof(struct links_entry)); if (le == NULL) { archive_set_error(&a->archive, ENOMEM, "Out of memory adding file to list"); return (ARCHIVE_FATAL); } if (cpio->links_head != NULL) cpio->links_head->previous = le; le->next = cpio->links_head; le->previous = NULL; cpio->links_head = le; le->dev = dev; le->ino = ino; le->links = archive_entry_nlink(entry) - 1; le->name = strdup(archive_entry_pathname(entry)); if (le->name == NULL) { archive_set_error(&a->archive, ENOMEM, "Out of memory adding file to list"); return (ARCHIVE_FATAL); } return (ARCHIVE_OK); } Index: stable/10/contrib/libarchive/libarchive/archive_read_support_format_mtree.c =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_read_support_format_mtree.c (revision 324417) +++ stable/10/contrib/libarchive/libarchive/archive_read_support_format_mtree.c (revision 324418) @@ -1,1980 +1,1990 @@ /*- * Copyright (c) 2003-2007 Tim Kientzle * Copyright (c) 2008 Joerg Sonnenberger * Copyright (c) 2011-2012 Michihiro NAKAJIMA * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "archive_platform.h" __FBSDID("$FreeBSD$"); #ifdef HAVE_SYS_STAT_H #include #endif #ifdef HAVE_ERRNO_H #include #endif #ifdef HAVE_FCNTL_H #include #endif #include /* #include */ /* See archive_platform.h */ #ifdef HAVE_STDLIB_H #include #endif #ifdef HAVE_STRING_H #include #endif #include "archive.h" #include "archive_entry.h" #include "archive_private.h" #include "archive_read_private.h" #include "archive_string.h" #include "archive_pack_dev.h" #ifndef O_BINARY #define O_BINARY 0 #endif #ifndef O_CLOEXEC #define O_CLOEXEC 0 #endif #define MTREE_HAS_DEVICE 0x0001 #define MTREE_HAS_FFLAGS 0x0002 #define MTREE_HAS_GID 0x0004 #define MTREE_HAS_GNAME 0x0008 #define MTREE_HAS_MTIME 0x0010 #define MTREE_HAS_NLINK 0x0020 #define MTREE_HAS_PERM 0x0040 #define MTREE_HAS_SIZE 0x0080 #define MTREE_HAS_TYPE 0x0100 #define MTREE_HAS_UID 0x0200 #define MTREE_HAS_UNAME 0x0400 #define MTREE_HAS_OPTIONAL 0x0800 #define MTREE_HAS_NOCHANGE 0x1000 /* FreeBSD specific */ #define MTREE_HASHTABLE_SIZE 1024 +#define MAX_LINE_LEN (1024 * 1024) + struct mtree_option { struct mtree_option *next; char *value; }; struct mtree_entry { struct mtree_entry *next; struct mtree_option *options; char *name; char full; char used; unsigned int name_hash; struct mtree_entry *hashtable_next; }; struct mtree { struct archive_string line; size_t buffsize; char *buff; int64_t offset; int fd; int archive_format; const char *archive_format_name; struct mtree_entry *entries; struct mtree_entry *this_entry; struct mtree_entry *entry_hashtable[MTREE_HASHTABLE_SIZE]; struct archive_string current_dir; struct archive_string contents_name; struct archive_entry_linkresolver *resolver; int64_t cur_size; char checkfs; }; static int bid_keycmp(const char *, const char *, ssize_t); static int cleanup(struct archive_read *); static int detect_form(struct archive_read *, int *); static unsigned int hash(const char *); static int mtree_bid(struct archive_read *, int); static int parse_file(struct archive_read *, struct archive_entry *, struct mtree *, struct mtree_entry *, int *); static void parse_escapes(char *, struct mtree_entry *); static int parse_line(struct archive_read *, struct archive_entry *, struct mtree *, struct mtree_entry *, int *); static int parse_keyword(struct archive_read *, struct mtree *, struct archive_entry *, struct mtree_option *, int *); static int read_data(struct archive_read *a, const void **buff, size_t *size, int64_t *offset); static ssize_t readline(struct archive_read *, struct mtree *, char **, ssize_t); static int skip(struct archive_read *a); static int read_header(struct archive_read *, struct archive_entry *); static int64_t mtree_atol(char **, int base); /* * There's no standard for TIME_T_MAX/TIME_T_MIN. So we compute them * here. TODO: Move this to configure time, but be careful * about cross-compile environments. */ static int64_t get_time_t_max(void) { #if defined(TIME_T_MAX) return TIME_T_MAX; #else /* ISO C allows time_t to be a floating-point type, but POSIX requires an integer type. The following should work on any system that follows the POSIX conventions. */ if (((time_t)0) < ((time_t)-1)) { /* Time_t is unsigned */ return (~(time_t)0); } else { /* Time_t is signed. */ /* Assume it's the same as int64_t or int32_t */ if (sizeof(time_t) == sizeof(int64_t)) { return (time_t)INT64_MAX; } else { return (time_t)INT32_MAX; } } #endif } static int64_t get_time_t_min(void) { #if defined(TIME_T_MIN) return TIME_T_MIN; #else if (((time_t)0) < ((time_t)-1)) { /* Time_t is unsigned */ return (time_t)0; } else { /* Time_t is signed. */ if (sizeof(time_t) == sizeof(int64_t)) { return (time_t)INT64_MIN; } else { return (time_t)INT32_MIN; } } #endif } static int archive_read_format_mtree_options(struct archive_read *a, const char *key, const char *val) { struct mtree *mtree; mtree = (struct mtree *)(a->format->data); if (strcmp(key, "checkfs") == 0) { /* Allows to read information missing from the mtree from the file system */ if (val == NULL || val[0] == 0) { mtree->checkfs = 0; } else { mtree->checkfs = 1; } return (ARCHIVE_OK); } /* Note: The "warn" return is just to inform the options * supervisor that we didn't handle it. It will generate * a suitable error if no one used this option. */ return (ARCHIVE_WARN); } static void free_options(struct mtree_option *head) { struct mtree_option *next; for (; head != NULL; head = next) { next = head->next; free(head->value); free(head); } } int archive_read_support_format_mtree(struct archive *_a) { struct archive_read *a = (struct archive_read *)_a; struct mtree *mtree; int r; archive_check_magic(_a, ARCHIVE_READ_MAGIC, ARCHIVE_STATE_NEW, "archive_read_support_format_mtree"); mtree = (struct mtree *)calloc(1, sizeof(*mtree)); if (mtree == NULL) { archive_set_error(&a->archive, ENOMEM, "Can't allocate mtree data"); return (ARCHIVE_FATAL); } mtree->fd = -1; r = __archive_read_register_format(a, mtree, "mtree", mtree_bid, archive_read_format_mtree_options, read_header, read_data, skip, NULL, cleanup, NULL, NULL); if (r != ARCHIVE_OK) free(mtree); return (ARCHIVE_OK); } static int cleanup(struct archive_read *a) { struct mtree *mtree; struct mtree_entry *p, *q; mtree = (struct mtree *)(a->format->data); p = mtree->entries; while (p != NULL) { q = p->next; free(p->name); free_options(p->options); free(p); p = q; } archive_string_free(&mtree->line); archive_string_free(&mtree->current_dir); archive_string_free(&mtree->contents_name); archive_entry_linkresolver_free(mtree->resolver); free(mtree->buff); free(mtree); (a->format->data) = NULL; return (ARCHIVE_OK); } static ssize_t get_line_size(const char *b, ssize_t avail, ssize_t *nlsize) { ssize_t len; len = 0; while (len < avail) { switch (*b) { case '\0':/* Non-ascii character or control character. */ if (nlsize != NULL) *nlsize = 0; return (-1); case '\r': if (avail-len > 1 && b[1] == '\n') { if (nlsize != NULL) *nlsize = 2; return (len+2); } /* FALL THROUGH */ case '\n': if (nlsize != NULL) *nlsize = 1; return (len+1); default: b++; len++; break; } } if (nlsize != NULL) *nlsize = 0; return (avail); } /* * <---------------- ravail ---------------------> * <-- diff ------> <--- avail -----------------> * <---- len -----------> * | Previous lines | line being parsed nl extra | * ^ * b * */ static ssize_t next_line(struct archive_read *a, const char **b, ssize_t *avail, ssize_t *ravail, ssize_t *nl) { ssize_t len; int quit; quit = 0; if (*avail == 0) { *nl = 0; len = 0; } else len = get_line_size(*b, *avail, nl); /* * Read bytes more while it does not reach the end of line. */ while (*nl == 0 && len == *avail && !quit) { ssize_t diff = *ravail - *avail; size_t nbytes_req = (*ravail+1023) & ~1023U; ssize_t tested; + + /* + * Place an arbitrary limit on the line length. + * mtree is almost free-form input and without line length limits, + * it can consume a lot of memory. + */ + if (len >= MAX_LINE_LEN) + return (-1); /* Increase reading bytes if it is not enough to at least * new two lines. */ if (nbytes_req < (size_t)*ravail + 160) nbytes_req <<= 1; *b = __archive_read_ahead(a, nbytes_req, avail); if (*b == NULL) { if (*ravail >= *avail) return (0); /* Reading bytes reaches the end of file. */ *b = __archive_read_ahead(a, *avail, avail); quit = 1; } *ravail = *avail; *b += diff; *avail -= diff; tested = len;/* Skip some bytes we already determinated. */ len = get_line_size(*b + len, *avail - len, nl); if (len >= 0) len += tested; } return (len); } /* * Compare characters with a mtree keyword. * Returns the length of a mtree keyword if matched. * Returns 0 if not matched. */ static int bid_keycmp(const char *p, const char *key, ssize_t len) { int match_len = 0; while (len > 0 && *p && *key) { if (*p == *key) { --len; ++p; ++key; ++match_len; continue; } return (0);/* Not match */ } if (*key != '\0') return (0);/* Not match */ /* A following character should be specified characters */ if (p[0] == '=' || p[0] == ' ' || p[0] == '\t' || p[0] == '\n' || p[0] == '\r' || (p[0] == '\\' && (p[1] == '\n' || p[1] == '\r'))) return (match_len); return (0);/* Not match */ } /* * Test whether the characters 'p' has is mtree keyword. * Returns the length of a detected keyword. * Returns 0 if any keywords were not found. */ static int bid_keyword(const char *p, ssize_t len) { static const char * const keys_c[] = { "content", "contents", "cksum", NULL }; static const char * const keys_df[] = { "device", "flags", NULL }; static const char * const keys_g[] = { "gid", "gname", NULL }; static const char * const keys_il[] = { "ignore", "inode", "link", NULL }; static const char * const keys_m[] = { "md5", "md5digest", "mode", NULL }; static const char * const keys_no[] = { "nlink", "nochange", "optional", NULL }; static const char * const keys_r[] = { "resdevice", "rmd160", "rmd160digest", NULL }; static const char * const keys_s[] = { "sha1", "sha1digest", "sha256", "sha256digest", "sha384", "sha384digest", "sha512", "sha512digest", "size", NULL }; static const char * const keys_t[] = { "tags", "time", "type", NULL }; static const char * const keys_u[] = { "uid", "uname", NULL }; const char * const *keys; int i; switch (*p) { case 'c': keys = keys_c; break; case 'd': case 'f': keys = keys_df; break; case 'g': keys = keys_g; break; case 'i': case 'l': keys = keys_il; break; case 'm': keys = keys_m; break; case 'n': case 'o': keys = keys_no; break; case 'r': keys = keys_r; break; case 's': keys = keys_s; break; case 't': keys = keys_t; break; case 'u': keys = keys_u; break; default: return (0);/* Unknown key */ } for (i = 0; keys[i] != NULL; i++) { int l = bid_keycmp(p, keys[i], len); if (l > 0) return (l); } return (0);/* Unknown key */ } /* * Test whether there is a set of mtree keywords. * Returns the number of keyword. * Returns -1 if we got incorrect sequence. * This function expects a set of "keyword=value". * When "unset" is specified, expects a set of "keyword". */ static int bid_keyword_list(const char *p, ssize_t len, int unset, int last_is_path) { int l; int keycnt = 0; while (len > 0 && *p) { int blank = 0; /* Test whether there are blank characters in the line. */ while (len >0 && (*p == ' ' || *p == '\t')) { ++p; --len; blank = 1; } if (*p == '\n' || *p == '\r') break; if (p[0] == '\\' && (p[1] == '\n' || p[1] == '\r')) break; if (!blank && !last_is_path) /* No blank character. */ return (-1); if (last_is_path && len == 0) return (keycnt); if (unset) { l = bid_keycmp(p, "all", len); if (l > 0) return (1); } /* Test whether there is a correct key in the line. */ l = bid_keyword(p, len); if (l == 0) return (-1);/* Unknown keyword was found. */ p += l; len -= l; keycnt++; /* Skip value */ if (*p == '=') { int value = 0; ++p; --len; while (len > 0 && *p != ' ' && *p != '\t') { ++p; --len; value = 1; } /* A keyword should have a its value unless * "/unset" operation. */ if (!unset && value == 0) return (-1); } } return (keycnt); } static int bid_entry(const char *p, ssize_t len, ssize_t nl, int *last_is_path) { int f = 0; static const unsigned char safe_char[256] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 00 - 0F */ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 10 - 1F */ /* !"$%&'()*+,-./ EXCLUSION:( )(#) */ 0, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 20 - 2F */ /* 0123456789:;<>? EXCLUSION:(=) */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, /* 30 - 3F */ /* @ABCDEFGHIJKLMNO */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 40 - 4F */ /* PQRSTUVWXYZ[\]^_ */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 50 - 5F */ /* `abcdefghijklmno */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 60 - 6F */ /* pqrstuvwxyz{|}~ */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, /* 70 - 7F */ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 80 - 8F */ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 90 - 9F */ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* A0 - AF */ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* B0 - BF */ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* C0 - CF */ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* D0 - DF */ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* E0 - EF */ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* F0 - FF */ }; ssize_t ll; const char *pp = p; const char * const pp_end = pp + len; *last_is_path = 0; /* * Skip the path-name which is quoted. */ for (;pp < pp_end; ++pp) { if (!safe_char[*(const unsigned char *)pp]) { if (*pp != ' ' && *pp != '\t' && *pp != '\r' && *pp != '\n') f = 0; break; } f = 1; } ll = pp_end - pp; /* If a path-name was not found at the first, try to check * a mtree format(a.k.a form D) ``NetBSD's mtree -D'' creates, * which places the path-name at the last. */ if (f == 0) { const char *pb = p + len - nl; int name_len = 0; int slash; /* The form D accepts only a single line for an entry. */ if (pb-2 >= p && pb[-1] == '\\' && (pb[-2] == ' ' || pb[-2] == '\t')) return (-1); if (pb-1 >= p && pb[-1] == '\\') return (-1); slash = 0; while (p <= --pb && *pb != ' ' && *pb != '\t') { if (!safe_char[*(const unsigned char *)pb]) return (-1); name_len++; /* The pathname should have a slash in this * format. */ if (*pb == '/') slash = 1; } if (name_len == 0 || slash == 0) return (-1); /* If '/' is placed at the first in this field, this is not * a valid filename. */ if (pb[1] == '/') return (-1); ll = len - nl - name_len; pp = p; *last_is_path = 1; } return (bid_keyword_list(pp, ll, 0, *last_is_path)); } #define MAX_BID_ENTRY 3 static int mtree_bid(struct archive_read *a, int best_bid) { const char *signature = "#mtree"; const char *p; (void)best_bid; /* UNUSED */ /* Now let's look at the actual header and see if it matches. */ p = __archive_read_ahead(a, strlen(signature), NULL); if (p == NULL) return (-1); if (memcmp(p, signature, strlen(signature)) == 0) return (8 * (int)strlen(signature)); /* * There is not a mtree signature. Let's try to detect mtree format. */ return (detect_form(a, NULL)); } static int detect_form(struct archive_read *a, int *is_form_d) { const char *p; ssize_t avail, ravail; ssize_t detected_bytes = 0, len, nl; int entry_cnt = 0, multiline = 0; int form_D = 0;/* The archive is generated by `NetBSD mtree -D' * (In this source we call it `form D') . */ if (is_form_d != NULL) *is_form_d = 0; p = __archive_read_ahead(a, 1, &avail); if (p == NULL) return (-1); ravail = avail; for (;;) { len = next_line(a, &p, &avail, &ravail, &nl); /* The terminal character of the line should be * a new line character, '\r\n' or '\n'. */ if (len <= 0 || nl == 0) break; if (!multiline) { /* Leading whitespace is never significant, * ignore it. */ while (len > 0 && (*p == ' ' || *p == '\t')) { ++p; --avail; --len; } /* Skip comment or empty line. */ if (p[0] == '#' || p[0] == '\n' || p[0] == '\r') { p += len; avail -= len; continue; } } else { /* A continuance line; the terminal * character of previous line was '\' character. */ if (bid_keyword_list(p, len, 0, 0) <= 0) break; if (multiline == 1) detected_bytes += len; if (p[len-nl-1] != '\\') { if (multiline == 1 && ++entry_cnt >= MAX_BID_ENTRY) break; multiline = 0; } p += len; avail -= len; continue; } if (p[0] != '/') { int last_is_path, keywords; keywords = bid_entry(p, len, nl, &last_is_path); if (keywords >= 0) { detected_bytes += len; if (form_D == 0) { if (last_is_path) form_D = 1; else if (keywords > 0) /* This line is not `form D'. */ form_D = -1; } else if (form_D == 1) { if (!last_is_path && keywords > 0) /* This this is not `form D' * and We cannot accept mixed * format. */ break; } if (!last_is_path && p[len-nl-1] == '\\') /* This line continues. */ multiline = 1; else { /* We've got plenty of correct lines * to assume that this file is a mtree * format. */ if (++entry_cnt >= MAX_BID_ENTRY) break; } } else break; } else if (len > 4 && strncmp(p, "/set", 4) == 0) { if (bid_keyword_list(p+4, len-4, 0, 0) <= 0) break; /* This line continues. */ if (p[len-nl-1] == '\\') multiline = 2; } else if (len > 6 && strncmp(p, "/unset", 6) == 0) { if (bid_keyword_list(p+6, len-6, 1, 0) <= 0) break; /* This line continues. */ if (p[len-nl-1] == '\\') multiline = 2; } else break; /* Test next line. */ p += len; avail -= len; } if (entry_cnt >= MAX_BID_ENTRY || (entry_cnt > 0 && len == 0)) { if (is_form_d != NULL) { if (form_D == 1) *is_form_d = 1; } return (32); } return (0); } /* * The extended mtree format permits multiple lines specifying * attributes for each file. For those entries, only the last line * is actually used. Practically speaking, that means we have * to read the entire mtree file into memory up front. * * The parsing is done in two steps. First, it is decided if a line * changes the global defaults and if it is, processed accordingly. * Otherwise, the options of the line are merged with the current * global options. */ static int add_option(struct archive_read *a, struct mtree_option **global, const char *value, size_t len) { struct mtree_option *opt; if ((opt = malloc(sizeof(*opt))) == NULL) { archive_set_error(&a->archive, errno, "Can't allocate memory"); return (ARCHIVE_FATAL); } if ((opt->value = malloc(len + 1)) == NULL) { free(opt); archive_set_error(&a->archive, errno, "Can't allocate memory"); return (ARCHIVE_FATAL); } memcpy(opt->value, value, len); opt->value[len] = '\0'; opt->next = *global; *global = opt; return (ARCHIVE_OK); } static void remove_option(struct mtree_option **global, const char *value, size_t len) { struct mtree_option *iter, *last; last = NULL; for (iter = *global; iter != NULL; last = iter, iter = iter->next) { if (strncmp(iter->value, value, len) == 0 && (iter->value[len] == '\0' || iter->value[len] == '=')) break; } if (iter == NULL) return; if (last == NULL) *global = iter->next; else last->next = iter->next; free(iter->value); free(iter); } static int process_global_set(struct archive_read *a, struct mtree_option **global, const char *line) { const char *next, *eq; size_t len; int r; line += 4; for (;;) { next = line + strspn(line, " \t\r\n"); if (*next == '\0') return (ARCHIVE_OK); line = next; next = line + strcspn(line, " \t\r\n"); eq = strchr(line, '='); if (eq > next) len = next - line; else len = eq - line; remove_option(global, line, len); r = add_option(a, global, line, next - line); if (r != ARCHIVE_OK) return (r); line = next; } } static int process_global_unset(struct archive_read *a, struct mtree_option **global, const char *line) { const char *next; size_t len; line += 6; if (strchr(line, '=') != NULL) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "/unset shall not contain `='"); return ARCHIVE_FATAL; } for (;;) { next = line + strspn(line, " \t\r\n"); if (*next == '\0') return (ARCHIVE_OK); line = next; len = strcspn(line, " \t\r\n"); if (len == 3 && strncmp(line, "all", 3) == 0) { free_options(*global); *global = NULL; } else { remove_option(global, line, len); } line += len; } } static int process_add_entry(struct archive_read *a, struct mtree *mtree, struct mtree_option **global, const char *line, ssize_t line_len, struct mtree_entry **last_entry, int is_form_d) { struct mtree_entry *entry, *ht_iter; struct mtree_option *iter; const char *next, *eq, *name, *end; size_t name_len, len; int r, i; unsigned int ht_idx; if ((entry = malloc(sizeof(*entry))) == NULL) { archive_set_error(&a->archive, errno, "Can't allocate memory"); return (ARCHIVE_FATAL); } entry->next = NULL; entry->options = NULL; entry->name = NULL; entry->used = 0; entry->full = 0; entry->name_hash = 0; entry->hashtable_next = NULL; /* Add this entry to list. */ if (*last_entry == NULL) mtree->entries = entry; else (*last_entry)->next = entry; *last_entry = entry; if (is_form_d) { /* Filename is last item on line. */ /* Adjust line_len to trim trailing whitespace */ while (line_len > 0) { char last_character = line[line_len - 1]; if (last_character == '\r' || last_character == '\n' || last_character == '\t' || last_character == ' ') { line_len--; } else { break; } } /* Name starts after the last whitespace separator */ name = line; for (i = 0; i < line_len; i++) { if (line[i] == '\r' || line[i] == '\n' || line[i] == '\t' || line[i] == ' ') { name = line + i + 1; } } name_len = line + line_len - name; end = name; } else { /* Filename is first item on line */ name_len = strcspn(line, " \t\r\n"); name = line; line += name_len; end = line + line_len; } /* name/name_len is the name within the line. */ /* line..end brackets the entire line except the name */ if ((entry->name = malloc(name_len + 1)) == NULL) { archive_set_error(&a->archive, errno, "Can't allocate memory"); return (ARCHIVE_FATAL); } memcpy(entry->name, name, name_len); entry->name[name_len] = '\0'; parse_escapes(entry->name, entry); entry->name_hash = hash(entry->name); ht_idx = entry->name_hash % MTREE_HASHTABLE_SIZE; if ((ht_iter = mtree->entry_hashtable[ht_idx]) != NULL) { while (ht_iter->hashtable_next) ht_iter = ht_iter->hashtable_next; ht_iter->hashtable_next = entry; } else { mtree->entry_hashtable[ht_idx] = entry; } for (iter = *global; iter != NULL; iter = iter->next) { r = add_option(a, &entry->options, iter->value, strlen(iter->value)); if (r != ARCHIVE_OK) return (r); } for (;;) { next = line + strspn(line, " \t\r\n"); if (*next == '\0') return (ARCHIVE_OK); if (next >= end) return (ARCHIVE_OK); line = next; next = line + strcspn(line, " \t\r\n"); eq = strchr(line, '='); if (eq == NULL || eq > next) len = next - line; else len = eq - line; remove_option(&entry->options, line, len); r = add_option(a, &entry->options, line, next - line); if (r != ARCHIVE_OK) return (r); line = next; } } static int read_mtree(struct archive_read *a, struct mtree *mtree) { ssize_t len; uintmax_t counter; char *p; struct mtree_option *global; struct mtree_entry *last_entry; int r, is_form_d; mtree->archive_format = ARCHIVE_FORMAT_MTREE; mtree->archive_format_name = "mtree"; global = NULL; last_entry = NULL; (void)detect_form(a, &is_form_d); for (counter = 1; ; ++counter) { len = readline(a, mtree, &p, 65536); if (len == 0) { mtree->this_entry = mtree->entries; free_options(global); return (ARCHIVE_OK); } if (len < 0) { free_options(global); return ((int)len); } /* Leading whitespace is never significant, ignore it. */ while (*p == ' ' || *p == '\t') { ++p; --len; } /* Skip content lines and blank lines. */ if (*p == '#') continue; if (*p == '\r' || *p == '\n' || *p == '\0') continue; if (*p != '/') { r = process_add_entry(a, mtree, &global, p, len, &last_entry, is_form_d); } else if (len > 4 && strncmp(p, "/set", 4) == 0) { if (p[4] != ' ' && p[4] != '\t') break; r = process_global_set(a, &global, p); } else if (len > 6 && strncmp(p, "/unset", 6) == 0) { if (p[6] != ' ' && p[6] != '\t') break; r = process_global_unset(a, &global, p); } else break; if (r != ARCHIVE_OK) { free_options(global); return r; } } archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Can't parse line %ju", counter); free_options(global); return (ARCHIVE_FATAL); } /* * Read in the entire mtree file into memory on the first request. * Then use the next unused file to satisfy each header request. */ static int read_header(struct archive_read *a, struct archive_entry *entry) { struct mtree *mtree; char *p; int r, use_next; mtree = (struct mtree *)(a->format->data); if (mtree->fd >= 0) { close(mtree->fd); mtree->fd = -1; } if (mtree->entries == NULL) { mtree->resolver = archive_entry_linkresolver_new(); if (mtree->resolver == NULL) return ARCHIVE_FATAL; archive_entry_linkresolver_set_strategy(mtree->resolver, ARCHIVE_FORMAT_MTREE); r = read_mtree(a, mtree); if (r != ARCHIVE_OK) return (r); } a->archive.archive_format = mtree->archive_format; a->archive.archive_format_name = mtree->archive_format_name; for (;;) { if (mtree->this_entry == NULL) return (ARCHIVE_EOF); if (strcmp(mtree->this_entry->name, "..") == 0) { mtree->this_entry->used = 1; if (archive_strlen(&mtree->current_dir) > 0) { /* Roll back current path. */ p = mtree->current_dir.s + mtree->current_dir.length - 1; while (p >= mtree->current_dir.s && *p != '/') --p; if (p >= mtree->current_dir.s) --p; mtree->current_dir.length = p - mtree->current_dir.s + 1; } } if (!mtree->this_entry->used) { use_next = 0; r = parse_file(a, entry, mtree, mtree->this_entry, &use_next); if (use_next == 0) return (r); } mtree->this_entry = mtree->this_entry->next; } } /* * A single file can have multiple lines contribute specifications. * Parse as many lines as necessary, then pull additional information * from a backing file on disk as necessary. */ static int parse_file(struct archive_read *a, struct archive_entry *entry, struct mtree *mtree, struct mtree_entry *mentry, int *use_next) { const char *path; struct stat st_storage, *st; struct mtree_entry *mp; struct archive_entry *sparse_entry; int r = ARCHIVE_OK, r1, parsed_kws; mentry->used = 1; /* Initialize reasonable defaults. */ archive_entry_set_filetype(entry, AE_IFREG); archive_entry_set_size(entry, 0); archive_string_empty(&mtree->contents_name); /* Parse options from this line. */ parsed_kws = 0; r = parse_line(a, entry, mtree, mentry, &parsed_kws); if (mentry->full) { archive_entry_copy_pathname(entry, mentry->name); /* * "Full" entries are allowed to have multiple lines * and those lines aren't required to be adjacent. We * don't support multiple lines for "relative" entries * nor do we make any attempt to merge data from * separate "relative" and "full" entries. (Merging * "relative" and "full" entries would require dealing * with pathname canonicalization, which is a very * tricky subject.) */ for (mp = mentry->hashtable_next; mp != NULL; mp = mp->hashtable_next) { if (mp->full && !mp->used && mentry->name_hash == mp->name_hash && strcmp(mentry->name, mp->name) == 0) { /* Later lines override earlier ones. */ mp->used = 1; r1 = parse_line(a, entry, mtree, mp, &parsed_kws); if (r1 < r) r = r1; } } } else { /* * Relative entries require us to construct * the full path and possibly update the * current directory. */ size_t n = archive_strlen(&mtree->current_dir); if (n > 0) archive_strcat(&mtree->current_dir, "/"); archive_strcat(&mtree->current_dir, mentry->name); archive_entry_copy_pathname(entry, mtree->current_dir.s); if (archive_entry_filetype(entry) != AE_IFDIR) mtree->current_dir.length = n; } if (mtree->checkfs) { /* * Try to open and stat the file to get the real size * and other file info. It would be nice to avoid * this here so that getting a listing of an mtree * wouldn't require opening every referenced contents * file. But then we wouldn't know the actual * contents size, so I don't see a really viable way * around this. (Also, we may want to someday pull * other unspecified info from the contents file on * disk.) */ mtree->fd = -1; if (archive_strlen(&mtree->contents_name) > 0) path = mtree->contents_name.s; else path = archive_entry_pathname(entry); if (archive_entry_filetype(entry) == AE_IFREG || archive_entry_filetype(entry) == AE_IFDIR) { mtree->fd = open(path, O_RDONLY | O_BINARY | O_CLOEXEC); __archive_ensure_cloexec_flag(mtree->fd); if (mtree->fd == -1 && (errno != ENOENT || archive_strlen(&mtree->contents_name) > 0)) { archive_set_error(&a->archive, errno, "Can't open %s", path); r = ARCHIVE_WARN; } } st = &st_storage; if (mtree->fd >= 0) { if (fstat(mtree->fd, st) == -1) { archive_set_error(&a->archive, errno, "Could not fstat %s", path); r = ARCHIVE_WARN; /* If we can't stat it, don't keep it open. */ close(mtree->fd); mtree->fd = -1; st = NULL; } } else if (lstat(path, st) == -1) { st = NULL; } /* * Check for a mismatch between the type in the specification * and the type of the contents object on disk. */ if (st != NULL) { if (((st->st_mode & S_IFMT) == S_IFREG && archive_entry_filetype(entry) == AE_IFREG) #ifdef S_IFLNK ||((st->st_mode & S_IFMT) == S_IFLNK && archive_entry_filetype(entry) == AE_IFLNK) #endif #ifdef S_IFSOCK ||((st->st_mode & S_IFSOCK) == S_IFSOCK && archive_entry_filetype(entry) == AE_IFSOCK) #endif #ifdef S_IFCHR ||((st->st_mode & S_IFMT) == S_IFCHR && archive_entry_filetype(entry) == AE_IFCHR) #endif #ifdef S_IFBLK ||((st->st_mode & S_IFMT) == S_IFBLK && archive_entry_filetype(entry) == AE_IFBLK) #endif ||((st->st_mode & S_IFMT) == S_IFDIR && archive_entry_filetype(entry) == AE_IFDIR) #ifdef S_IFIFO ||((st->st_mode & S_IFMT) == S_IFIFO && archive_entry_filetype(entry) == AE_IFIFO) #endif ) { /* Types match. */ } else { /* Types don't match; bail out gracefully. */ if (mtree->fd >= 0) close(mtree->fd); mtree->fd = -1; if (parsed_kws & MTREE_HAS_OPTIONAL) { /* It's not an error for an optional * entry to not match disk. */ *use_next = 1; } else if (r == ARCHIVE_OK) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "mtree specification has different" " type for %s", archive_entry_pathname(entry)); r = ARCHIVE_WARN; } return (r); } } /* * If there is a contents file on disk, pick some of the * metadata from that file. For most of these, we only * set it from the contents if it wasn't already parsed * from the specification. */ if (st != NULL) { if (((parsed_kws & MTREE_HAS_DEVICE) == 0 || (parsed_kws & MTREE_HAS_NOCHANGE) != 0) && (archive_entry_filetype(entry) == AE_IFCHR || archive_entry_filetype(entry) == AE_IFBLK)) archive_entry_set_rdev(entry, st->st_rdev); if ((parsed_kws & (MTREE_HAS_GID | MTREE_HAS_GNAME)) == 0 || (parsed_kws & MTREE_HAS_NOCHANGE) != 0) archive_entry_set_gid(entry, st->st_gid); if ((parsed_kws & (MTREE_HAS_UID | MTREE_HAS_UNAME)) == 0 || (parsed_kws & MTREE_HAS_NOCHANGE) != 0) archive_entry_set_uid(entry, st->st_uid); if ((parsed_kws & MTREE_HAS_MTIME) == 0 || (parsed_kws & MTREE_HAS_NOCHANGE) != 0) { #if HAVE_STRUCT_STAT_ST_MTIMESPEC_TV_NSEC archive_entry_set_mtime(entry, st->st_mtime, st->st_mtimespec.tv_nsec); #elif HAVE_STRUCT_STAT_ST_MTIM_TV_NSEC archive_entry_set_mtime(entry, st->st_mtime, st->st_mtim.tv_nsec); #elif HAVE_STRUCT_STAT_ST_MTIME_N archive_entry_set_mtime(entry, st->st_mtime, st->st_mtime_n); #elif HAVE_STRUCT_STAT_ST_UMTIME archive_entry_set_mtime(entry, st->st_mtime, st->st_umtime*1000); #elif HAVE_STRUCT_STAT_ST_MTIME_USEC archive_entry_set_mtime(entry, st->st_mtime, st->st_mtime_usec*1000); #else archive_entry_set_mtime(entry, st->st_mtime, 0); #endif } if ((parsed_kws & MTREE_HAS_NLINK) == 0 || (parsed_kws & MTREE_HAS_NOCHANGE) != 0) archive_entry_set_nlink(entry, st->st_nlink); if ((parsed_kws & MTREE_HAS_PERM) == 0 || (parsed_kws & MTREE_HAS_NOCHANGE) != 0) archive_entry_set_perm(entry, st->st_mode); if ((parsed_kws & MTREE_HAS_SIZE) == 0 || (parsed_kws & MTREE_HAS_NOCHANGE) != 0) archive_entry_set_size(entry, st->st_size); archive_entry_set_ino(entry, st->st_ino); archive_entry_set_dev(entry, st->st_dev); archive_entry_linkify(mtree->resolver, &entry, &sparse_entry); } else if (parsed_kws & MTREE_HAS_OPTIONAL) { /* * Couldn't open the entry, stat it or the on-disk type * didn't match. If this entry is optional, just * ignore it and read the next header entry. */ *use_next = 1; return ARCHIVE_OK; } } mtree->cur_size = archive_entry_size(entry); mtree->offset = 0; return r; } /* * Each line contains a sequence of keywords. */ static int parse_line(struct archive_read *a, struct archive_entry *entry, struct mtree *mtree, struct mtree_entry *mp, int *parsed_kws) { struct mtree_option *iter; int r = ARCHIVE_OK, r1; for (iter = mp->options; iter != NULL; iter = iter->next) { r1 = parse_keyword(a, mtree, entry, iter, parsed_kws); if (r1 < r) r = r1; } if (r == ARCHIVE_OK && (*parsed_kws & MTREE_HAS_TYPE) == 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Missing type keyword in mtree specification"); return (ARCHIVE_WARN); } return (r); } /* * Device entries have one of the following forms: * - raw dev_t * - format,major,minor[,subdevice] * When parsing succeeded, `pdev' will contain the appropriate dev_t value. */ /* strsep() is not in C90, but strcspn() is. */ /* Taken from http://unixpapa.com/incnote/string.html */ static char * la_strsep(char **sp, const char *sep) { char *p, *s; if (sp == NULL || *sp == NULL || **sp == '\0') return(NULL); s = *sp; p = s + strcspn(s, sep); if (*p != '\0') *p++ = '\0'; *sp = p; return(s); } static int parse_device(dev_t *pdev, struct archive *a, char *val) { #define MAX_PACK_ARGS 3 unsigned long numbers[MAX_PACK_ARGS]; char *p, *dev; int argc; pack_t *pack; dev_t result; const char *error = NULL; memset(pdev, 0, sizeof(*pdev)); if ((dev = strchr(val, ',')) != NULL) { /* * Device's major/minor are given in a specified format. * Decode and pack it accordingly. */ *dev++ = '\0'; if ((pack = pack_find(val)) == NULL) { archive_set_error(a, ARCHIVE_ERRNO_FILE_FORMAT, "Unknown format `%s'", val); return ARCHIVE_WARN; } argc = 0; while ((p = la_strsep(&dev, ",")) != NULL) { if (*p == '\0') { archive_set_error(a, ARCHIVE_ERRNO_FILE_FORMAT, "Missing number"); return ARCHIVE_WARN; } if (argc >= MAX_PACK_ARGS) { archive_set_error(a, ARCHIVE_ERRNO_FILE_FORMAT, "Too many arguments"); return ARCHIVE_WARN; } numbers[argc++] = (unsigned long)mtree_atol(&p, 0); } if (argc < 2) { archive_set_error(a, ARCHIVE_ERRNO_FILE_FORMAT, "Not enough arguments"); return ARCHIVE_WARN; } result = (*pack)(argc, numbers, &error); if (error != NULL) { archive_set_error(a, ARCHIVE_ERRNO_FILE_FORMAT, "%s", error); return ARCHIVE_WARN; } } else { /* file system raw value. */ result = (dev_t)mtree_atol(&val, 0); } *pdev = result; return ARCHIVE_OK; #undef MAX_PACK_ARGS } /* * Parse a single keyword and its value. */ static int parse_keyword(struct archive_read *a, struct mtree *mtree, struct archive_entry *entry, struct mtree_option *opt, int *parsed_kws) { char *val, *key; key = opt->value; if (*key == '\0') return (ARCHIVE_OK); if (strcmp(key, "nochange") == 0) { *parsed_kws |= MTREE_HAS_NOCHANGE; return (ARCHIVE_OK); } if (strcmp(key, "optional") == 0) { *parsed_kws |= MTREE_HAS_OPTIONAL; return (ARCHIVE_OK); } if (strcmp(key, "ignore") == 0) { /* * The mtree processing is not recursive, so * recursion will only happen for explicitly listed * entries. */ return (ARCHIVE_OK); } val = strchr(key, '='); if (val == NULL) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Malformed attribute \"%s\" (%d)", key, key[0]); return (ARCHIVE_WARN); } *val = '\0'; ++val; switch (key[0]) { case 'c': if (strcmp(key, "content") == 0 || strcmp(key, "contents") == 0) { parse_escapes(val, NULL); archive_strcpy(&mtree->contents_name, val); break; } if (strcmp(key, "cksum") == 0) break; case 'd': if (strcmp(key, "device") == 0) { /* stat(2) st_rdev field, e.g. the major/minor IDs * of a char/block special file */ int r; dev_t dev; *parsed_kws |= MTREE_HAS_DEVICE; r = parse_device(&dev, &a->archive, val); if (r == ARCHIVE_OK) archive_entry_set_rdev(entry, dev); return r; } case 'f': if (strcmp(key, "flags") == 0) { *parsed_kws |= MTREE_HAS_FFLAGS; archive_entry_copy_fflags_text(entry, val); break; } case 'g': if (strcmp(key, "gid") == 0) { *parsed_kws |= MTREE_HAS_GID; archive_entry_set_gid(entry, mtree_atol(&val, 10)); break; } if (strcmp(key, "gname") == 0) { *parsed_kws |= MTREE_HAS_GNAME; archive_entry_copy_gname(entry, val); break; } case 'i': if (strcmp(key, "inode") == 0) { archive_entry_set_ino(entry, mtree_atol(&val, 10)); break; } case 'l': if (strcmp(key, "link") == 0) { archive_entry_copy_symlink(entry, val); break; } case 'm': if (strcmp(key, "md5") == 0 || strcmp(key, "md5digest") == 0) break; if (strcmp(key, "mode") == 0) { if (val[0] >= '0' && val[0] <= '7') { *parsed_kws |= MTREE_HAS_PERM; archive_entry_set_perm(entry, (mode_t)mtree_atol(&val, 8)); } else { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Symbolic or non-octal mode \"%s\" unsupported", val); return ARCHIVE_WARN; } break; } case 'n': if (strcmp(key, "nlink") == 0) { *parsed_kws |= MTREE_HAS_NLINK; archive_entry_set_nlink(entry, (unsigned int)mtree_atol(&val, 10)); break; } case 'r': if (strcmp(key, "resdevice") == 0) { /* stat(2) st_dev field, e.g. the device ID where the * inode resides */ int r; dev_t dev; r = parse_device(&dev, &a->archive, val); if (r == ARCHIVE_OK) archive_entry_set_dev(entry, dev); return r; } if (strcmp(key, "rmd160") == 0 || strcmp(key, "rmd160digest") == 0) break; case 's': if (strcmp(key, "sha1") == 0 || strcmp(key, "sha1digest") == 0) break; if (strcmp(key, "sha256") == 0 || strcmp(key, "sha256digest") == 0) break; if (strcmp(key, "sha384") == 0 || strcmp(key, "sha384digest") == 0) break; if (strcmp(key, "sha512") == 0 || strcmp(key, "sha512digest") == 0) break; if (strcmp(key, "size") == 0) { archive_entry_set_size(entry, mtree_atol(&val, 10)); break; } case 't': if (strcmp(key, "tags") == 0) { /* * Comma delimited list of tags. * Ignore the tags for now, but the interface * should be extended to allow inclusion/exclusion. */ break; } if (strcmp(key, "time") == 0) { int64_t m; int64_t my_time_t_max = get_time_t_max(); int64_t my_time_t_min = get_time_t_min(); long ns = 0; *parsed_kws |= MTREE_HAS_MTIME; m = mtree_atol(&val, 10); /* Replicate an old mtree bug: * 123456789.1 represents 123456789 * seconds and 1 nanosecond. */ if (*val == '.') { ++val; ns = (long)mtree_atol(&val, 10); if (ns < 0) ns = 0; else if (ns > 999999999) ns = 999999999; } if (m > my_time_t_max) m = my_time_t_max; else if (m < my_time_t_min) m = my_time_t_min; archive_entry_set_mtime(entry, (time_t)m, ns); break; } if (strcmp(key, "type") == 0) { switch (val[0]) { case 'b': if (strcmp(val, "block") == 0) { archive_entry_set_filetype(entry, AE_IFBLK); break; } case 'c': if (strcmp(val, "char") == 0) { archive_entry_set_filetype(entry, AE_IFCHR); break; } case 'd': if (strcmp(val, "dir") == 0) { archive_entry_set_filetype(entry, AE_IFDIR); break; } case 'f': if (strcmp(val, "fifo") == 0) { archive_entry_set_filetype(entry, AE_IFIFO); break; } if (strcmp(val, "file") == 0) { archive_entry_set_filetype(entry, AE_IFREG); break; } case 'l': if (strcmp(val, "link") == 0) { archive_entry_set_filetype(entry, AE_IFLNK); break; } default: archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Unrecognized file type \"%s\"; " "assuming \"file\"", val); archive_entry_set_filetype(entry, AE_IFREG); return (ARCHIVE_WARN); } *parsed_kws |= MTREE_HAS_TYPE; break; } case 'u': if (strcmp(key, "uid") == 0) { *parsed_kws |= MTREE_HAS_UID; archive_entry_set_uid(entry, mtree_atol(&val, 10)); break; } if (strcmp(key, "uname") == 0) { *parsed_kws |= MTREE_HAS_UNAME; archive_entry_copy_uname(entry, val); break; } default: archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Unrecognized key %s=%s", key, val); return (ARCHIVE_WARN); } return (ARCHIVE_OK); } static int read_data(struct archive_read *a, const void **buff, size_t *size, int64_t *offset) { size_t bytes_to_read; ssize_t bytes_read; struct mtree *mtree; mtree = (struct mtree *)(a->format->data); if (mtree->fd < 0) { *buff = NULL; *offset = 0; *size = 0; return (ARCHIVE_EOF); } if (mtree->buff == NULL) { mtree->buffsize = 64 * 1024; mtree->buff = malloc(mtree->buffsize); if (mtree->buff == NULL) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory"); return (ARCHIVE_FATAL); } } *buff = mtree->buff; *offset = mtree->offset; if ((int64_t)mtree->buffsize > mtree->cur_size - mtree->offset) bytes_to_read = (size_t)(mtree->cur_size - mtree->offset); else bytes_to_read = mtree->buffsize; bytes_read = read(mtree->fd, mtree->buff, bytes_to_read); if (bytes_read < 0) { archive_set_error(&a->archive, errno, "Can't read"); return (ARCHIVE_WARN); } if (bytes_read == 0) { *size = 0; return (ARCHIVE_EOF); } mtree->offset += bytes_read; *size = bytes_read; return (ARCHIVE_OK); } /* Skip does nothing except possibly close the contents file. */ static int skip(struct archive_read *a) { struct mtree *mtree; mtree = (struct mtree *)(a->format->data); if (mtree->fd >= 0) { close(mtree->fd); mtree->fd = -1; } return (ARCHIVE_OK); } /* * Since parsing backslash sequences always makes strings shorter, * we can always do this conversion in-place. */ static void parse_escapes(char *src, struct mtree_entry *mentry) { char *dest = src; char c; if (mentry != NULL && strcmp(src, ".") == 0) mentry->full = 1; while (*src != '\0') { c = *src++; if (c == '/' && mentry != NULL) mentry->full = 1; if (c == '\\') { switch (src[0]) { case '0': if (src[1] < '0' || src[1] > '7') { c = 0; ++src; break; } /* FALLTHROUGH */ case '1': case '2': case '3': if (src[1] >= '0' && src[1] <= '7' && src[2] >= '0' && src[2] <= '7') { c = (src[0] - '0') << 6; c |= (src[1] - '0') << 3; c |= (src[2] - '0'); src += 3; } break; case 'a': c = '\a'; ++src; break; case 'b': c = '\b'; ++src; break; case 'f': c = '\f'; ++src; break; case 'n': c = '\n'; ++src; break; case 'r': c = '\r'; ++src; break; case 's': c = ' '; ++src; break; case 't': c = '\t'; ++src; break; case 'v': c = '\v'; ++src; break; case '\\': c = '\\'; ++src; break; } } *dest++ = c; } *dest = '\0'; } /* Parse a hex digit. */ static int parsedigit(char c) { if (c >= '0' && c <= '9') return c - '0'; else if (c >= 'a' && c <= 'f') return c - 'a'; else if (c >= 'A' && c <= 'F') return c - 'A'; else return -1; } /* * Note that this implementation does not (and should not!) obey * locale settings; you cannot simply substitute strtol here, since * it does obey locale. */ static int64_t mtree_atol(char **p, int base) { int64_t l, limit; int digit, last_digit_limit; if (base == 0) { if (**p != '0') base = 10; else if ((*p)[1] == 'x' || (*p)[1] == 'X') { *p += 2; base = 16; } else { base = 8; } } if (**p == '-') { limit = INT64_MIN / base; last_digit_limit = INT64_MIN % base; ++(*p); l = 0; digit = parsedigit(**p); while (digit >= 0 && digit < base) { if (l < limit || (l == limit && digit > last_digit_limit)) return INT64_MIN; l = (l * base) - digit; digit = parsedigit(*++(*p)); } return l; } else { limit = INT64_MAX / base; last_digit_limit = INT64_MAX % base; l = 0; digit = parsedigit(**p); while (digit >= 0 && digit < base) { if (l > limit || (l == limit && digit > last_digit_limit)) return INT64_MAX; l = (l * base) + digit; digit = parsedigit(*++(*p)); } return l; } } /* * Returns length of line (including trailing newline) * or negative on error. 'start' argument is updated to * point to first character of line. */ static ssize_t readline(struct archive_read *a, struct mtree *mtree, char **start, ssize_t limit) { ssize_t bytes_read; ssize_t total_size = 0; ssize_t find_off = 0; const void *t; void *nl; char *u; /* Accumulate line in a line buffer. */ for (;;) { /* Read some more. */ t = __archive_read_ahead(a, 1, &bytes_read); if (t == NULL) return (0); if (bytes_read < 0) return (ARCHIVE_FATAL); nl = memchr(t, '\n', bytes_read); /* If we found '\n', trim the read to end exactly there. */ if (nl != NULL) { bytes_read = ((const char *)nl) - ((const char *)t) + 1; } if (total_size + bytes_read + 1 > limit) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Line too long"); return (ARCHIVE_FATAL); } if (archive_string_ensure(&mtree->line, total_size + bytes_read + 1) == NULL) { archive_set_error(&a->archive, ENOMEM, "Can't allocate working buffer"); return (ARCHIVE_FATAL); } /* Append new bytes to string. */ memcpy(mtree->line.s + total_size, t, bytes_read); __archive_read_consume(a, bytes_read); total_size += bytes_read; mtree->line.s[total_size] = '\0'; for (u = mtree->line.s + find_off; *u; ++u) { if (u[0] == '\n') { /* Ends with unescaped newline. */ *start = mtree->line.s; return total_size; } else if (u[0] == '#') { /* Ends with comment sequence #...\n */ if (nl == NULL) { /* But we've not found the \n yet */ break; } } else if (u[0] == '\\') { if (u[1] == '\n') { /* Trim escaped newline. */ total_size -= 2; mtree->line.s[total_size] = '\0'; break; } else if (u[1] != '\0') { /* Skip the two-char escape sequence */ ++u; } } } find_off = u - mtree->line.s; } } static unsigned int hash(const char *p) { /* A 32-bit version of Peter Weinberger's (PJW) hash algorithm, as used by ELF for hashing function names. */ unsigned g, h = 0; while (*p != '\0') { h = (h << 4) + *p++; if ((g = h & 0xF0000000) != 0) { h ^= g >> 24; h &= 0x0FFFFFFF; } } return h; } Index: stable/10/contrib/libarchive/libarchive/archive_read_support_format_rar.c =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_read_support_format_rar.c (revision 324417) +++ stable/10/contrib/libarchive/libarchive/archive_read_support_format_rar.c (revision 324418) @@ -1,2953 +1,2957 @@ /*- * Copyright (c) 2003-2007 Tim Kientzle * Copyright (c) 2011 Andres Mejia * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "archive_platform.h" #ifdef HAVE_ERRNO_H #include #endif #include #include #ifdef HAVE_ZLIB_H #include /* crc32 */ #endif #include "archive.h" #ifndef HAVE_ZLIB_H #include "archive_crc32.h" #endif #include "archive_endian.h" #include "archive_entry.h" #include "archive_entry_locale.h" #include "archive_ppmd7_private.h" #include "archive_private.h" #include "archive_read_private.h" /* RAR signature, also known as the mark header */ #define RAR_SIGNATURE "\x52\x61\x72\x21\x1A\x07\x00" /* Header types */ #define MARK_HEAD 0x72 #define MAIN_HEAD 0x73 #define FILE_HEAD 0x74 #define COMM_HEAD 0x75 #define AV_HEAD 0x76 #define SUB_HEAD 0x77 #define PROTECT_HEAD 0x78 #define SIGN_HEAD 0x79 #define NEWSUB_HEAD 0x7a #define ENDARC_HEAD 0x7b /* Main Header Flags */ #define MHD_VOLUME 0x0001 #define MHD_COMMENT 0x0002 #define MHD_LOCK 0x0004 #define MHD_SOLID 0x0008 #define MHD_NEWNUMBERING 0x0010 #define MHD_AV 0x0020 #define MHD_PROTECT 0x0040 #define MHD_PASSWORD 0x0080 #define MHD_FIRSTVOLUME 0x0100 #define MHD_ENCRYPTVER 0x0200 /* Flags common to all headers */ #define HD_MARKDELETION 0x4000 #define HD_ADD_SIZE_PRESENT 0x8000 /* File Header Flags */ #define FHD_SPLIT_BEFORE 0x0001 #define FHD_SPLIT_AFTER 0x0002 #define FHD_PASSWORD 0x0004 #define FHD_COMMENT 0x0008 #define FHD_SOLID 0x0010 #define FHD_LARGE 0x0100 #define FHD_UNICODE 0x0200 #define FHD_SALT 0x0400 #define FHD_VERSION 0x0800 #define FHD_EXTTIME 0x1000 #define FHD_EXTFLAGS 0x2000 /* File dictionary sizes */ #define DICTIONARY_SIZE_64 0x00 #define DICTIONARY_SIZE_128 0x20 #define DICTIONARY_SIZE_256 0x40 #define DICTIONARY_SIZE_512 0x60 #define DICTIONARY_SIZE_1024 0x80 #define DICTIONARY_SIZE_2048 0xA0 #define DICTIONARY_SIZE_4096 0xC0 #define FILE_IS_DIRECTORY 0xE0 #define DICTIONARY_MASK FILE_IS_DIRECTORY /* OS Flags */ #define OS_MSDOS 0 #define OS_OS2 1 #define OS_WIN32 2 #define OS_UNIX 3 #define OS_MAC_OS 4 #define OS_BEOS 5 /* Compression Methods */ #define COMPRESS_METHOD_STORE 0x30 /* LZSS */ #define COMPRESS_METHOD_FASTEST 0x31 #define COMPRESS_METHOD_FAST 0x32 #define COMPRESS_METHOD_NORMAL 0x33 /* PPMd Variant H */ #define COMPRESS_METHOD_GOOD 0x34 #define COMPRESS_METHOD_BEST 0x35 #define CRC_POLYNOMIAL 0xEDB88320 #define NS_UNIT 10000000 #define DICTIONARY_MAX_SIZE 0x400000 #define MAINCODE_SIZE 299 #define OFFSETCODE_SIZE 60 #define LOWOFFSETCODE_SIZE 17 #define LENGTHCODE_SIZE 28 #define HUFFMAN_TABLE_SIZE \ MAINCODE_SIZE + OFFSETCODE_SIZE + LOWOFFSETCODE_SIZE + LENGTHCODE_SIZE #define MAX_SYMBOL_LENGTH 0xF #define MAX_SYMBOLS 20 /* * Considering L1,L2 cache miss and a calling of write system-call, * the best size of the output buffer(uncompressed buffer) is 128K. * If the structure of extracting process is changed, this value * might be researched again. */ #define UNP_BUFFER_SIZE (128 * 1024) /* Define this here for non-Windows platforms */ #if !((defined(__WIN32__) || defined(_WIN32) || defined(__WIN32)) && !defined(__CYGWIN__)) #define FILE_ATTRIBUTE_DIRECTORY 0x10 #endif /* Fields common to all headers */ struct rar_header { char crc[2]; char type; char flags[2]; char size[2]; }; /* Fields common to all file headers */ struct rar_file_header { char pack_size[4]; char unp_size[4]; char host_os; char file_crc[4]; char file_time[4]; char unp_ver; char method; char name_size[2]; char file_attr[4]; }; struct huffman_tree_node { int branches[2]; }; struct huffman_table_entry { unsigned int length; int value; }; struct huffman_code { struct huffman_tree_node *tree; int numentries; int numallocatedentries; int minlength; int maxlength; int tablesize; struct huffman_table_entry *table; }; struct lzss { unsigned char *window; int mask; int64_t position; }; struct data_block_offsets { int64_t header_size; int64_t start_offset; int64_t end_offset; }; struct rar { /* Entries from main RAR header */ unsigned main_flags; unsigned long file_crc; char reserved1[2]; char reserved2[4]; char encryptver; /* File header entries */ char compression_method; unsigned file_flags; int64_t packed_size; int64_t unp_size; time_t mtime; long mnsec; mode_t mode; char *filename; char *filename_save; size_t filename_save_size; size_t filename_allocated; /* File header optional entries */ char salt[8]; time_t atime; long ansec; time_t ctime; long cnsec; time_t arctime; long arcnsec; /* Fields to help with tracking decompression of files. */ int64_t bytes_unconsumed; int64_t bytes_remaining; int64_t bytes_uncopied; int64_t offset; int64_t offset_outgoing; int64_t offset_seek; char valid; unsigned int unp_offset; unsigned int unp_buffer_size; unsigned char *unp_buffer; unsigned int dictionary_size; char start_new_block; char entry_eof; unsigned long crc_calculated; int found_first_header; char has_endarc_header; struct data_block_offsets *dbo; unsigned int cursor; unsigned int nodes; /* LZSS members */ struct huffman_code maincode; struct huffman_code offsetcode; struct huffman_code lowoffsetcode; struct huffman_code lengthcode; unsigned char lengthtable[HUFFMAN_TABLE_SIZE]; struct lzss lzss; char output_last_match; unsigned int lastlength; unsigned int lastoffset; unsigned int oldoffset[4]; unsigned int lastlowoffset; unsigned int numlowoffsetrepeats; int64_t filterstart; char start_new_table; /* PPMd Variant H members */ char ppmd_valid; char ppmd_eod; char is_ppmd_block; int ppmd_escape; CPpmd7 ppmd7_context; CPpmd7z_RangeDec range_dec; IByteIn bytein; /* * String conversion object. */ int init_default_conversion; struct archive_string_conv *sconv_default; struct archive_string_conv *opt_sconv; struct archive_string_conv *sconv_utf8; struct archive_string_conv *sconv_utf16be; /* * Bit stream reader. */ struct rar_br { #define CACHE_TYPE uint64_t #define CACHE_BITS (8 * sizeof(CACHE_TYPE)) /* Cache buffer. */ CACHE_TYPE cache_buffer; /* Indicates how many bits avail in cache_buffer. */ int cache_avail; ssize_t avail_in; const unsigned char *next_in; } br; /* * Custom field to denote that this archive contains encrypted entries */ int has_encrypted_entries; }; static int archive_read_support_format_rar_capabilities(struct archive_read *); static int archive_read_format_rar_has_encrypted_entries(struct archive_read *); static int archive_read_format_rar_bid(struct archive_read *, int); static int archive_read_format_rar_options(struct archive_read *, const char *, const char *); static int archive_read_format_rar_read_header(struct archive_read *, struct archive_entry *); static int archive_read_format_rar_read_data(struct archive_read *, const void **, size_t *, int64_t *); static int archive_read_format_rar_read_data_skip(struct archive_read *a); static int64_t archive_read_format_rar_seek_data(struct archive_read *, int64_t, int); static int archive_read_format_rar_cleanup(struct archive_read *); /* Support functions */ static int read_header(struct archive_read *, struct archive_entry *, char); static time_t get_time(int); static int read_exttime(const char *, struct rar *, const char *); static int read_symlink_stored(struct archive_read *, struct archive_entry *, struct archive_string_conv *); static int read_data_stored(struct archive_read *, const void **, size_t *, int64_t *); static int read_data_compressed(struct archive_read *, const void **, size_t *, int64_t *); static int rar_br_preparation(struct archive_read *, struct rar_br *); static int parse_codes(struct archive_read *); static void free_codes(struct archive_read *); static int read_next_symbol(struct archive_read *, struct huffman_code *); static int create_code(struct archive_read *, struct huffman_code *, unsigned char *, int, char); static int add_value(struct archive_read *, struct huffman_code *, int, int, int); static int new_node(struct huffman_code *); static int make_table(struct archive_read *, struct huffman_code *); static int make_table_recurse(struct archive_read *, struct huffman_code *, int, struct huffman_table_entry *, int, int); static int64_t expand(struct archive_read *, int64_t); static int copy_from_lzss_window(struct archive_read *, const void **, int64_t, int); static const void *rar_read_ahead(struct archive_read *, size_t, ssize_t *); /* * Bit stream reader. */ /* Check that the cache buffer has enough bits. */ #define rar_br_has(br, n) ((br)->cache_avail >= n) /* Get compressed data by bit. */ #define rar_br_bits(br, n) \ (((uint32_t)((br)->cache_buffer >> \ ((br)->cache_avail - (n)))) & cache_masks[n]) #define rar_br_bits_forced(br, n) \ (((uint32_t)((br)->cache_buffer << \ ((n) - (br)->cache_avail))) & cache_masks[n]) /* Read ahead to make sure the cache buffer has enough compressed data we * will use. * True : completed, there is enough data in the cache buffer. * False : there is no data in the stream. */ #define rar_br_read_ahead(a, br, n) \ ((rar_br_has(br, (n)) || rar_br_fillup(a, br)) || rar_br_has(br, (n))) /* Notify how many bits we consumed. */ #define rar_br_consume(br, n) ((br)->cache_avail -= (n)) #define rar_br_consume_unalined_bits(br) ((br)->cache_avail &= ~7) static const uint32_t cache_masks[] = { 0x00000000, 0x00000001, 0x00000003, 0x00000007, 0x0000000F, 0x0000001F, 0x0000003F, 0x0000007F, 0x000000FF, 0x000001FF, 0x000003FF, 0x000007FF, 0x00000FFF, 0x00001FFF, 0x00003FFF, 0x00007FFF, 0x0000FFFF, 0x0001FFFF, 0x0003FFFF, 0x0007FFFF, 0x000FFFFF, 0x001FFFFF, 0x003FFFFF, 0x007FFFFF, 0x00FFFFFF, 0x01FFFFFF, 0x03FFFFFF, 0x07FFFFFF, 0x0FFFFFFF, 0x1FFFFFFF, 0x3FFFFFFF, 0x7FFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF }; /* * Shift away used bits in the cache data and fill it up with following bits. * Call this when cache buffer does not have enough bits you need. * * Returns 1 if the cache buffer is full. * Returns 0 if the cache buffer is not full; input buffer is empty. */ static int rar_br_fillup(struct archive_read *a, struct rar_br *br) { struct rar *rar = (struct rar *)(a->format->data); int n = CACHE_BITS - br->cache_avail; for (;;) { switch (n >> 3) { case 8: if (br->avail_in >= 8) { br->cache_buffer = ((uint64_t)br->next_in[0]) << 56 | ((uint64_t)br->next_in[1]) << 48 | ((uint64_t)br->next_in[2]) << 40 | ((uint64_t)br->next_in[3]) << 32 | ((uint32_t)br->next_in[4]) << 24 | ((uint32_t)br->next_in[5]) << 16 | ((uint32_t)br->next_in[6]) << 8 | (uint32_t)br->next_in[7]; br->next_in += 8; br->avail_in -= 8; br->cache_avail += 8 * 8; rar->bytes_unconsumed += 8; rar->bytes_remaining -= 8; return (1); } break; case 7: if (br->avail_in >= 7) { br->cache_buffer = (br->cache_buffer << 56) | ((uint64_t)br->next_in[0]) << 48 | ((uint64_t)br->next_in[1]) << 40 | ((uint64_t)br->next_in[2]) << 32 | ((uint32_t)br->next_in[3]) << 24 | ((uint32_t)br->next_in[4]) << 16 | ((uint32_t)br->next_in[5]) << 8 | (uint32_t)br->next_in[6]; br->next_in += 7; br->avail_in -= 7; br->cache_avail += 7 * 8; rar->bytes_unconsumed += 7; rar->bytes_remaining -= 7; return (1); } break; case 6: if (br->avail_in >= 6) { br->cache_buffer = (br->cache_buffer << 48) | ((uint64_t)br->next_in[0]) << 40 | ((uint64_t)br->next_in[1]) << 32 | ((uint32_t)br->next_in[2]) << 24 | ((uint32_t)br->next_in[3]) << 16 | ((uint32_t)br->next_in[4]) << 8 | (uint32_t)br->next_in[5]; br->next_in += 6; br->avail_in -= 6; br->cache_avail += 6 * 8; rar->bytes_unconsumed += 6; rar->bytes_remaining -= 6; return (1); } break; case 0: /* We have enough compressed data in * the cache buffer.*/ return (1); default: break; } if (br->avail_in <= 0) { if (rar->bytes_unconsumed > 0) { /* Consume as much as the decompressor * actually used. */ __archive_read_consume(a, rar->bytes_unconsumed); rar->bytes_unconsumed = 0; } br->next_in = rar_read_ahead(a, 1, &(br->avail_in)); if (br->next_in == NULL) return (0); if (br->avail_in == 0) return (0); } br->cache_buffer = (br->cache_buffer << 8) | *br->next_in++; br->avail_in--; br->cache_avail += 8; n -= 8; rar->bytes_unconsumed++; rar->bytes_remaining--; } } static int rar_br_preparation(struct archive_read *a, struct rar_br *br) { struct rar *rar = (struct rar *)(a->format->data); if (rar->bytes_remaining > 0) { br->next_in = rar_read_ahead(a, 1, &(br->avail_in)); if (br->next_in == NULL) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated RAR file data"); return (ARCHIVE_FATAL); } if (br->cache_avail == 0) (void)rar_br_fillup(a, br); } return (ARCHIVE_OK); } /* Find last bit set */ static inline int rar_fls(unsigned int word) { word |= (word >> 1); word |= (word >> 2); word |= (word >> 4); word |= (word >> 8); word |= (word >> 16); return word - (word >> 1); } /* LZSS functions */ static inline int64_t lzss_position(struct lzss *lzss) { return lzss->position; } static inline int lzss_mask(struct lzss *lzss) { return lzss->mask; } static inline int lzss_size(struct lzss *lzss) { return lzss->mask + 1; } static inline int lzss_offset_for_position(struct lzss *lzss, int64_t pos) { return (int)(pos & lzss->mask); } static inline unsigned char * lzss_pointer_for_position(struct lzss *lzss, int64_t pos) { return &lzss->window[lzss_offset_for_position(lzss, pos)]; } static inline int lzss_current_offset(struct lzss *lzss) { return lzss_offset_for_position(lzss, lzss->position); } static inline uint8_t * lzss_current_pointer(struct lzss *lzss) { return lzss_pointer_for_position(lzss, lzss->position); } static inline void lzss_emit_literal(struct rar *rar, uint8_t literal) { *lzss_current_pointer(&rar->lzss) = literal; rar->lzss.position++; } static inline void lzss_emit_match(struct rar *rar, int offset, int length) { int dstoffs = lzss_current_offset(&rar->lzss); int srcoffs = (dstoffs - offset) & lzss_mask(&rar->lzss); int l, li, remaining; unsigned char *d, *s; remaining = length; while (remaining > 0) { l = remaining; if (dstoffs > srcoffs) { if (l > lzss_size(&rar->lzss) - dstoffs) l = lzss_size(&rar->lzss) - dstoffs; } else { if (l > lzss_size(&rar->lzss) - srcoffs) l = lzss_size(&rar->lzss) - srcoffs; } d = &(rar->lzss.window[dstoffs]); s = &(rar->lzss.window[srcoffs]); if ((dstoffs + l < srcoffs) || (srcoffs + l < dstoffs)) memcpy(d, s, l); else { for (li = 0; li < l; li++) d[li] = s[li]; } remaining -= l; dstoffs = (dstoffs + l) & lzss_mask(&(rar->lzss)); srcoffs = (srcoffs + l) & lzss_mask(&(rar->lzss)); } rar->lzss.position += length; } static void * ppmd_alloc(void *p, size_t size) { (void)p; return malloc(size); } static void ppmd_free(void *p, void *address) { (void)p; free(address); } static ISzAlloc g_szalloc = { ppmd_alloc, ppmd_free }; static Byte ppmd_read(void *p) { struct archive_read *a = ((IByteIn*)p)->a; struct rar *rar = (struct rar *)(a->format->data); struct rar_br *br = &(rar->br); Byte b; if (!rar_br_read_ahead(a, br, 8)) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated RAR file data"); rar->valid = 0; return 0; } b = rar_br_bits(br, 8); rar_br_consume(br, 8); return b; } int archive_read_support_format_rar(struct archive *_a) { struct archive_read *a = (struct archive_read *)_a; struct rar *rar; int r; archive_check_magic(_a, ARCHIVE_READ_MAGIC, ARCHIVE_STATE_NEW, "archive_read_support_format_rar"); rar = (struct rar *)calloc(sizeof(*rar), 1); if (rar == NULL) { archive_set_error(&a->archive, ENOMEM, "Can't allocate rar data"); return (ARCHIVE_FATAL); } /* * Until enough data has been read, we cannot tell about * any encrypted entries yet. */ rar->has_encrypted_entries = ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW; r = __archive_read_register_format(a, rar, "rar", archive_read_format_rar_bid, archive_read_format_rar_options, archive_read_format_rar_read_header, archive_read_format_rar_read_data, archive_read_format_rar_read_data_skip, archive_read_format_rar_seek_data, archive_read_format_rar_cleanup, archive_read_support_format_rar_capabilities, archive_read_format_rar_has_encrypted_entries); if (r != ARCHIVE_OK) free(rar); return (r); } static int archive_read_support_format_rar_capabilities(struct archive_read * a) { (void)a; /* UNUSED */ return (ARCHIVE_READ_FORMAT_CAPS_ENCRYPT_DATA | ARCHIVE_READ_FORMAT_CAPS_ENCRYPT_METADATA); } static int archive_read_format_rar_has_encrypted_entries(struct archive_read *_a) { if (_a && _a->format) { struct rar * rar = (struct rar *)_a->format->data; if (rar) { return rar->has_encrypted_entries; } } return ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW; } static int archive_read_format_rar_bid(struct archive_read *a, int best_bid) { const char *p; /* If there's already a bid > 30, we'll never win. */ if (best_bid > 30) return (-1); if ((p = __archive_read_ahead(a, 7, NULL)) == NULL) return (-1); if (memcmp(p, RAR_SIGNATURE, 7) == 0) return (30); if ((p[0] == 'M' && p[1] == 'Z') || memcmp(p, "\x7F\x45LF", 4) == 0) { /* This is a PE file */ ssize_t offset = 0x10000; ssize_t window = 4096; ssize_t bytes_avail; while (offset + window <= (1024 * 128)) { const char *buff = __archive_read_ahead(a, offset + window, &bytes_avail); if (buff == NULL) { /* Remaining bytes are less than window. */ window >>= 1; if (window < 0x40) return (0); continue; } p = buff + offset; while (p + 7 < buff + bytes_avail) { if (memcmp(p, RAR_SIGNATURE, 7) == 0) return (30); p += 0x10; } offset = p - buff; } } return (0); } static int skip_sfx(struct archive_read *a) { const void *h; const char *p, *q; size_t skip, total; ssize_t bytes, window; total = 0; window = 4096; while (total + window <= (1024 * 128)) { h = __archive_read_ahead(a, window, &bytes); if (h == NULL) { /* Remaining bytes are less than window. */ window >>= 1; if (window < 0x40) goto fatal; continue; } if (bytes < 0x40) goto fatal; p = h; q = p + bytes; /* * Scan ahead until we find something that looks * like the RAR header. */ while (p + 7 < q) { if (memcmp(p, RAR_SIGNATURE, 7) == 0) { skip = p - (const char *)h; __archive_read_consume(a, skip); return (ARCHIVE_OK); } p += 0x10; } skip = p - (const char *)h; __archive_read_consume(a, skip); total += skip; } fatal: archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Couldn't find out RAR header"); return (ARCHIVE_FATAL); } static int archive_read_format_rar_options(struct archive_read *a, const char *key, const char *val) { struct rar *rar; int ret = ARCHIVE_FAILED; rar = (struct rar *)(a->format->data); if (strcmp(key, "hdrcharset") == 0) { if (val == NULL || val[0] == 0) archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "rar: hdrcharset option needs a character-set name"); else { rar->opt_sconv = archive_string_conversion_from_charset( &a->archive, val, 0); if (rar->opt_sconv != NULL) ret = ARCHIVE_OK; else ret = ARCHIVE_FATAL; } return (ret); } /* Note: The "warn" return is just to inform the options * supervisor that we didn't handle it. It will generate * a suitable error if no one used this option. */ return (ARCHIVE_WARN); } static int archive_read_format_rar_read_header(struct archive_read *a, struct archive_entry *entry) { const void *h; const char *p; struct rar *rar; size_t skip; char head_type; int ret; unsigned flags; unsigned long crc32_expected; a->archive.archive_format = ARCHIVE_FORMAT_RAR; if (a->archive.archive_format_name == NULL) a->archive.archive_format_name = "RAR"; rar = (struct rar *)(a->format->data); /* * It should be sufficient to call archive_read_next_header() for * a reader to determine if an entry is encrypted or not. If the * encryption of an entry is only detectable when calling * archive_read_data(), so be it. We'll do the same check there * as well. */ if (rar->has_encrypted_entries == ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW) { rar->has_encrypted_entries = 0; } /* RAR files can be generated without EOF headers, so return ARCHIVE_EOF if * this fails. */ if ((h = __archive_read_ahead(a, 7, NULL)) == NULL) return (ARCHIVE_EOF); p = h; if (rar->found_first_header == 0 && ((p[0] == 'M' && p[1] == 'Z') || memcmp(p, "\x7F\x45LF", 4) == 0)) { /* This is an executable ? Must be self-extracting... */ ret = skip_sfx(a); if (ret < ARCHIVE_WARN) return (ret); } rar->found_first_header = 1; while (1) { unsigned long crc32_val; if ((h = __archive_read_ahead(a, 7, NULL)) == NULL) return (ARCHIVE_FATAL); p = h; head_type = p[2]; switch(head_type) { case MARK_HEAD: if (memcmp(p, RAR_SIGNATURE, 7) != 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid marker header"); return (ARCHIVE_FATAL); } __archive_read_consume(a, 7); break; case MAIN_HEAD: rar->main_flags = archive_le16dec(p + 3); skip = archive_le16dec(p + 5); if (skip < 7 + sizeof(rar->reserved1) + sizeof(rar->reserved2)) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid header size"); return (ARCHIVE_FATAL); } if ((h = __archive_read_ahead(a, skip, NULL)) == NULL) return (ARCHIVE_FATAL); p = h; memcpy(rar->reserved1, p + 7, sizeof(rar->reserved1)); memcpy(rar->reserved2, p + 7 + sizeof(rar->reserved1), sizeof(rar->reserved2)); if (rar->main_flags & MHD_ENCRYPTVER) { if (skip < 7 + sizeof(rar->reserved1) + sizeof(rar->reserved2)+1) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid header size"); return (ARCHIVE_FATAL); } rar->encryptver = *(p + 7 + sizeof(rar->reserved1) + sizeof(rar->reserved2)); } /* Main header is password encrypted, so we cannot read any file names or any other info about files from the header. */ if (rar->main_flags & MHD_PASSWORD) { archive_entry_set_is_metadata_encrypted(entry, 1); archive_entry_set_is_data_encrypted(entry, 1); rar->has_encrypted_entries = 1; archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "RAR encryption support unavailable."); return (ARCHIVE_FATAL); } crc32_val = crc32(0, (const unsigned char *)p + 2, (unsigned)skip - 2); if ((crc32_val & 0xffff) != archive_le16dec(p)) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Header CRC error"); return (ARCHIVE_FATAL); } __archive_read_consume(a, skip); break; case FILE_HEAD: return read_header(a, entry, head_type); case COMM_HEAD: case AV_HEAD: case SUB_HEAD: case PROTECT_HEAD: case SIGN_HEAD: case ENDARC_HEAD: flags = archive_le16dec(p + 3); skip = archive_le16dec(p + 5); if (skip < 7) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid header size too small"); return (ARCHIVE_FATAL); } if (flags & HD_ADD_SIZE_PRESENT) { if (skip < 7 + 4) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid header size too small"); return (ARCHIVE_FATAL); } if ((h = __archive_read_ahead(a, skip, NULL)) == NULL) return (ARCHIVE_FATAL); p = h; skip += archive_le32dec(p + 7); } /* Skip over the 2-byte CRC at the beginning of the header. */ crc32_expected = archive_le16dec(p); __archive_read_consume(a, 2); skip -= 2; /* Skim the entire header and compute the CRC. */ crc32_val = 0; while (skip > 0) { size_t to_read = skip; ssize_t did_read; if (to_read > 32 * 1024) { to_read = 32 * 1024; } if ((h = __archive_read_ahead(a, to_read, &did_read)) == NULL) { return (ARCHIVE_FATAL); } p = h; crc32_val = crc32(crc32_val, (const unsigned char *)p, (unsigned)did_read); __archive_read_consume(a, did_read); skip -= did_read; } if ((crc32_val & 0xffff) != crc32_expected) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Header CRC error"); return (ARCHIVE_FATAL); } if (head_type == ENDARC_HEAD) return (ARCHIVE_EOF); break; case NEWSUB_HEAD: if ((ret = read_header(a, entry, head_type)) < ARCHIVE_WARN) return ret; break; default: archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Bad RAR file"); return (ARCHIVE_FATAL); } } } static int archive_read_format_rar_read_data(struct archive_read *a, const void **buff, size_t *size, int64_t *offset) { struct rar *rar = (struct rar *)(a->format->data); int ret; if (rar->has_encrypted_entries == ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW) { rar->has_encrypted_entries = 0; } if (rar->bytes_unconsumed > 0) { /* Consume as much as the decompressor actually used. */ __archive_read_consume(a, rar->bytes_unconsumed); rar->bytes_unconsumed = 0; } *buff = NULL; if (rar->entry_eof || rar->offset_seek >= rar->unp_size) { *size = 0; *offset = rar->offset; if (*offset < rar->unp_size) *offset = rar->unp_size; return (ARCHIVE_EOF); } switch (rar->compression_method) { case COMPRESS_METHOD_STORE: ret = read_data_stored(a, buff, size, offset); break; case COMPRESS_METHOD_FASTEST: case COMPRESS_METHOD_FAST: case COMPRESS_METHOD_NORMAL: case COMPRESS_METHOD_GOOD: case COMPRESS_METHOD_BEST: ret = read_data_compressed(a, buff, size, offset); if (ret != ARCHIVE_OK && ret != ARCHIVE_WARN) __archive_ppmd7_functions.Ppmd7_Free(&rar->ppmd7_context, &g_szalloc); break; default: archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Unsupported compression method for RAR file."); ret = ARCHIVE_FATAL; break; } return (ret); } static int archive_read_format_rar_read_data_skip(struct archive_read *a) { struct rar *rar; int64_t bytes_skipped; int ret; rar = (struct rar *)(a->format->data); if (rar->bytes_unconsumed > 0) { /* Consume as much as the decompressor actually used. */ __archive_read_consume(a, rar->bytes_unconsumed); rar->bytes_unconsumed = 0; } if (rar->bytes_remaining > 0) { bytes_skipped = __archive_read_consume(a, rar->bytes_remaining); if (bytes_skipped < 0) return (ARCHIVE_FATAL); } /* Compressed data to skip must be read from each header in a multivolume * archive. */ if (rar->main_flags & MHD_VOLUME && rar->file_flags & FHD_SPLIT_AFTER) { ret = archive_read_format_rar_read_header(a, a->entry); if (ret == (ARCHIVE_EOF)) ret = archive_read_format_rar_read_header(a, a->entry); if (ret != (ARCHIVE_OK)) return ret; return archive_read_format_rar_read_data_skip(a); } return (ARCHIVE_OK); } static int64_t archive_read_format_rar_seek_data(struct archive_read *a, int64_t offset, int whence) { int64_t client_offset, ret; unsigned int i; struct rar *rar = (struct rar *)(a->format->data); if (rar->compression_method == COMPRESS_METHOD_STORE) { /* Modify the offset for use with SEEK_SET */ switch (whence) { case SEEK_CUR: client_offset = rar->offset_seek; break; case SEEK_END: client_offset = rar->unp_size; break; case SEEK_SET: default: client_offset = 0; } client_offset += offset; if (client_offset < 0) { /* Can't seek past beginning of data block */ return -1; } else if (client_offset > rar->unp_size) { /* * Set the returned offset but only seek to the end of * the data block. */ rar->offset_seek = client_offset; client_offset = rar->unp_size; } client_offset += rar->dbo[0].start_offset; i = 0; while (i < rar->cursor) { i++; client_offset += rar->dbo[i].start_offset - rar->dbo[i-1].end_offset; } if (rar->main_flags & MHD_VOLUME) { /* Find the appropriate offset among the multivolume archive */ while (1) { if (client_offset < rar->dbo[rar->cursor].start_offset && rar->file_flags & FHD_SPLIT_BEFORE) { /* Search backwards for the correct data block */ if (rar->cursor == 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Attempt to seek past beginning of RAR data block"); return (ARCHIVE_FAILED); } rar->cursor--; client_offset -= rar->dbo[rar->cursor+1].start_offset - rar->dbo[rar->cursor].end_offset; if (client_offset < rar->dbo[rar->cursor].start_offset) continue; ret = __archive_read_seek(a, rar->dbo[rar->cursor].start_offset - rar->dbo[rar->cursor].header_size, SEEK_SET); if (ret < (ARCHIVE_OK)) return ret; ret = archive_read_format_rar_read_header(a, a->entry); if (ret != (ARCHIVE_OK)) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Error during seek of RAR file"); return (ARCHIVE_FAILED); } rar->cursor--; break; } else if (client_offset > rar->dbo[rar->cursor].end_offset && rar->file_flags & FHD_SPLIT_AFTER) { /* Search forward for the correct data block */ rar->cursor++; if (rar->cursor < rar->nodes && client_offset > rar->dbo[rar->cursor].end_offset) { client_offset += rar->dbo[rar->cursor].start_offset - rar->dbo[rar->cursor-1].end_offset; continue; } rar->cursor--; ret = __archive_read_seek(a, rar->dbo[rar->cursor].end_offset, SEEK_SET); if (ret < (ARCHIVE_OK)) return ret; ret = archive_read_format_rar_read_header(a, a->entry); if (ret == (ARCHIVE_EOF)) { rar->has_endarc_header = 1; ret = archive_read_format_rar_read_header(a, a->entry); } if (ret != (ARCHIVE_OK)) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Error during seek of RAR file"); return (ARCHIVE_FAILED); } client_offset += rar->dbo[rar->cursor].start_offset - rar->dbo[rar->cursor-1].end_offset; continue; } break; } } ret = __archive_read_seek(a, client_offset, SEEK_SET); if (ret < (ARCHIVE_OK)) return ret; rar->bytes_remaining = rar->dbo[rar->cursor].end_offset - ret; i = rar->cursor; while (i > 0) { i--; ret -= rar->dbo[i+1].start_offset - rar->dbo[i].end_offset; } ret -= rar->dbo[0].start_offset; /* Always restart reading the file after a seek */ __archive_reset_read_data(&a->archive); rar->bytes_unconsumed = 0; rar->offset = 0; /* * If a seek past the end of file was requested, return the requested * offset. */ if (ret == rar->unp_size && rar->offset_seek > rar->unp_size) return rar->offset_seek; /* Return the new offset */ rar->offset_seek = ret; return rar->offset_seek; } else { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Seeking of compressed RAR files is unsupported"); } return (ARCHIVE_FAILED); } static int archive_read_format_rar_cleanup(struct archive_read *a) { struct rar *rar; rar = (struct rar *)(a->format->data); free_codes(a); free(rar->filename); free(rar->filename_save); free(rar->dbo); free(rar->unp_buffer); free(rar->lzss.window); __archive_ppmd7_functions.Ppmd7_Free(&rar->ppmd7_context, &g_szalloc); free(rar); (a->format->data) = NULL; return (ARCHIVE_OK); } static int read_header(struct archive_read *a, struct archive_entry *entry, char head_type) { const void *h; const char *p, *endp; struct rar *rar; struct rar_header rar_header; struct rar_file_header file_header; int64_t header_size; unsigned filename_size, end; char *filename; char *strp; char packed_size[8]; char unp_size[8]; int ttime; struct archive_string_conv *sconv, *fn_sconv; unsigned long crc32_val; int ret = (ARCHIVE_OK), ret2; rar = (struct rar *)(a->format->data); /* Setup a string conversion object for non-rar-unicode filenames. */ sconv = rar->opt_sconv; if (sconv == NULL) { if (!rar->init_default_conversion) { rar->sconv_default = archive_string_default_conversion_for_read( &(a->archive)); rar->init_default_conversion = 1; } sconv = rar->sconv_default; } if ((h = __archive_read_ahead(a, 7, NULL)) == NULL) return (ARCHIVE_FATAL); p = h; memcpy(&rar_header, p, sizeof(rar_header)); rar->file_flags = archive_le16dec(rar_header.flags); header_size = archive_le16dec(rar_header.size); if (header_size < (int64_t)sizeof(file_header) + 7) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid header size"); return (ARCHIVE_FATAL); } crc32_val = crc32(0, (const unsigned char *)p + 2, 7 - 2); __archive_read_consume(a, 7); if (!(rar->file_flags & FHD_SOLID)) { rar->compression_method = 0; rar->packed_size = 0; rar->unp_size = 0; rar->mtime = 0; rar->ctime = 0; rar->atime = 0; rar->arctime = 0; rar->mode = 0; memset(&rar->salt, 0, sizeof(rar->salt)); rar->atime = 0; rar->ansec = 0; rar->ctime = 0; rar->cnsec = 0; rar->mtime = 0; rar->mnsec = 0; rar->arctime = 0; rar->arcnsec = 0; } else { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "RAR solid archive support unavailable."); return (ARCHIVE_FATAL); } if ((h = __archive_read_ahead(a, (size_t)header_size - 7, NULL)) == NULL) return (ARCHIVE_FATAL); /* File Header CRC check. */ crc32_val = crc32(crc32_val, h, (unsigned)(header_size - 7)); if ((crc32_val & 0xffff) != archive_le16dec(rar_header.crc)) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Header CRC error"); return (ARCHIVE_FATAL); } /* If no CRC error, Go on parsing File Header. */ p = h; endp = p + header_size - 7; memcpy(&file_header, p, sizeof(file_header)); p += sizeof(file_header); rar->compression_method = file_header.method; ttime = archive_le32dec(file_header.file_time); rar->mtime = get_time(ttime); rar->file_crc = archive_le32dec(file_header.file_crc); if (rar->file_flags & FHD_PASSWORD) { archive_entry_set_is_data_encrypted(entry, 1); rar->has_encrypted_entries = 1; archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "RAR encryption support unavailable."); /* Since it is only the data part itself that is encrypted we can at least extract information about the currently processed entry and don't need to return ARCHIVE_FATAL here. */ /*return (ARCHIVE_FATAL);*/ } if (rar->file_flags & FHD_LARGE) { memcpy(packed_size, file_header.pack_size, 4); memcpy(packed_size + 4, p, 4); /* High pack size */ p += 4; memcpy(unp_size, file_header.unp_size, 4); memcpy(unp_size + 4, p, 4); /* High unpack size */ p += 4; rar->packed_size = archive_le64dec(&packed_size); rar->unp_size = archive_le64dec(&unp_size); } else { rar->packed_size = archive_le32dec(file_header.pack_size); rar->unp_size = archive_le32dec(file_header.unp_size); } if (rar->packed_size < 0 || rar->unp_size < 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid sizes specified."); return (ARCHIVE_FATAL); } rar->bytes_remaining = rar->packed_size; /* TODO: RARv3 subblocks contain comments. For now the complete block is * consumed at the end. */ if (head_type == NEWSUB_HEAD) { size_t distance = p - (const char *)h; header_size += rar->packed_size; /* Make sure we have the extended data. */ if ((h = __archive_read_ahead(a, (size_t)header_size - 7, NULL)) == NULL) return (ARCHIVE_FATAL); p = h; endp = p + header_size - 7; p += distance; } filename_size = archive_le16dec(file_header.name_size); if (p + filename_size > endp) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid filename size"); return (ARCHIVE_FATAL); } if (rar->filename_allocated < filename_size * 2 + 2) { char *newptr; size_t newsize = filename_size * 2 + 2; newptr = realloc(rar->filename, newsize); if (newptr == NULL) { archive_set_error(&a->archive, ENOMEM, "Couldn't allocate memory."); return (ARCHIVE_FATAL); } rar->filename = newptr; rar->filename_allocated = newsize; } filename = rar->filename; memcpy(filename, p, filename_size); filename[filename_size] = '\0'; if (rar->file_flags & FHD_UNICODE) { if (filename_size != strlen(filename)) { unsigned char highbyte, flagbits, flagbyte; unsigned fn_end, offset; end = filename_size; fn_end = filename_size * 2; filename_size = 0; offset = (unsigned)strlen(filename) + 1; highbyte = *(p + offset++); flagbits = 0; flagbyte = 0; while (offset < end && filename_size < fn_end) { if (!flagbits) { flagbyte = *(p + offset++); flagbits = 8; } flagbits -= 2; switch((flagbyte >> flagbits) & 3) { case 0: filename[filename_size++] = '\0'; filename[filename_size++] = *(p + offset++); break; case 1: filename[filename_size++] = highbyte; filename[filename_size++] = *(p + offset++); break; case 2: filename[filename_size++] = *(p + offset + 1); filename[filename_size++] = *(p + offset); offset += 2; break; case 3: { char extra, high; uint8_t length = *(p + offset++); if (length & 0x80) { extra = *(p + offset++); high = (char)highbyte; } else extra = high = 0; length = (length & 0x7f) + 2; while (length && filename_size < fn_end) { unsigned cp = filename_size >> 1; filename[filename_size++] = high; filename[filename_size++] = p[cp] + extra; length--; } } break; } } if (filename_size > fn_end) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid filename"); return (ARCHIVE_FATAL); } filename[filename_size++] = '\0'; - filename[filename_size++] = '\0'; + /* + * Do not increment filename_size here as the computations below + * add the space for the terminating NUL explicitly. + */ + filename[filename_size] = '\0'; /* Decoded unicode form is UTF-16BE, so we have to update a string * conversion object for it. */ if (rar->sconv_utf16be == NULL) { rar->sconv_utf16be = archive_string_conversion_from_charset( &a->archive, "UTF-16BE", 1); if (rar->sconv_utf16be == NULL) return (ARCHIVE_FATAL); } fn_sconv = rar->sconv_utf16be; strp = filename; while (memcmp(strp, "\x00\x00", 2)) { if (!memcmp(strp, "\x00\\", 2)) *(strp + 1) = '/'; strp += 2; } p += offset; } else { /* * If FHD_UNICODE is set but no unicode data, this file name form * is UTF-8, so we have to update a string conversion object for * it accordingly. */ if (rar->sconv_utf8 == NULL) { rar->sconv_utf8 = archive_string_conversion_from_charset( &a->archive, "UTF-8", 1); if (rar->sconv_utf8 == NULL) return (ARCHIVE_FATAL); } fn_sconv = rar->sconv_utf8; while ((strp = strchr(filename, '\\')) != NULL) *strp = '/'; p += filename_size; } } else { fn_sconv = sconv; while ((strp = strchr(filename, '\\')) != NULL) *strp = '/'; p += filename_size; } /* Split file in multivolume RAR. No more need to process header. */ if (rar->filename_save && filename_size == rar->filename_save_size && !memcmp(rar->filename, rar->filename_save, filename_size + 1)) { __archive_read_consume(a, header_size - 7); rar->cursor++; if (rar->cursor >= rar->nodes) { rar->nodes++; if ((rar->dbo = realloc(rar->dbo, sizeof(*rar->dbo) * rar->nodes)) == NULL) { archive_set_error(&a->archive, ENOMEM, "Couldn't allocate memory."); return (ARCHIVE_FATAL); } rar->dbo[rar->cursor].header_size = header_size; rar->dbo[rar->cursor].start_offset = -1; rar->dbo[rar->cursor].end_offset = -1; } if (rar->dbo[rar->cursor].start_offset < 0) { rar->dbo[rar->cursor].start_offset = a->filter->position; rar->dbo[rar->cursor].end_offset = rar->dbo[rar->cursor].start_offset + rar->packed_size; } return ret; } rar->filename_save = (char*)realloc(rar->filename_save, filename_size + 1); memcpy(rar->filename_save, rar->filename, filename_size + 1); rar->filename_save_size = filename_size; /* Set info for seeking */ free(rar->dbo); if ((rar->dbo = calloc(1, sizeof(*rar->dbo))) == NULL) { archive_set_error(&a->archive, ENOMEM, "Couldn't allocate memory."); return (ARCHIVE_FATAL); } rar->dbo[0].header_size = header_size; rar->dbo[0].start_offset = -1; rar->dbo[0].end_offset = -1; rar->cursor = 0; rar->nodes = 1; if (rar->file_flags & FHD_SALT) { if (p + 8 > endp) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid header size"); return (ARCHIVE_FATAL); } memcpy(rar->salt, p, 8); p += 8; } if (rar->file_flags & FHD_EXTTIME) { if (read_exttime(p, rar, endp) < 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid header size"); return (ARCHIVE_FATAL); } } __archive_read_consume(a, header_size - 7); rar->dbo[0].start_offset = a->filter->position; rar->dbo[0].end_offset = rar->dbo[0].start_offset + rar->packed_size; switch(file_header.host_os) { case OS_MSDOS: case OS_OS2: case OS_WIN32: rar->mode = archive_le32dec(file_header.file_attr); if (rar->mode & FILE_ATTRIBUTE_DIRECTORY) rar->mode = AE_IFDIR | S_IXUSR | S_IXGRP | S_IXOTH; else rar->mode = AE_IFREG; rar->mode |= S_IRUSR | S_IWUSR | S_IRGRP | S_IROTH; break; case OS_UNIX: case OS_MAC_OS: case OS_BEOS: rar->mode = archive_le32dec(file_header.file_attr); break; default: archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Unknown file attributes from RAR file's host OS"); return (ARCHIVE_FATAL); } rar->bytes_uncopied = rar->bytes_unconsumed = 0; rar->lzss.position = rar->offset = 0; rar->offset_seek = 0; rar->dictionary_size = 0; rar->offset_outgoing = 0; rar->br.cache_avail = 0; rar->br.avail_in = 0; rar->crc_calculated = 0; rar->entry_eof = 0; rar->valid = 1; rar->is_ppmd_block = 0; rar->start_new_table = 1; free(rar->unp_buffer); rar->unp_buffer = NULL; rar->unp_offset = 0; rar->unp_buffer_size = UNP_BUFFER_SIZE; memset(rar->lengthtable, 0, sizeof(rar->lengthtable)); __archive_ppmd7_functions.Ppmd7_Free(&rar->ppmd7_context, &g_szalloc); rar->ppmd_valid = rar->ppmd_eod = 0; /* Don't set any archive entries for non-file header types */ if (head_type == NEWSUB_HEAD) return ret; archive_entry_set_mtime(entry, rar->mtime, rar->mnsec); archive_entry_set_ctime(entry, rar->ctime, rar->cnsec); archive_entry_set_atime(entry, rar->atime, rar->ansec); archive_entry_set_size(entry, rar->unp_size); archive_entry_set_mode(entry, rar->mode); if (archive_entry_copy_pathname_l(entry, filename, filename_size, fn_sconv)) { if (errno == ENOMEM) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory for Pathname"); return (ARCHIVE_FATAL); } archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Pathname cannot be converted from %s to current locale.", archive_string_conversion_charset_name(fn_sconv)); ret = (ARCHIVE_WARN); } if (((rar->mode) & AE_IFMT) == AE_IFLNK) { /* Make sure a symbolic-link file does not have its body. */ rar->bytes_remaining = 0; archive_entry_set_size(entry, 0); /* Read a symbolic-link name. */ if ((ret2 = read_symlink_stored(a, entry, sconv)) < (ARCHIVE_WARN)) return ret2; if (ret > ret2) ret = ret2; } if (rar->bytes_remaining == 0) rar->entry_eof = 1; return ret; } static time_t get_time(int ttime) { struct tm tm; tm.tm_sec = 2 * (ttime & 0x1f); tm.tm_min = (ttime >> 5) & 0x3f; tm.tm_hour = (ttime >> 11) & 0x1f; tm.tm_mday = (ttime >> 16) & 0x1f; tm.tm_mon = ((ttime >> 21) & 0x0f) - 1; tm.tm_year = ((ttime >> 25) & 0x7f) + 80; tm.tm_isdst = -1; return mktime(&tm); } static int read_exttime(const char *p, struct rar *rar, const char *endp) { unsigned rmode, flags, rem, j, count; int ttime, i; struct tm *tm; time_t t; long nsec; if (p + 2 > endp) return (-1); flags = archive_le16dec(p); p += 2; for (i = 3; i >= 0; i--) { t = 0; if (i == 3) t = rar->mtime; rmode = flags >> i * 4; if (rmode & 8) { if (!t) { if (p + 4 > endp) return (-1); ttime = archive_le32dec(p); t = get_time(ttime); p += 4; } rem = 0; count = rmode & 3; if (p + count > endp) return (-1); for (j = 0; j < count; j++) { rem = (((unsigned)(unsigned char)*p) << 16) | (rem >> 8); p++; } tm = localtime(&t); nsec = tm->tm_sec + rem / NS_UNIT; if (rmode & 4) { tm->tm_sec++; t = mktime(tm); } if (i == 3) { rar->mtime = t; rar->mnsec = nsec; } else if (i == 2) { rar->ctime = t; rar->cnsec = nsec; } else if (i == 1) { rar->atime = t; rar->ansec = nsec; } else { rar->arctime = t; rar->arcnsec = nsec; } } } return (0); } static int read_symlink_stored(struct archive_read *a, struct archive_entry *entry, struct archive_string_conv *sconv) { const void *h; const char *p; struct rar *rar; int ret = (ARCHIVE_OK); rar = (struct rar *)(a->format->data); if ((h = rar_read_ahead(a, (size_t)rar->packed_size, NULL)) == NULL) return (ARCHIVE_FATAL); p = h; if (archive_entry_copy_symlink_l(entry, p, (size_t)rar->packed_size, sconv)) { if (errno == ENOMEM) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory for link"); return (ARCHIVE_FATAL); } archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "link cannot be converted from %s to current locale.", archive_string_conversion_charset_name(sconv)); ret = (ARCHIVE_WARN); } __archive_read_consume(a, rar->packed_size); return ret; } static int read_data_stored(struct archive_read *a, const void **buff, size_t *size, int64_t *offset) { struct rar *rar; ssize_t bytes_avail; rar = (struct rar *)(a->format->data); if (rar->bytes_remaining == 0 && !(rar->main_flags & MHD_VOLUME && rar->file_flags & FHD_SPLIT_AFTER)) { *buff = NULL; *size = 0; *offset = rar->offset; if (rar->file_crc != rar->crc_calculated) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "File CRC error"); return (ARCHIVE_FATAL); } rar->entry_eof = 1; return (ARCHIVE_EOF); } *buff = rar_read_ahead(a, 1, &bytes_avail); if (bytes_avail <= 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated RAR file data"); return (ARCHIVE_FATAL); } *size = bytes_avail; *offset = rar->offset; rar->offset += bytes_avail; rar->offset_seek += bytes_avail; rar->bytes_remaining -= bytes_avail; rar->bytes_unconsumed = bytes_avail; /* Calculate File CRC. */ rar->crc_calculated = crc32(rar->crc_calculated, *buff, (unsigned)bytes_avail); return (ARCHIVE_OK); } static int read_data_compressed(struct archive_read *a, const void **buff, size_t *size, int64_t *offset) { struct rar *rar; int64_t start, end, actualend; size_t bs; int ret = (ARCHIVE_OK), sym, code, lzss_offset, length, i; rar = (struct rar *)(a->format->data); do { if (!rar->valid) return (ARCHIVE_FATAL); if (rar->ppmd_eod || (rar->dictionary_size && rar->offset >= rar->unp_size)) { if (rar->unp_offset > 0) { /* * We have unprocessed extracted data. write it out. */ *buff = rar->unp_buffer; *size = rar->unp_offset; *offset = rar->offset_outgoing; rar->offset_outgoing += *size; /* Calculate File CRC. */ rar->crc_calculated = crc32(rar->crc_calculated, *buff, (unsigned)*size); rar->unp_offset = 0; return (ARCHIVE_OK); } *buff = NULL; *size = 0; *offset = rar->offset; if (rar->file_crc != rar->crc_calculated) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "File CRC error"); return (ARCHIVE_FATAL); } rar->entry_eof = 1; return (ARCHIVE_EOF); } if (!rar->is_ppmd_block && rar->dictionary_size && rar->bytes_uncopied > 0) { if (rar->bytes_uncopied > (rar->unp_buffer_size - rar->unp_offset)) bs = rar->unp_buffer_size - rar->unp_offset; else bs = (size_t)rar->bytes_uncopied; ret = copy_from_lzss_window(a, buff, rar->offset, (int)bs); if (ret != ARCHIVE_OK) return (ret); rar->offset += bs; rar->bytes_uncopied -= bs; if (*buff != NULL) { rar->unp_offset = 0; *size = rar->unp_buffer_size; *offset = rar->offset_outgoing; rar->offset_outgoing += *size; /* Calculate File CRC. */ rar->crc_calculated = crc32(rar->crc_calculated, *buff, (unsigned)*size); return (ret); } continue; } if (!rar->br.next_in && (ret = rar_br_preparation(a, &(rar->br))) < ARCHIVE_WARN) return (ret); if (rar->start_new_table && ((ret = parse_codes(a)) < (ARCHIVE_WARN))) return (ret); if (rar->is_ppmd_block) { if ((sym = __archive_ppmd7_functions.Ppmd7_DecodeSymbol( &rar->ppmd7_context, &rar->range_dec.p)) < 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid symbol"); return (ARCHIVE_FATAL); } if(sym != rar->ppmd_escape) { lzss_emit_literal(rar, sym); rar->bytes_uncopied++; } else { if ((code = __archive_ppmd7_functions.Ppmd7_DecodeSymbol( &rar->ppmd7_context, &rar->range_dec.p)) < 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid symbol"); return (ARCHIVE_FATAL); } switch(code) { case 0: rar->start_new_table = 1; return read_data_compressed(a, buff, size, offset); case 2: rar->ppmd_eod = 1;/* End Of ppmd Data. */ continue; case 3: archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Parsing filters is unsupported."); return (ARCHIVE_FAILED); case 4: lzss_offset = 0; for (i = 2; i >= 0; i--) { if ((code = __archive_ppmd7_functions.Ppmd7_DecodeSymbol( &rar->ppmd7_context, &rar->range_dec.p)) < 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid symbol"); return (ARCHIVE_FATAL); } lzss_offset |= code << (i * 8); } if ((length = __archive_ppmd7_functions.Ppmd7_DecodeSymbol( &rar->ppmd7_context, &rar->range_dec.p)) < 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid symbol"); return (ARCHIVE_FATAL); } lzss_emit_match(rar, lzss_offset + 2, length + 32); rar->bytes_uncopied += length + 32; break; case 5: if ((length = __archive_ppmd7_functions.Ppmd7_DecodeSymbol( &rar->ppmd7_context, &rar->range_dec.p)) < 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid symbol"); return (ARCHIVE_FATAL); } lzss_emit_match(rar, 1, length + 4); rar->bytes_uncopied += length + 4; break; default: lzss_emit_literal(rar, sym); rar->bytes_uncopied++; } } } else { start = rar->offset; end = start + rar->dictionary_size; rar->filterstart = INT64_MAX; if ((actualend = expand(a, end)) < 0) return ((int)actualend); rar->bytes_uncopied = actualend - start; if (rar->bytes_uncopied == 0) { /* Broken RAR files cause this case. * NOTE: If this case were possible on a normal RAR file * we would find out where it was actually bad and * what we would do to solve it. */ archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Internal error extracting RAR file"); return (ARCHIVE_FATAL); } } if (rar->bytes_uncopied > (rar->unp_buffer_size - rar->unp_offset)) bs = rar->unp_buffer_size - rar->unp_offset; else bs = (size_t)rar->bytes_uncopied; ret = copy_from_lzss_window(a, buff, rar->offset, (int)bs); if (ret != ARCHIVE_OK) return (ret); rar->offset += bs; rar->bytes_uncopied -= bs; /* * If *buff is NULL, it means unp_buffer is not full. * So we have to continue extracting a RAR file. */ } while (*buff == NULL); rar->unp_offset = 0; *size = rar->unp_buffer_size; *offset = rar->offset_outgoing; rar->offset_outgoing += *size; /* Calculate File CRC. */ rar->crc_calculated = crc32(rar->crc_calculated, *buff, (unsigned)*size); return ret; } static int parse_codes(struct archive_read *a) { int i, j, val, n, r; unsigned char bitlengths[MAX_SYMBOLS], zerocount, ppmd_flags; unsigned int maxorder; struct huffman_code precode; struct rar *rar = (struct rar *)(a->format->data); struct rar_br *br = &(rar->br); free_codes(a); /* Skip to the next byte */ rar_br_consume_unalined_bits(br); /* PPMd block flag */ if (!rar_br_read_ahead(a, br, 1)) goto truncated_data; if ((rar->is_ppmd_block = rar_br_bits(br, 1)) != 0) { rar_br_consume(br, 1); if (!rar_br_read_ahead(a, br, 7)) goto truncated_data; ppmd_flags = rar_br_bits(br, 7); rar_br_consume(br, 7); /* Memory is allocated in MB */ if (ppmd_flags & 0x20) { if (!rar_br_read_ahead(a, br, 8)) goto truncated_data; rar->dictionary_size = (rar_br_bits(br, 8) + 1) << 20; rar_br_consume(br, 8); } if (ppmd_flags & 0x40) { if (!rar_br_read_ahead(a, br, 8)) goto truncated_data; rar->ppmd_escape = rar->ppmd7_context.InitEsc = rar_br_bits(br, 8); rar_br_consume(br, 8); } else rar->ppmd_escape = 2; if (ppmd_flags & 0x20) { maxorder = (ppmd_flags & 0x1F) + 1; if(maxorder > 16) maxorder = 16 + (maxorder - 16) * 3; if (maxorder == 1) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated RAR file data"); return (ARCHIVE_FATAL); } /* Make sure ppmd7_contest is freed before Ppmd7_Construct * because reading a broken file cause this abnormal sequence. */ __archive_ppmd7_functions.Ppmd7_Free(&rar->ppmd7_context, &g_szalloc); rar->bytein.a = a; rar->bytein.Read = &ppmd_read; __archive_ppmd7_functions.PpmdRAR_RangeDec_CreateVTable(&rar->range_dec); rar->range_dec.Stream = &rar->bytein; __archive_ppmd7_functions.Ppmd7_Construct(&rar->ppmd7_context); if (rar->dictionary_size == 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid zero dictionary size"); return (ARCHIVE_FATAL); } if (!__archive_ppmd7_functions.Ppmd7_Alloc(&rar->ppmd7_context, rar->dictionary_size, &g_szalloc)) { archive_set_error(&a->archive, ENOMEM, "Out of memory"); return (ARCHIVE_FATAL); } if (!__archive_ppmd7_functions.PpmdRAR_RangeDec_Init(&rar->range_dec)) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Unable to initialize PPMd range decoder"); return (ARCHIVE_FATAL); } __archive_ppmd7_functions.Ppmd7_Init(&rar->ppmd7_context, maxorder); rar->ppmd_valid = 1; } else { if (!rar->ppmd_valid) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid PPMd sequence"); return (ARCHIVE_FATAL); } if (!__archive_ppmd7_functions.PpmdRAR_RangeDec_Init(&rar->range_dec)) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Unable to initialize PPMd range decoder"); return (ARCHIVE_FATAL); } } } else { rar_br_consume(br, 1); /* Keep existing table flag */ if (!rar_br_read_ahead(a, br, 1)) goto truncated_data; if (!rar_br_bits(br, 1)) memset(rar->lengthtable, 0, sizeof(rar->lengthtable)); rar_br_consume(br, 1); memset(&bitlengths, 0, sizeof(bitlengths)); for (i = 0; i < MAX_SYMBOLS;) { if (!rar_br_read_ahead(a, br, 4)) goto truncated_data; bitlengths[i++] = rar_br_bits(br, 4); rar_br_consume(br, 4); if (bitlengths[i-1] == 0xF) { if (!rar_br_read_ahead(a, br, 4)) goto truncated_data; zerocount = rar_br_bits(br, 4); rar_br_consume(br, 4); if (zerocount) { i--; for (j = 0; j < zerocount + 2 && i < MAX_SYMBOLS; j++) bitlengths[i++] = 0; } } } memset(&precode, 0, sizeof(precode)); r = create_code(a, &precode, bitlengths, MAX_SYMBOLS, MAX_SYMBOL_LENGTH); if (r != ARCHIVE_OK) { free(precode.tree); free(precode.table); return (r); } for (i = 0; i < HUFFMAN_TABLE_SIZE;) { if ((val = read_next_symbol(a, &precode)) < 0) { free(precode.tree); free(precode.table); return (ARCHIVE_FATAL); } if (val < 16) { rar->lengthtable[i] = (rar->lengthtable[i] + val) & 0xF; i++; } else if (val < 18) { if (i == 0) { free(precode.tree); free(precode.table); archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Internal error extracting RAR file."); return (ARCHIVE_FATAL); } if(val == 16) { if (!rar_br_read_ahead(a, br, 3)) { free(precode.tree); free(precode.table); goto truncated_data; } n = rar_br_bits(br, 3) + 3; rar_br_consume(br, 3); } else { if (!rar_br_read_ahead(a, br, 7)) { free(precode.tree); free(precode.table); goto truncated_data; } n = rar_br_bits(br, 7) + 11; rar_br_consume(br, 7); } for (j = 0; j < n && i < HUFFMAN_TABLE_SIZE; j++) { rar->lengthtable[i] = rar->lengthtable[i-1]; i++; } } else { if(val == 18) { if (!rar_br_read_ahead(a, br, 3)) { free(precode.tree); free(precode.table); goto truncated_data; } n = rar_br_bits(br, 3) + 3; rar_br_consume(br, 3); } else { if (!rar_br_read_ahead(a, br, 7)) { free(precode.tree); free(precode.table); goto truncated_data; } n = rar_br_bits(br, 7) + 11; rar_br_consume(br, 7); } for(j = 0; j < n && i < HUFFMAN_TABLE_SIZE; j++) rar->lengthtable[i++] = 0; } } free(precode.tree); free(precode.table); r = create_code(a, &rar->maincode, &rar->lengthtable[0], MAINCODE_SIZE, MAX_SYMBOL_LENGTH); if (r != ARCHIVE_OK) return (r); r = create_code(a, &rar->offsetcode, &rar->lengthtable[MAINCODE_SIZE], OFFSETCODE_SIZE, MAX_SYMBOL_LENGTH); if (r != ARCHIVE_OK) return (r); r = create_code(a, &rar->lowoffsetcode, &rar->lengthtable[MAINCODE_SIZE + OFFSETCODE_SIZE], LOWOFFSETCODE_SIZE, MAX_SYMBOL_LENGTH); if (r != ARCHIVE_OK) return (r); r = create_code(a, &rar->lengthcode, &rar->lengthtable[MAINCODE_SIZE + OFFSETCODE_SIZE + LOWOFFSETCODE_SIZE], LENGTHCODE_SIZE, MAX_SYMBOL_LENGTH); if (r != ARCHIVE_OK) return (r); } if (!rar->dictionary_size || !rar->lzss.window) { /* Seems as though dictionary sizes are not used. Even so, minimize * memory usage as much as possible. */ void *new_window; unsigned int new_size; if (rar->unp_size >= DICTIONARY_MAX_SIZE) new_size = DICTIONARY_MAX_SIZE; else new_size = rar_fls((unsigned int)rar->unp_size) << 1; new_window = realloc(rar->lzss.window, new_size); if (new_window == NULL) { archive_set_error(&a->archive, ENOMEM, "Unable to allocate memory for uncompressed data."); return (ARCHIVE_FATAL); } rar->lzss.window = (unsigned char *)new_window; rar->dictionary_size = new_size; memset(rar->lzss.window, 0, rar->dictionary_size); rar->lzss.mask = rar->dictionary_size - 1; } rar->start_new_table = 0; return (ARCHIVE_OK); truncated_data: archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated RAR file data"); rar->valid = 0; return (ARCHIVE_FATAL); } static void free_codes(struct archive_read *a) { struct rar *rar = (struct rar *)(a->format->data); free(rar->maincode.tree); free(rar->offsetcode.tree); free(rar->lowoffsetcode.tree); free(rar->lengthcode.tree); free(rar->maincode.table); free(rar->offsetcode.table); free(rar->lowoffsetcode.table); free(rar->lengthcode.table); memset(&rar->maincode, 0, sizeof(rar->maincode)); memset(&rar->offsetcode, 0, sizeof(rar->offsetcode)); memset(&rar->lowoffsetcode, 0, sizeof(rar->lowoffsetcode)); memset(&rar->lengthcode, 0, sizeof(rar->lengthcode)); } static int read_next_symbol(struct archive_read *a, struct huffman_code *code) { unsigned char bit; unsigned int bits; int length, value, node; struct rar *rar; struct rar_br *br; if (!code->table) { if (make_table(a, code) != (ARCHIVE_OK)) return -1; } rar = (struct rar *)(a->format->data); br = &(rar->br); /* Look ahead (peek) at bits */ if (!rar_br_read_ahead(a, br, code->tablesize)) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated RAR file data"); rar->valid = 0; return -1; } bits = rar_br_bits(br, code->tablesize); length = code->table[bits].length; value = code->table[bits].value; if (length < 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid prefix code in bitstream"); return -1; } if (length <= code->tablesize) { /* Skip length bits */ rar_br_consume(br, length); return value; } /* Skip tablesize bits */ rar_br_consume(br, code->tablesize); node = value; while (!(code->tree[node].branches[0] == code->tree[node].branches[1])) { if (!rar_br_read_ahead(a, br, 1)) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated RAR file data"); rar->valid = 0; return -1; } bit = rar_br_bits(br, 1); rar_br_consume(br, 1); if (code->tree[node].branches[bit] < 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid prefix code in bitstream"); return -1; } node = code->tree[node].branches[bit]; } return code->tree[node].branches[0]; } static int create_code(struct archive_read *a, struct huffman_code *code, unsigned char *lengths, int numsymbols, char maxlength) { int i, j, codebits = 0, symbolsleft = numsymbols; code->numentries = 0; code->numallocatedentries = 0; if (new_node(code) < 0) { archive_set_error(&a->archive, ENOMEM, "Unable to allocate memory for node data."); return (ARCHIVE_FATAL); } code->numentries = 1; code->minlength = INT_MAX; code->maxlength = INT_MIN; codebits = 0; for(i = 1; i <= maxlength; i++) { for(j = 0; j < numsymbols; j++) { if (lengths[j] != i) continue; if (add_value(a, code, j, codebits, i) != ARCHIVE_OK) return (ARCHIVE_FATAL); codebits++; if (--symbolsleft <= 0) { break; break; } } codebits <<= 1; } return (ARCHIVE_OK); } static int add_value(struct archive_read *a, struct huffman_code *code, int value, int codebits, int length) { int repeatpos, lastnode, bitpos, bit, repeatnode, nextnode; free(code->table); code->table = NULL; if(length > code->maxlength) code->maxlength = length; if(length < code->minlength) code->minlength = length; repeatpos = -1; if (repeatpos == 0 || (repeatpos >= 0 && (((codebits >> (repeatpos - 1)) & 3) == 0 || ((codebits >> (repeatpos - 1)) & 3) == 3))) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid repeat position"); return (ARCHIVE_FATAL); } lastnode = 0; for (bitpos = length - 1; bitpos >= 0; bitpos--) { bit = (codebits >> bitpos) & 1; /* Leaf node check */ if (code->tree[lastnode].branches[0] == code->tree[lastnode].branches[1]) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Prefix found"); return (ARCHIVE_FATAL); } if (bitpos == repeatpos) { /* Open branch check */ if (!(code->tree[lastnode].branches[bit] < 0)) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid repeating code"); return (ARCHIVE_FATAL); } if ((repeatnode = new_node(code)) < 0) { archive_set_error(&a->archive, ENOMEM, "Unable to allocate memory for node data."); return (ARCHIVE_FATAL); } if ((nextnode = new_node(code)) < 0) { archive_set_error(&a->archive, ENOMEM, "Unable to allocate memory for node data."); return (ARCHIVE_FATAL); } /* Set branches */ code->tree[lastnode].branches[bit] = repeatnode; code->tree[repeatnode].branches[bit] = repeatnode; code->tree[repeatnode].branches[bit^1] = nextnode; lastnode = nextnode; bitpos++; /* terminating bit already handled, skip it */ } else { /* Open branch check */ if (code->tree[lastnode].branches[bit] < 0) { if (new_node(code) < 0) { archive_set_error(&a->archive, ENOMEM, "Unable to allocate memory for node data."); return (ARCHIVE_FATAL); } code->tree[lastnode].branches[bit] = code->numentries++; } /* set to branch */ lastnode = code->tree[lastnode].branches[bit]; } } if (!(code->tree[lastnode].branches[0] == -1 && code->tree[lastnode].branches[1] == -2)) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Prefix found"); return (ARCHIVE_FATAL); } /* Set leaf value */ code->tree[lastnode].branches[0] = value; code->tree[lastnode].branches[1] = value; return (ARCHIVE_OK); } static int new_node(struct huffman_code *code) { void *new_tree; if (code->numallocatedentries == code->numentries) { int new_num_entries = 256; if (code->numentries > 0) { new_num_entries = code->numentries * 2; } new_tree = realloc(code->tree, new_num_entries * sizeof(*code->tree)); if (new_tree == NULL) return (-1); code->tree = (struct huffman_tree_node *)new_tree; code->numallocatedentries = new_num_entries; } code->tree[code->numentries].branches[0] = -1; code->tree[code->numentries].branches[1] = -2; return 1; } static int make_table(struct archive_read *a, struct huffman_code *code) { if (code->maxlength < code->minlength || code->maxlength > 10) code->tablesize = 10; else code->tablesize = code->maxlength; code->table = (struct huffman_table_entry *)calloc(1, sizeof(*code->table) * ((size_t)1 << code->tablesize)); return make_table_recurse(a, code, 0, code->table, 0, code->tablesize); } static int make_table_recurse(struct archive_read *a, struct huffman_code *code, int node, struct huffman_table_entry *table, int depth, int maxdepth) { int currtablesize, i, ret = (ARCHIVE_OK); if (!code->tree) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Huffman tree was not created."); return (ARCHIVE_FATAL); } if (node < 0 || node >= code->numentries) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid location to Huffman tree specified."); return (ARCHIVE_FATAL); } currtablesize = 1 << (maxdepth - depth); if (code->tree[node].branches[0] == code->tree[node].branches[1]) { for(i = 0; i < currtablesize; i++) { table[i].length = depth; table[i].value = code->tree[node].branches[0]; } } else if (node < 0) { for(i = 0; i < currtablesize; i++) table[i].length = -1; } else { if(depth == maxdepth) { table[0].length = maxdepth + 1; table[0].value = node; } else { ret |= make_table_recurse(a, code, code->tree[node].branches[0], table, depth + 1, maxdepth); ret |= make_table_recurse(a, code, code->tree[node].branches[1], table + currtablesize / 2, depth + 1, maxdepth); } } return ret; } static int64_t expand(struct archive_read *a, int64_t end) { static const unsigned char lengthbases[] = { 0, 1, 2, 3, 4, 5, 6, 7, 8, 10, 12, 14, 16, 20, 24, 28, 32, 40, 48, 56, 64, 80, 96, 112, 128, 160, 192, 224 }; static const unsigned char lengthbits[] = { 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5 }; static const unsigned int offsetbases[] = { 0, 1, 2, 3, 4, 6, 8, 12, 16, 24, 32, 48, 64, 96, 128, 192, 256, 384, 512, 768, 1024, 1536, 2048, 3072, 4096, 6144, 8192, 12288, 16384, 24576, 32768, 49152, 65536, 98304, 131072, 196608, 262144, 327680, 393216, 458752, 524288, 589824, 655360, 720896, 786432, 851968, 917504, 983040, 1048576, 1310720, 1572864, 1835008, 2097152, 2359296, 2621440, 2883584, 3145728, 3407872, 3670016, 3932160 }; static const unsigned char offsetbits[] = { 0, 0, 0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8, 8, 9, 9, 10, 10, 11, 11, 12, 12, 13, 13, 14, 14, 15, 15, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18 }; static const unsigned char shortbases[] = { 0, 4, 8, 16, 32, 64, 128, 192 }; static const unsigned char shortbits[] = { 2, 2, 3, 4, 5, 6, 6, 6 }; int symbol, offs, len, offsindex, lensymbol, i, offssymbol, lowoffsetsymbol; unsigned char newfile; struct rar *rar = (struct rar *)(a->format->data); struct rar_br *br = &(rar->br); if (rar->filterstart < end) end = rar->filterstart; while (1) { if (rar->output_last_match && lzss_position(&rar->lzss) + rar->lastlength <= end) { lzss_emit_match(rar, rar->lastoffset, rar->lastlength); rar->output_last_match = 0; } if(rar->is_ppmd_block || rar->output_last_match || lzss_position(&rar->lzss) >= end) return lzss_position(&rar->lzss); if ((symbol = read_next_symbol(a, &rar->maincode)) < 0) return (ARCHIVE_FATAL); rar->output_last_match = 0; if (symbol < 256) { lzss_emit_literal(rar, symbol); continue; } else if (symbol == 256) { if (!rar_br_read_ahead(a, br, 1)) goto truncated_data; newfile = !rar_br_bits(br, 1); rar_br_consume(br, 1); if(newfile) { rar->start_new_block = 1; if (!rar_br_read_ahead(a, br, 1)) goto truncated_data; rar->start_new_table = rar_br_bits(br, 1); rar_br_consume(br, 1); return lzss_position(&rar->lzss); } else { if (parse_codes(a) != ARCHIVE_OK) return (ARCHIVE_FATAL); continue; } } else if(symbol==257) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Parsing filters is unsupported."); return (ARCHIVE_FAILED); } else if(symbol==258) { if(rar->lastlength == 0) continue; offs = rar->lastoffset; len = rar->lastlength; } else if (symbol <= 262) { offsindex = symbol - 259; offs = rar->oldoffset[offsindex]; if ((lensymbol = read_next_symbol(a, &rar->lengthcode)) < 0) goto bad_data; if (lensymbol > (int)(sizeof(lengthbases)/sizeof(lengthbases[0]))) goto bad_data; if (lensymbol > (int)(sizeof(lengthbits)/sizeof(lengthbits[0]))) goto bad_data; len = lengthbases[lensymbol] + 2; if (lengthbits[lensymbol] > 0) { if (!rar_br_read_ahead(a, br, lengthbits[lensymbol])) goto truncated_data; len += rar_br_bits(br, lengthbits[lensymbol]); rar_br_consume(br, lengthbits[lensymbol]); } for (i = offsindex; i > 0; i--) rar->oldoffset[i] = rar->oldoffset[i-1]; rar->oldoffset[0] = offs; } else if(symbol<=270) { offs = shortbases[symbol-263] + 1; if(shortbits[symbol-263] > 0) { if (!rar_br_read_ahead(a, br, shortbits[symbol-263])) goto truncated_data; offs += rar_br_bits(br, shortbits[symbol-263]); rar_br_consume(br, shortbits[symbol-263]); } len = 2; for(i = 3; i > 0; i--) rar->oldoffset[i] = rar->oldoffset[i-1]; rar->oldoffset[0] = offs; } else { if (symbol-271 > (int)(sizeof(lengthbases)/sizeof(lengthbases[0]))) goto bad_data; if (symbol-271 > (int)(sizeof(lengthbits)/sizeof(lengthbits[0]))) goto bad_data; len = lengthbases[symbol-271]+3; if(lengthbits[symbol-271] > 0) { if (!rar_br_read_ahead(a, br, lengthbits[symbol-271])) goto truncated_data; len += rar_br_bits(br, lengthbits[symbol-271]); rar_br_consume(br, lengthbits[symbol-271]); } if ((offssymbol = read_next_symbol(a, &rar->offsetcode)) < 0) goto bad_data; if (offssymbol > (int)(sizeof(offsetbases)/sizeof(offsetbases[0]))) goto bad_data; if (offssymbol > (int)(sizeof(offsetbits)/sizeof(offsetbits[0]))) goto bad_data; offs = offsetbases[offssymbol]+1; if(offsetbits[offssymbol] > 0) { if(offssymbol > 9) { if(offsetbits[offssymbol] > 4) { if (!rar_br_read_ahead(a, br, offsetbits[offssymbol] - 4)) goto truncated_data; offs += rar_br_bits(br, offsetbits[offssymbol] - 4) << 4; rar_br_consume(br, offsetbits[offssymbol] - 4); } if(rar->numlowoffsetrepeats > 0) { rar->numlowoffsetrepeats--; offs += rar->lastlowoffset; } else { if ((lowoffsetsymbol = read_next_symbol(a, &rar->lowoffsetcode)) < 0) return (ARCHIVE_FATAL); if(lowoffsetsymbol == 16) { rar->numlowoffsetrepeats = 15; offs += rar->lastlowoffset; } else { offs += lowoffsetsymbol; rar->lastlowoffset = lowoffsetsymbol; } } } else { if (!rar_br_read_ahead(a, br, offsetbits[offssymbol])) goto truncated_data; offs += rar_br_bits(br, offsetbits[offssymbol]); rar_br_consume(br, offsetbits[offssymbol]); } } if (offs >= 0x40000) len++; if (offs >= 0x2000) len++; for(i = 3; i > 0; i--) rar->oldoffset[i] = rar->oldoffset[i-1]; rar->oldoffset[0] = offs; } rar->lastoffset = offs; rar->lastlength = len; rar->output_last_match = 1; } truncated_data: archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated RAR file data"); rar->valid = 0; return (ARCHIVE_FATAL); bad_data: archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Bad RAR file data"); return (ARCHIVE_FATAL); } static int copy_from_lzss_window(struct archive_read *a, const void **buffer, int64_t startpos, int length) { int windowoffs, firstpart; struct rar *rar = (struct rar *)(a->format->data); if (!rar->unp_buffer) { if ((rar->unp_buffer = malloc(rar->unp_buffer_size)) == NULL) { archive_set_error(&a->archive, ENOMEM, "Unable to allocate memory for uncompressed data."); return (ARCHIVE_FATAL); } } windowoffs = lzss_offset_for_position(&rar->lzss, startpos); if(windowoffs + length <= lzss_size(&rar->lzss)) { memcpy(&rar->unp_buffer[rar->unp_offset], &rar->lzss.window[windowoffs], length); } else if (length <= lzss_size(&rar->lzss)) { firstpart = lzss_size(&rar->lzss) - windowoffs; if (firstpart < 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Bad RAR file data"); return (ARCHIVE_FATAL); } if (firstpart < length) { memcpy(&rar->unp_buffer[rar->unp_offset], &rar->lzss.window[windowoffs], firstpart); memcpy(&rar->unp_buffer[rar->unp_offset + firstpart], &rar->lzss.window[0], length - firstpart); } else { memcpy(&rar->unp_buffer[rar->unp_offset], &rar->lzss.window[windowoffs], length); } } else { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Bad RAR file data"); return (ARCHIVE_FATAL); } rar->unp_offset += length; if (rar->unp_offset >= rar->unp_buffer_size) *buffer = rar->unp_buffer; else *buffer = NULL; return (ARCHIVE_OK); } static const void * rar_read_ahead(struct archive_read *a, size_t min, ssize_t *avail) { struct rar *rar = (struct rar *)(a->format->data); const void *h = __archive_read_ahead(a, min, avail); int ret; if (avail) { if (a->archive.read_data_is_posix_read && *avail > (ssize_t)a->archive.read_data_requested) *avail = a->archive.read_data_requested; if (*avail > rar->bytes_remaining) *avail = (ssize_t)rar->bytes_remaining; if (*avail < 0) return NULL; else if (*avail == 0 && rar->main_flags & MHD_VOLUME && rar->file_flags & FHD_SPLIT_AFTER) { ret = archive_read_format_rar_read_header(a, a->entry); if (ret == (ARCHIVE_EOF)) { rar->has_endarc_header = 1; ret = archive_read_format_rar_read_header(a, a->entry); } if (ret != (ARCHIVE_OK)) return NULL; return rar_read_ahead(a, min, avail); } } return h; } Index: stable/10/contrib/libarchive/libarchive/archive_read_support_format_tar.c =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_read_support_format_tar.c (revision 324417) +++ stable/10/contrib/libarchive/libarchive/archive_read_support_format_tar.c (revision 324418) @@ -1,2885 +1,2885 @@ /*- * Copyright (c) 2003-2007 Tim Kientzle * Copyright (c) 2011-2012 Michihiro NAKAJIMA * Copyright (c) 2016 Martin Matuska * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "archive_platform.h" __FBSDID("$FreeBSD$"); #ifdef HAVE_ERRNO_H #include #endif #include #ifdef HAVE_STDLIB_H #include #endif #ifdef HAVE_STRING_H #include #endif #include "archive.h" #include "archive_acl_private.h" /* For ACL parsing routines. */ #include "archive_entry.h" #include "archive_entry_locale.h" #include "archive_private.h" #include "archive_read_private.h" #define tar_min(a,b) ((a) < (b) ? (a) : (b)) /* * Layout of POSIX 'ustar' tar header. */ struct archive_entry_header_ustar { char name[100]; char mode[8]; char uid[8]; char gid[8]; char size[12]; char mtime[12]; char checksum[8]; char typeflag[1]; char linkname[100]; /* "old format" header ends here */ char magic[6]; /* For POSIX: "ustar\0" */ char version[2]; /* For POSIX: "00" */ char uname[32]; char gname[32]; char rdevmajor[8]; char rdevminor[8]; char prefix[155]; }; /* * Structure of GNU tar header */ struct gnu_sparse { char offset[12]; char numbytes[12]; }; struct archive_entry_header_gnutar { char name[100]; char mode[8]; char uid[8]; char gid[8]; char size[12]; char mtime[12]; char checksum[8]; char typeflag[1]; char linkname[100]; char magic[8]; /* "ustar \0" (note blank/blank/null at end) */ char uname[32]; char gname[32]; char rdevmajor[8]; char rdevminor[8]; char atime[12]; char ctime[12]; char offset[12]; char longnames[4]; char unused[1]; struct gnu_sparse sparse[4]; char isextended[1]; char realsize[12]; /* * Old GNU format doesn't use POSIX 'prefix' field; they use * the 'L' (longname) entry instead. */ }; /* * Data specific to this format. */ struct sparse_block { struct sparse_block *next; int64_t offset; int64_t remaining; int hole; }; struct tar { struct archive_string acl_text; struct archive_string entry_pathname; /* For "GNU.sparse.name" and other similar path extensions. */ struct archive_string entry_pathname_override; struct archive_string entry_linkpath; struct archive_string entry_uname; struct archive_string entry_gname; struct archive_string longlink; struct archive_string longname; struct archive_string pax_header; struct archive_string pax_global; struct archive_string line; int pax_hdrcharset_binary; int header_recursion_depth; int64_t entry_bytes_remaining; int64_t entry_offset; int64_t entry_padding; int64_t entry_bytes_unconsumed; int64_t realsize; int sparse_allowed; struct sparse_block *sparse_list; struct sparse_block *sparse_last; int64_t sparse_offset; int64_t sparse_numbytes; int sparse_gnu_major; int sparse_gnu_minor; char sparse_gnu_pending; struct archive_string localname; struct archive_string_conv *opt_sconv; struct archive_string_conv *sconv; struct archive_string_conv *sconv_acl; struct archive_string_conv *sconv_default; int init_default_conversion; int compat_2x; int process_mac_extensions; int read_concatenated_archives; int realsize_override; }; static int archive_block_is_null(const char *p); static char *base64_decode(const char *, size_t, size_t *); static int gnu_add_sparse_entry(struct archive_read *, struct tar *, int64_t offset, int64_t remaining); static void gnu_clear_sparse_list(struct tar *); static int gnu_sparse_old_read(struct archive_read *, struct tar *, const struct archive_entry_header_gnutar *header, size_t *); static int gnu_sparse_old_parse(struct archive_read *, struct tar *, const struct gnu_sparse *sparse, int length); static int gnu_sparse_01_parse(struct archive_read *, struct tar *, const char *); static ssize_t gnu_sparse_10_read(struct archive_read *, struct tar *, size_t *); static int header_Solaris_ACL(struct archive_read *, struct tar *, struct archive_entry *, const void *, size_t *); static int header_common(struct archive_read *, struct tar *, struct archive_entry *, const void *); static int header_old_tar(struct archive_read *, struct tar *, struct archive_entry *, const void *); static int header_pax_extensions(struct archive_read *, struct tar *, struct archive_entry *, const void *, size_t *); static int header_pax_global(struct archive_read *, struct tar *, struct archive_entry *, const void *h, size_t *); static int header_longlink(struct archive_read *, struct tar *, struct archive_entry *, const void *h, size_t *); static int header_longname(struct archive_read *, struct tar *, struct archive_entry *, const void *h, size_t *); static int read_mac_metadata_blob(struct archive_read *, struct tar *, struct archive_entry *, const void *h, size_t *); static int header_volume(struct archive_read *, struct tar *, struct archive_entry *, const void *h, size_t *); static int header_ustar(struct archive_read *, struct tar *, struct archive_entry *, const void *h); static int header_gnutar(struct archive_read *, struct tar *, struct archive_entry *, const void *h, size_t *); static int archive_read_format_tar_bid(struct archive_read *, int); static int archive_read_format_tar_options(struct archive_read *, const char *, const char *); static int archive_read_format_tar_cleanup(struct archive_read *); static int archive_read_format_tar_read_data(struct archive_read *a, const void **buff, size_t *size, int64_t *offset); static int archive_read_format_tar_skip(struct archive_read *a); static int archive_read_format_tar_read_header(struct archive_read *, struct archive_entry *); static int checksum(struct archive_read *, const void *); static int pax_attribute(struct archive_read *, struct tar *, struct archive_entry *, const char *key, const char *value, size_t value_length); static int pax_attribute_acl(struct archive_read *, struct tar *, struct archive_entry *, const char *, int); static int pax_attribute_xattr(struct archive_entry *, const char *, const char *); static int pax_header(struct archive_read *, struct tar *, struct archive_entry *, struct archive_string *); static void pax_time(const char *, int64_t *sec, long *nanos); static ssize_t readline(struct archive_read *, struct tar *, const char **, ssize_t limit, size_t *); static int read_body_to_string(struct archive_read *, struct tar *, struct archive_string *, const void *h, size_t *); static int solaris_sparse_parse(struct archive_read *, struct tar *, struct archive_entry *, const char *); static int64_t tar_atol(const char *, size_t); static int64_t tar_atol10(const char *, size_t); static int64_t tar_atol256(const char *, size_t); static int64_t tar_atol8(const char *, size_t); static int tar_read_header(struct archive_read *, struct tar *, struct archive_entry *, size_t *); static int tohex(int c); static char *url_decode(const char *); static void tar_flush_unconsumed(struct archive_read *, size_t *); int archive_read_support_format_gnutar(struct archive *a) { archive_check_magic(a, ARCHIVE_READ_MAGIC, ARCHIVE_STATE_NEW, "archive_read_support_format_gnutar"); return (archive_read_support_format_tar(a)); } int archive_read_support_format_tar(struct archive *_a) { struct archive_read *a = (struct archive_read *)_a; struct tar *tar; int r; archive_check_magic(_a, ARCHIVE_READ_MAGIC, ARCHIVE_STATE_NEW, "archive_read_support_format_tar"); tar = (struct tar *)calloc(1, sizeof(*tar)); #ifdef HAVE_COPYFILE_H /* Set this by default on Mac OS. */ tar->process_mac_extensions = 1; #endif if (tar == NULL) { archive_set_error(&a->archive, ENOMEM, "Can't allocate tar data"); return (ARCHIVE_FATAL); } r = __archive_read_register_format(a, tar, "tar", archive_read_format_tar_bid, archive_read_format_tar_options, archive_read_format_tar_read_header, archive_read_format_tar_read_data, archive_read_format_tar_skip, NULL, archive_read_format_tar_cleanup, NULL, NULL); if (r != ARCHIVE_OK) free(tar); return (ARCHIVE_OK); } static int archive_read_format_tar_cleanup(struct archive_read *a) { struct tar *tar; tar = (struct tar *)(a->format->data); gnu_clear_sparse_list(tar); archive_string_free(&tar->acl_text); archive_string_free(&tar->entry_pathname); archive_string_free(&tar->entry_pathname_override); archive_string_free(&tar->entry_linkpath); archive_string_free(&tar->entry_uname); archive_string_free(&tar->entry_gname); archive_string_free(&tar->line); archive_string_free(&tar->pax_global); archive_string_free(&tar->pax_header); archive_string_free(&tar->longname); archive_string_free(&tar->longlink); archive_string_free(&tar->localname); free(tar); (a->format->data) = NULL; return (ARCHIVE_OK); } /* * Validate number field * * This has to be pretty lenient in order to accommodate the enormous * variety of tar writers in the world: * = POSIX (IEEE Std 1003.1-1988) ustar requires octal values with leading * zeros and allows fields to be terminated with space or null characters * = Many writers use different termination (in particular, libarchive * omits terminator bytes to squeeze one or two more digits) * = Many writers pad with space and omit leading zeros * = GNU tar and star write base-256 values if numbers are too * big to be represented in octal * * Examples of specific tar headers that we should support: * = Perl Archive::Tar terminates uid, gid, devminor and devmajor with two * null bytes, pads size with spaces and other numeric fields with zeroes * = plexus-archiver prior to 2.6.3 (before switching to commons-compress) * may have uid and gid fields filled with spaces without any octal digits * at all and pads all numeric fields with spaces * * This should tolerate all variants in use. It will reject a field * where the writer just left garbage after a trailing NUL. */ static int validate_number_field(const char* p_field, size_t i_size) { unsigned char marker = (unsigned char)p_field[0]; if (marker == 128 || marker == 255 || marker == 0) { /* Base-256 marker, there's nothing we can check. */ return 1; } else { /* Must be octal */ size_t i = 0; /* Skip any leading spaces */ while (i < i_size && p_field[i] == ' ') { ++i; } /* Skip octal digits. */ while (i < i_size && p_field[i] >= '0' && p_field[i] <= '7') { ++i; } /* Any remaining characters must be space or NUL padding. */ while (i < i_size) { if (p_field[i] != ' ' && p_field[i] != 0) { return 0; } ++i; } return 1; } } static int archive_read_format_tar_bid(struct archive_read *a, int best_bid) { int bid; const char *h; const struct archive_entry_header_ustar *header; (void)best_bid; /* UNUSED */ bid = 0; /* Now let's look at the actual header and see if it matches. */ h = __archive_read_ahead(a, 512, NULL); if (h == NULL) return (-1); /* If it's an end-of-archive mark, we can handle it. */ if (h[0] == 0 && archive_block_is_null(h)) { /* * Usually, I bid the number of bits verified, but * in this case, 4096 seems excessive so I picked 10 as * an arbitrary but reasonable-seeming value. */ return (10); } /* If it's not an end-of-archive mark, it must have a valid checksum.*/ if (!checksum(a, h)) return (0); bid += 48; /* Checksum is usually 6 octal digits. */ header = (const struct archive_entry_header_ustar *)h; /* Recognize POSIX formats. */ if ((memcmp(header->magic, "ustar\0", 6) == 0) && (memcmp(header->version, "00", 2) == 0)) bid += 56; /* Recognize GNU tar format. */ if ((memcmp(header->magic, "ustar ", 6) == 0) && (memcmp(header->version, " \0", 2) == 0)) bid += 56; /* Type flag must be null, digit or A-Z, a-z. */ if (header->typeflag[0] != 0 && !( header->typeflag[0] >= '0' && header->typeflag[0] <= '9') && !( header->typeflag[0] >= 'A' && header->typeflag[0] <= 'Z') && !( header->typeflag[0] >= 'a' && header->typeflag[0] <= 'z') ) return (0); bid += 2; /* 6 bits of variation in an 8-bit field leaves 2 bits. */ /* * Check format of mode/uid/gid/mtime/size/rdevmajor/rdevminor fields. */ if (bid > 0 && ( validate_number_field(header->mode, sizeof(header->mode)) == 0 || validate_number_field(header->uid, sizeof(header->uid)) == 0 || validate_number_field(header->gid, sizeof(header->gid)) == 0 || validate_number_field(header->mtime, sizeof(header->mtime)) == 0 || validate_number_field(header->size, sizeof(header->size)) == 0 || validate_number_field(header->rdevmajor, sizeof(header->rdevmajor)) == 0 || validate_number_field(header->rdevminor, sizeof(header->rdevminor)) == 0)) { bid = 0; } return (bid); } static int archive_read_format_tar_options(struct archive_read *a, const char *key, const char *val) { struct tar *tar; int ret = ARCHIVE_FAILED; tar = (struct tar *)(a->format->data); if (strcmp(key, "compat-2x") == 0) { /* Handle UTF-8 filenames as libarchive 2.x */ tar->compat_2x = (val != NULL && val[0] != 0); tar->init_default_conversion = tar->compat_2x; return (ARCHIVE_OK); } else if (strcmp(key, "hdrcharset") == 0) { if (val == NULL || val[0] == 0) archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "tar: hdrcharset option needs a character-set name"); else { tar->opt_sconv = archive_string_conversion_from_charset( &a->archive, val, 0); if (tar->opt_sconv != NULL) ret = ARCHIVE_OK; else ret = ARCHIVE_FATAL; } return (ret); } else if (strcmp(key, "mac-ext") == 0) { tar->process_mac_extensions = (val != NULL && val[0] != 0); return (ARCHIVE_OK); } else if (strcmp(key, "read_concatenated_archives") == 0) { tar->read_concatenated_archives = (val != NULL && val[0] != 0); return (ARCHIVE_OK); } /* Note: The "warn" return is just to inform the options * supervisor that we didn't handle it. It will generate * a suitable error if no one used this option. */ return (ARCHIVE_WARN); } /* utility function- this exists to centralize the logic of tracking * how much unconsumed data we have floating around, and to consume * anything outstanding since we're going to do read_aheads */ static void tar_flush_unconsumed(struct archive_read *a, size_t *unconsumed) { if (*unconsumed) { /* void *data = (void *)__archive_read_ahead(a, *unconsumed, NULL); * this block of code is to poison claimed unconsumed space, ensuring * things break if it is in use still. * currently it WILL break things, so enable it only for debugging this issue if (data) { memset(data, 0xff, *unconsumed); } */ __archive_read_consume(a, *unconsumed); *unconsumed = 0; } } /* * The function invoked by archive_read_next_header(). This * just sets up a few things and then calls the internal * tar_read_header() function below. */ static int archive_read_format_tar_read_header(struct archive_read *a, struct archive_entry *entry) { /* * When converting tar archives to cpio archives, it is * essential that each distinct file have a distinct inode * number. To simplify this, we keep a static count here to * assign fake dev/inode numbers to each tar entry. Note that * pax format archives may overwrite this with something more * useful. * * Ideally, we would track every file read from the archive so * that we could assign the same dev/ino pair to hardlinks, * but the memory required to store a complete lookup table is * probably not worthwhile just to support the relatively * obscure tar->cpio conversion case. */ static int default_inode; static int default_dev; struct tar *tar; const char *p; const wchar_t *wp; int r; size_t l, unconsumed = 0; /* Assign default device/inode values. */ archive_entry_set_dev(entry, 1 + default_dev); /* Don't use zero. */ archive_entry_set_ino(entry, ++default_inode); /* Don't use zero. */ /* Limit generated st_ino number to 16 bits. */ if (default_inode >= 0xffff) { ++default_dev; default_inode = 0; } tar = (struct tar *)(a->format->data); tar->entry_offset = 0; gnu_clear_sparse_list(tar); tar->realsize = -1; /* Mark this as "unset" */ tar->realsize_override = 0; /* Setup default string conversion. */ tar->sconv = tar->opt_sconv; if (tar->sconv == NULL) { if (!tar->init_default_conversion) { tar->sconv_default = archive_string_default_conversion_for_read(&(a->archive)); tar->init_default_conversion = 1; } tar->sconv = tar->sconv_default; } r = tar_read_header(a, tar, entry, &unconsumed); tar_flush_unconsumed(a, &unconsumed); /* * "non-sparse" files are really just sparse files with * a single block. */ if (tar->sparse_list == NULL) { if (gnu_add_sparse_entry(a, tar, 0, tar->entry_bytes_remaining) != ARCHIVE_OK) return (ARCHIVE_FATAL); } else { struct sparse_block *sb; for (sb = tar->sparse_list; sb != NULL; sb = sb->next) { if (!sb->hole) archive_entry_sparse_add_entry(entry, sb->offset, sb->remaining); } } if (r == ARCHIVE_OK && archive_entry_filetype(entry) == AE_IFREG) { /* * "Regular" entry with trailing '/' is really * directory: This is needed for certain old tar * variants and even for some broken newer ones. */ if ((wp = archive_entry_pathname_w(entry)) != NULL) { l = wcslen(wp); if (l > 0 && wp[l - 1] == L'/') { archive_entry_set_filetype(entry, AE_IFDIR); } } else if ((p = archive_entry_pathname(entry)) != NULL) { l = strlen(p); if (l > 0 && p[l - 1] == '/') { archive_entry_set_filetype(entry, AE_IFDIR); } } } return (r); } static int archive_read_format_tar_read_data(struct archive_read *a, const void **buff, size_t *size, int64_t *offset) { ssize_t bytes_read; struct tar *tar; struct sparse_block *p; tar = (struct tar *)(a->format->data); for (;;) { /* Remove exhausted entries from sparse list. */ while (tar->sparse_list != NULL && tar->sparse_list->remaining == 0) { p = tar->sparse_list; tar->sparse_list = p->next; free(p); } if (tar->entry_bytes_unconsumed) { __archive_read_consume(a, tar->entry_bytes_unconsumed); tar->entry_bytes_unconsumed = 0; } /* If we're at end of file, return EOF. */ if (tar->sparse_list == NULL || tar->entry_bytes_remaining == 0) { if (__archive_read_consume(a, tar->entry_padding) < 0) return (ARCHIVE_FATAL); tar->entry_padding = 0; *buff = NULL; *size = 0; *offset = tar->realsize; return (ARCHIVE_EOF); } *buff = __archive_read_ahead(a, 1, &bytes_read); if (bytes_read < 0) return (ARCHIVE_FATAL); if (*buff == NULL) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Truncated tar archive"); return (ARCHIVE_FATAL); } if (bytes_read > tar->entry_bytes_remaining) bytes_read = (ssize_t)tar->entry_bytes_remaining; /* Don't read more than is available in the * current sparse block. */ if (tar->sparse_list->remaining < bytes_read) bytes_read = (ssize_t)tar->sparse_list->remaining; *size = bytes_read; *offset = tar->sparse_list->offset; tar->sparse_list->remaining -= bytes_read; tar->sparse_list->offset += bytes_read; tar->entry_bytes_remaining -= bytes_read; tar->entry_bytes_unconsumed = bytes_read; if (!tar->sparse_list->hole) return (ARCHIVE_OK); /* Current is hole data and skip this. */ } } static int archive_read_format_tar_skip(struct archive_read *a) { int64_t bytes_skipped; int64_t request; struct sparse_block *p; struct tar* tar; tar = (struct tar *)(a->format->data); /* Do not consume the hole of a sparse file. */ request = 0; for (p = tar->sparse_list; p != NULL; p = p->next) { if (!p->hole) { if (p->remaining >= INT64_MAX - request) { return ARCHIVE_FATAL; } request += p->remaining; } } if (request > tar->entry_bytes_remaining) request = tar->entry_bytes_remaining; request += tar->entry_padding + tar->entry_bytes_unconsumed; bytes_skipped = __archive_read_consume(a, request); if (bytes_skipped < 0) return (ARCHIVE_FATAL); tar->entry_bytes_remaining = 0; tar->entry_bytes_unconsumed = 0; tar->entry_padding = 0; /* Free the sparse list. */ gnu_clear_sparse_list(tar); return (ARCHIVE_OK); } /* * This function recursively interprets all of the headers associated * with a single entry. */ static int tar_read_header(struct archive_read *a, struct tar *tar, struct archive_entry *entry, size_t *unconsumed) { ssize_t bytes; int err; const char *h; const struct archive_entry_header_ustar *header; const struct archive_entry_header_gnutar *gnuheader; /* Loop until we find a workable header record. */ for (;;) { tar_flush_unconsumed(a, unconsumed); /* Read 512-byte header record */ h = __archive_read_ahead(a, 512, &bytes); if (bytes < 0) return ((int)bytes); if (bytes == 0) { /* EOF at a block boundary. */ /* Some writers do omit the block of nulls. */ return (ARCHIVE_EOF); } if (bytes < 512) { /* Short block at EOF; this is bad. */ archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated tar archive"); return (ARCHIVE_FATAL); } *unconsumed = 512; /* Header is workable if it's not an end-of-archive mark. */ if (h[0] != 0 || !archive_block_is_null(h)) break; /* Ensure format is set for archives with only null blocks. */ if (a->archive.archive_format_name == NULL) { a->archive.archive_format = ARCHIVE_FORMAT_TAR; a->archive.archive_format_name = "tar"; } if (!tar->read_concatenated_archives) { /* Try to consume a second all-null record, as well. */ tar_flush_unconsumed(a, unconsumed); h = __archive_read_ahead(a, 512, NULL); if (h != NULL && h[0] == 0 && archive_block_is_null(h)) __archive_read_consume(a, 512); archive_clear_error(&a->archive); return (ARCHIVE_EOF); } /* * We're reading concatenated archives, ignore this block and * loop to get the next. */ } /* * Note: If the checksum fails and we return ARCHIVE_RETRY, * then the client is likely to just retry. This is a very * crude way to search for the next valid header! * * TODO: Improve this by implementing a real header scan. */ if (!checksum(a, h)) { tar_flush_unconsumed(a, unconsumed); archive_set_error(&a->archive, EINVAL, "Damaged tar archive"); return (ARCHIVE_RETRY); /* Retryable: Invalid header */ } if (++tar->header_recursion_depth > 32) { tar_flush_unconsumed(a, unconsumed); archive_set_error(&a->archive, EINVAL, "Too many special headers"); return (ARCHIVE_WARN); } /* Determine the format variant. */ header = (const struct archive_entry_header_ustar *)h; switch(header->typeflag[0]) { case 'A': /* Solaris tar ACL */ a->archive.archive_format = ARCHIVE_FORMAT_TAR_PAX_INTERCHANGE; a->archive.archive_format_name = "Solaris tar"; err = header_Solaris_ACL(a, tar, entry, h, unconsumed); break; case 'g': /* POSIX-standard 'g' header. */ a->archive.archive_format = ARCHIVE_FORMAT_TAR_PAX_INTERCHANGE; a->archive.archive_format_name = "POSIX pax interchange format"; err = header_pax_global(a, tar, entry, h, unconsumed); if (err == ARCHIVE_EOF) return (err); break; case 'K': /* Long link name (GNU tar, others) */ err = header_longlink(a, tar, entry, h, unconsumed); break; case 'L': /* Long filename (GNU tar, others) */ err = header_longname(a, tar, entry, h, unconsumed); break; case 'V': /* GNU volume header */ err = header_volume(a, tar, entry, h, unconsumed); break; case 'X': /* Used by SUN tar; same as 'x'. */ a->archive.archive_format = ARCHIVE_FORMAT_TAR_PAX_INTERCHANGE; a->archive.archive_format_name = "POSIX pax interchange format (Sun variant)"; err = header_pax_extensions(a, tar, entry, h, unconsumed); break; case 'x': /* POSIX-standard 'x' header. */ a->archive.archive_format = ARCHIVE_FORMAT_TAR_PAX_INTERCHANGE; a->archive.archive_format_name = "POSIX pax interchange format"; err = header_pax_extensions(a, tar, entry, h, unconsumed); break; default: gnuheader = (const struct archive_entry_header_gnutar *)h; if (memcmp(gnuheader->magic, "ustar \0", 8) == 0) { a->archive.archive_format = ARCHIVE_FORMAT_TAR_GNUTAR; a->archive.archive_format_name = "GNU tar format"; err = header_gnutar(a, tar, entry, h, unconsumed); } else if (memcmp(header->magic, "ustar", 5) == 0) { if (a->archive.archive_format != ARCHIVE_FORMAT_TAR_PAX_INTERCHANGE) { a->archive.archive_format = ARCHIVE_FORMAT_TAR_USTAR; a->archive.archive_format_name = "POSIX ustar format"; } err = header_ustar(a, tar, entry, h); } else { a->archive.archive_format = ARCHIVE_FORMAT_TAR; a->archive.archive_format_name = "tar (non-POSIX)"; err = header_old_tar(a, tar, entry, h); } } if (err == ARCHIVE_FATAL) return (err); tar_flush_unconsumed(a, unconsumed); h = NULL; header = NULL; --tar->header_recursion_depth; /* Yuck. Apple's design here ends up storing long pathname * extensions for both the AppleDouble extension entry and the * regular entry. */ if ((err == ARCHIVE_WARN || err == ARCHIVE_OK) && tar->header_recursion_depth == 0 && tar->process_mac_extensions) { int err2 = read_mac_metadata_blob(a, tar, entry, h, unconsumed); if (err2 < err) err = err2; } /* We return warnings or success as-is. Anything else is fatal. */ if (err == ARCHIVE_WARN || err == ARCHIVE_OK) { if (tar->sparse_gnu_pending) { if (tar->sparse_gnu_major == 1 && tar->sparse_gnu_minor == 0) { ssize_t bytes_read; tar->sparse_gnu_pending = 0; /* Read initial sparse map. */ bytes_read = gnu_sparse_10_read(a, tar, unconsumed); if (bytes_read < 0) return ((int)bytes_read); tar->entry_bytes_remaining -= bytes_read; } else { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Unrecognized GNU sparse file format"); return (ARCHIVE_WARN); } tar->sparse_gnu_pending = 0; } return (err); } if (err == ARCHIVE_EOF) /* EOF when recursively reading a header is bad. */ archive_set_error(&a->archive, EINVAL, "Damaged tar archive"); return (ARCHIVE_FATAL); } /* * Return true if block checksum is correct. */ static int checksum(struct archive_read *a, const void *h) { const unsigned char *bytes; const struct archive_entry_header_ustar *header; int check, sum; size_t i; (void)a; /* UNUSED */ bytes = (const unsigned char *)h; header = (const struct archive_entry_header_ustar *)h; /* Checksum field must hold an octal number */ for (i = 0; i < sizeof(header->checksum); ++i) { char c = header->checksum[i]; if (c != ' ' && c != '\0' && (c < '0' || c > '7')) return 0; } /* * Test the checksum. Note that POSIX specifies _unsigned_ * bytes for this calculation. */ sum = (int)tar_atol(header->checksum, sizeof(header->checksum)); check = 0; for (i = 0; i < 148; i++) check += (unsigned char)bytes[i]; for (; i < 156; i++) check += 32; for (; i < 512; i++) check += (unsigned char)bytes[i]; if (sum == check) return (1); /* * Repeat test with _signed_ bytes, just in case this archive * was created by an old BSD, Solaris, or HP-UX tar with a * broken checksum calculation. */ check = 0; for (i = 0; i < 148; i++) check += (signed char)bytes[i]; for (; i < 156; i++) check += 32; for (; i < 512; i++) check += (signed char)bytes[i]; if (sum == check) return (1); return (0); } /* * Return true if this block contains only nulls. */ static int archive_block_is_null(const char *p) { unsigned i; for (i = 0; i < 512; i++) if (*p++) return (0); return (1); } /* * Interpret 'A' Solaris ACL header */ static int header_Solaris_ACL(struct archive_read *a, struct tar *tar, struct archive_entry *entry, const void *h, size_t *unconsumed) { const struct archive_entry_header_ustar *header; size_t size; int err, acl_type; int64_t type; char *acl, *p; /* * read_body_to_string adds a NUL terminator, but we need a little * more to make sure that we don't overrun acl_text later. */ header = (const struct archive_entry_header_ustar *)h; size = (size_t)tar_atol(header->size, sizeof(header->size)); err = read_body_to_string(a, tar, &(tar->acl_text), h, unconsumed); if (err != ARCHIVE_OK) return (err); /* Recursively read next header */ err = tar_read_header(a, tar, entry, unconsumed); if ((err != ARCHIVE_OK) && (err != ARCHIVE_WARN)) return (err); /* TODO: Examine the first characters to see if this * is an AIX ACL descriptor. We'll likely never support * them, but it would be polite to recognize and warn when * we do see them. */ /* Leading octal number indicates ACL type and number of entries. */ p = acl = tar->acl_text.s; type = 0; while (*p != '\0' && p < acl + size) { if (*p < '0' || *p > '7') { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Malformed Solaris ACL attribute (invalid digit)"); return(ARCHIVE_WARN); } type <<= 3; type += *p - '0'; if (type > 077777777) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Malformed Solaris ACL attribute (count too large)"); return (ARCHIVE_WARN); } p++; } switch ((int)type & ~0777777) { case 01000000: /* POSIX.1e ACL */ acl_type = ARCHIVE_ENTRY_ACL_TYPE_ACCESS; break; case 03000000: /* NFSv4 ACL */ acl_type = ARCHIVE_ENTRY_ACL_TYPE_NFS4; break; default: archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Malformed Solaris ACL attribute (unsupported type %o)", (int)type); return (ARCHIVE_WARN); } p++; if (p >= acl + size) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Malformed Solaris ACL attribute (body overflow)"); return(ARCHIVE_WARN); } /* ACL text is null-terminated; find the end. */ size -= (p - acl); acl = p; while (*p != '\0' && p < acl + size) p++; if (tar->sconv_acl == NULL) { tar->sconv_acl = archive_string_conversion_from_charset( &(a->archive), "UTF-8", 1); if (tar->sconv_acl == NULL) return (ARCHIVE_FATAL); } archive_strncpy(&(tar->localname), acl, p - acl); err = archive_acl_from_text_l(archive_entry_acl(entry), tar->localname.s, acl_type, tar->sconv_acl); if (err != ARCHIVE_OK) { if (errno == ENOMEM) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory for ACL"); } else archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Malformed Solaris ACL attribute (unparsable)"); } return (err); } /* * Interpret 'K' long linkname header. */ static int header_longlink(struct archive_read *a, struct tar *tar, struct archive_entry *entry, const void *h, size_t *unconsumed) { int err; err = read_body_to_string(a, tar, &(tar->longlink), h, unconsumed); if (err != ARCHIVE_OK) return (err); err = tar_read_header(a, tar, entry, unconsumed); if ((err != ARCHIVE_OK) && (err != ARCHIVE_WARN)) return (err); /* Set symlink if symlink already set, else hardlink. */ archive_entry_copy_link(entry, tar->longlink.s); return (ARCHIVE_OK); } static int set_conversion_failed_error(struct archive_read *a, struct archive_string_conv *sconv, const char *name) { if (errno == ENOMEM) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory for %s", name); return (ARCHIVE_FATAL); } archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "%s can't be converted from %s to current locale.", name, archive_string_conversion_charset_name(sconv)); return (ARCHIVE_WARN); } /* * Interpret 'L' long filename header. */ static int header_longname(struct archive_read *a, struct tar *tar, struct archive_entry *entry, const void *h, size_t *unconsumed) { int err; err = read_body_to_string(a, tar, &(tar->longname), h, unconsumed); if (err != ARCHIVE_OK) return (err); /* Read and parse "real" header, then override name. */ err = tar_read_header(a, tar, entry, unconsumed); if ((err != ARCHIVE_OK) && (err != ARCHIVE_WARN)) return (err); if (archive_entry_copy_pathname_l(entry, tar->longname.s, archive_strlen(&(tar->longname)), tar->sconv) != 0) err = set_conversion_failed_error(a, tar->sconv, "Pathname"); return (err); } /* * Interpret 'V' GNU tar volume header. */ static int header_volume(struct archive_read *a, struct tar *tar, struct archive_entry *entry, const void *h, size_t *unconsumed) { (void)h; /* Just skip this and read the next header. */ return (tar_read_header(a, tar, entry, unconsumed)); } /* * Read body of an archive entry into an archive_string object. */ static int read_body_to_string(struct archive_read *a, struct tar *tar, struct archive_string *as, const void *h, size_t *unconsumed) { int64_t size; const struct archive_entry_header_ustar *header; const void *src; (void)tar; /* UNUSED */ header = (const struct archive_entry_header_ustar *)h; size = tar_atol(header->size, sizeof(header->size)); if ((size > 1048576) || (size < 0)) { archive_set_error(&a->archive, EINVAL, "Special header too large"); return (ARCHIVE_FATAL); } /* Fail if we can't make our buffer big enough. */ if (archive_string_ensure(as, (size_t)size+1) == NULL) { archive_set_error(&a->archive, ENOMEM, "No memory"); return (ARCHIVE_FATAL); } tar_flush_unconsumed(a, unconsumed); /* Read the body into the string. */ *unconsumed = (size_t)((size + 511) & ~ 511); src = __archive_read_ahead(a, *unconsumed, NULL); if (src == NULL) { *unconsumed = 0; return (ARCHIVE_FATAL); } memcpy(as->s, src, (size_t)size); as->s[size] = '\0'; as->length = (size_t)size; return (ARCHIVE_OK); } /* * Parse out common header elements. * * This would be the same as header_old_tar, except that the * filename is handled slightly differently for old and POSIX * entries (POSIX entries support a 'prefix'). This factoring * allows header_old_tar and header_ustar * to handle filenames differently, while still putting most of the * common parsing into one place. */ static int header_common(struct archive_read *a, struct tar *tar, struct archive_entry *entry, const void *h) { const struct archive_entry_header_ustar *header; char tartype; int err = ARCHIVE_OK; header = (const struct archive_entry_header_ustar *)h; if (header->linkname[0]) archive_strncpy(&(tar->entry_linkpath), header->linkname, sizeof(header->linkname)); else archive_string_empty(&(tar->entry_linkpath)); /* Parse out the numeric fields (all are octal) */ archive_entry_set_mode(entry, (mode_t)tar_atol(header->mode, sizeof(header->mode))); archive_entry_set_uid(entry, tar_atol(header->uid, sizeof(header->uid))); archive_entry_set_gid(entry, tar_atol(header->gid, sizeof(header->gid))); tar->entry_bytes_remaining = tar_atol(header->size, sizeof(header->size)); if (tar->entry_bytes_remaining < 0) { tar->entry_bytes_remaining = 0; archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Tar entry has negative size"); return (ARCHIVE_FATAL); } if (tar->entry_bytes_remaining == INT64_MAX) { /* Note: tar_atol returns INT64_MAX on overflow */ tar->entry_bytes_remaining = 0; archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Tar entry size overflow"); return (ARCHIVE_FATAL); } tar->realsize = tar->entry_bytes_remaining; archive_entry_set_size(entry, tar->entry_bytes_remaining); archive_entry_set_mtime(entry, tar_atol(header->mtime, sizeof(header->mtime)), 0); /* Handle the tar type flag appropriately. */ tartype = header->typeflag[0]; switch (tartype) { case '1': /* Hard link */ if (archive_entry_copy_hardlink_l(entry, tar->entry_linkpath.s, archive_strlen(&(tar->entry_linkpath)), tar->sconv) != 0) { err = set_conversion_failed_error(a, tar->sconv, "Linkname"); if (err == ARCHIVE_FATAL) return (err); } /* * The following may seem odd, but: Technically, tar * does not store the file type for a "hard link" * entry, only the fact that it is a hard link. So, I * leave the type zero normally. But, pax interchange * format allows hard links to have data, which * implies that the underlying entry is a regular * file. */ if (archive_entry_size(entry) > 0) archive_entry_set_filetype(entry, AE_IFREG); /* * A tricky point: Traditionally, tar readers have * ignored the size field when reading hardlink * entries, and some writers put non-zero sizes even * though the body is empty. POSIX blessed this * convention in the 1988 standard, but broke with * this tradition in 2001 by permitting hardlink * entries to store valid bodies in pax interchange * format, but not in ustar format. Since there is no * hard and fast way to distinguish pax interchange * from earlier archives (the 'x' and 'g' entries are * optional, after all), we need a heuristic. */ if (archive_entry_size(entry) == 0) { /* If the size is already zero, we're done. */ } else if (a->archive.archive_format == ARCHIVE_FORMAT_TAR_PAX_INTERCHANGE) { /* Definitely pax extended; must obey hardlink size. */ } else if (a->archive.archive_format == ARCHIVE_FORMAT_TAR || a->archive.archive_format == ARCHIVE_FORMAT_TAR_GNUTAR) { /* Old-style or GNU tar: we must ignore the size. */ archive_entry_set_size(entry, 0); tar->entry_bytes_remaining = 0; } else if (archive_read_format_tar_bid(a, 50) > 50) { /* * We don't know if it's pax: If the bid * function sees a valid ustar header * immediately following, then let's ignore * the hardlink size. */ archive_entry_set_size(entry, 0); tar->entry_bytes_remaining = 0; } /* * TODO: There are still two cases I'd like to handle: * = a ustar non-pax archive with a hardlink entry at * end-of-archive. (Look for block of nulls following?) * = a pax archive that has not seen any pax headers * and has an entry which is a hardlink entry storing * a body containing an uncompressed tar archive. * The first is worth addressing; I don't see any reliable * way to deal with the second possibility. */ break; case '2': /* Symlink */ archive_entry_set_filetype(entry, AE_IFLNK); archive_entry_set_size(entry, 0); tar->entry_bytes_remaining = 0; if (archive_entry_copy_symlink_l(entry, tar->entry_linkpath.s, archive_strlen(&(tar->entry_linkpath)), tar->sconv) != 0) { err = set_conversion_failed_error(a, tar->sconv, "Linkname"); if (err == ARCHIVE_FATAL) return (err); } break; case '3': /* Character device */ archive_entry_set_filetype(entry, AE_IFCHR); archive_entry_set_size(entry, 0); tar->entry_bytes_remaining = 0; break; case '4': /* Block device */ archive_entry_set_filetype(entry, AE_IFBLK); archive_entry_set_size(entry, 0); tar->entry_bytes_remaining = 0; break; case '5': /* Dir */ archive_entry_set_filetype(entry, AE_IFDIR); archive_entry_set_size(entry, 0); tar->entry_bytes_remaining = 0; break; case '6': /* FIFO device */ archive_entry_set_filetype(entry, AE_IFIFO); archive_entry_set_size(entry, 0); tar->entry_bytes_remaining = 0; break; case 'D': /* GNU incremental directory type */ /* * No special handling is actually required here. * It might be nice someday to preprocess the file list and * provide it to the client, though. */ archive_entry_set_filetype(entry, AE_IFDIR); break; case 'M': /* GNU "Multi-volume" (remainder of file from last archive)*/ /* * As far as I can tell, this is just like a regular file * entry, except that the contents should be _appended_ to * the indicated file at the indicated offset. This may * require some API work to fully support. */ break; case 'N': /* Old GNU "long filename" entry. */ /* The body of this entry is a script for renaming * previously-extracted entries. Ugh. It will never * be supported by libarchive. */ archive_entry_set_filetype(entry, AE_IFREG); break; case 'S': /* GNU sparse files */ /* * Sparse files are really just regular files with * sparse information in the extended area. */ /* FALLTHROUGH */ case '0': /* * Enable sparse file "read" support only for regular * files and explicit GNU sparse files. However, we * don't allow non-standard file types to be sparse. */ tar->sparse_allowed = 1; /* FALLTHROUGH */ default: /* Regular file and non-standard types */ /* * Per POSIX: non-recognized types should always be * treated as regular files. */ archive_entry_set_filetype(entry, AE_IFREG); break; } return (err); } /* * Parse out header elements for "old-style" tar archives. */ static int header_old_tar(struct archive_read *a, struct tar *tar, struct archive_entry *entry, const void *h) { const struct archive_entry_header_ustar *header; int err = ARCHIVE_OK, err2; /* Copy filename over (to ensure null termination). */ header = (const struct archive_entry_header_ustar *)h; if (archive_entry_copy_pathname_l(entry, header->name, sizeof(header->name), tar->sconv) != 0) { err = set_conversion_failed_error(a, tar->sconv, "Pathname"); if (err == ARCHIVE_FATAL) return (err); } /* Grab rest of common fields */ err2 = header_common(a, tar, entry, h); if (err > err2) err = err2; tar->entry_padding = 0x1ff & (-tar->entry_bytes_remaining); return (err); } /* * Read a Mac AppleDouble-encoded blob of file metadata, * if there is one. */ static int read_mac_metadata_blob(struct archive_read *a, struct tar *tar, struct archive_entry *entry, const void *h, size_t *unconsumed) { int64_t size; const void *data; const char *p, *name; const wchar_t *wp, *wname; (void)h; /* UNUSED */ wname = wp = archive_entry_pathname_w(entry); if (wp != NULL) { /* Find the last path element. */ for (; *wp != L'\0'; ++wp) { if (wp[0] == '/' && wp[1] != L'\0') wname = wp + 1; } /* * If last path element starts with "._", then * this is a Mac extension. */ if (wname[0] != L'.' || wname[1] != L'_' || wname[2] == L'\0') return ARCHIVE_OK; } else { /* Find the last path element. */ name = p = archive_entry_pathname(entry); if (p == NULL) return (ARCHIVE_FAILED); for (; *p != '\0'; ++p) { if (p[0] == '/' && p[1] != '\0') name = p + 1; } /* * If last path element starts with "._", then * this is a Mac extension. */ if (name[0] != '.' || name[1] != '_' || name[2] == '\0') return ARCHIVE_OK; } /* Read the body as a Mac OS metadata blob. */ size = archive_entry_size(entry); /* * TODO: Look beyond the body here to peek at the next header. * If it's a regular header (not an extension header) * that has the wrong name, just return the current * entry as-is, without consuming the body here. * That would reduce the risk of us mis-identifying * an ordinary file that just happened to have * a name starting with "._". * * Q: Is the above idea really possible? Even * when there are GNU or pax extension entries? */ data = __archive_read_ahead(a, (size_t)size, NULL); if (data == NULL) { *unconsumed = 0; return (ARCHIVE_FATAL); } archive_entry_copy_mac_metadata(entry, data, (size_t)size); *unconsumed = (size_t)((size + 511) & ~ 511); tar_flush_unconsumed(a, unconsumed); return (tar_read_header(a, tar, entry, unconsumed)); } /* * Parse a file header for a pax extended archive entry. */ static int header_pax_global(struct archive_read *a, struct tar *tar, struct archive_entry *entry, const void *h, size_t *unconsumed) { int err; err = read_body_to_string(a, tar, &(tar->pax_global), h, unconsumed); if (err != ARCHIVE_OK) return (err); err = tar_read_header(a, tar, entry, unconsumed); return (err); } static int header_pax_extensions(struct archive_read *a, struct tar *tar, struct archive_entry *entry, const void *h, size_t *unconsumed) { int err, err2; err = read_body_to_string(a, tar, &(tar->pax_header), h, unconsumed); if (err != ARCHIVE_OK) return (err); /* Parse the next header. */ err = tar_read_header(a, tar, entry, unconsumed); if ((err != ARCHIVE_OK) && (err != ARCHIVE_WARN)) return (err); /* * TODO: Parse global/default options into 'entry' struct here * before handling file-specific options. * * This design (parse standard header, then overwrite with pax * extended attribute data) usually works well, but isn't ideal; * it would be better to parse the pax extended attributes first * and then skip any fields in the standard header that were * defined in the pax header. */ err2 = pax_header(a, tar, entry, &tar->pax_header); err = err_combine(err, err2); tar->entry_padding = 0x1ff & (-tar->entry_bytes_remaining); return (err); } /* * Parse a file header for a Posix "ustar" archive entry. This also * handles "pax" or "extended ustar" entries. */ static int header_ustar(struct archive_read *a, struct tar *tar, struct archive_entry *entry, const void *h) { const struct archive_entry_header_ustar *header; struct archive_string *as; int err = ARCHIVE_OK, r; header = (const struct archive_entry_header_ustar *)h; /* Copy name into an internal buffer to ensure null-termination. */ as = &(tar->entry_pathname); if (header->prefix[0]) { archive_strncpy(as, header->prefix, sizeof(header->prefix)); if (as->s[archive_strlen(as) - 1] != '/') archive_strappend_char(as, '/'); archive_strncat(as, header->name, sizeof(header->name)); } else { archive_strncpy(as, header->name, sizeof(header->name)); } if (archive_entry_copy_pathname_l(entry, as->s, archive_strlen(as), tar->sconv) != 0) { err = set_conversion_failed_error(a, tar->sconv, "Pathname"); if (err == ARCHIVE_FATAL) return (err); } /* Handle rest of common fields. */ r = header_common(a, tar, entry, h); if (r == ARCHIVE_FATAL) return (r); if (r < err) err = r; /* Handle POSIX ustar fields. */ if (archive_entry_copy_uname_l(entry, header->uname, sizeof(header->uname), tar->sconv) != 0) { err = set_conversion_failed_error(a, tar->sconv, "Uname"); if (err == ARCHIVE_FATAL) return (err); } if (archive_entry_copy_gname_l(entry, header->gname, sizeof(header->gname), tar->sconv) != 0) { err = set_conversion_failed_error(a, tar->sconv, "Gname"); if (err == ARCHIVE_FATAL) return (err); } /* Parse out device numbers only for char and block specials. */ if (header->typeflag[0] == '3' || header->typeflag[0] == '4') { archive_entry_set_rdevmajor(entry, (dev_t) tar_atol(header->rdevmajor, sizeof(header->rdevmajor))); archive_entry_set_rdevminor(entry, (dev_t) tar_atol(header->rdevminor, sizeof(header->rdevminor))); } tar->entry_padding = 0x1ff & (-tar->entry_bytes_remaining); return (err); } /* * Parse the pax extended attributes record. * * Returns non-zero if there's an error in the data. */ static int pax_header(struct archive_read *a, struct tar *tar, struct archive_entry *entry, struct archive_string *in_as) { size_t attr_length, l, line_length, value_length; char *p; char *key, *value; struct archive_string *as; struct archive_string_conv *sconv; int err, err2; char *attr = in_as->s; attr_length = in_as->length; tar->pax_hdrcharset_binary = 0; archive_string_empty(&(tar->entry_gname)); archive_string_empty(&(tar->entry_linkpath)); archive_string_empty(&(tar->entry_pathname)); archive_string_empty(&(tar->entry_pathname_override)); archive_string_empty(&(tar->entry_uname)); err = ARCHIVE_OK; while (attr_length > 0) { /* Parse decimal length field at start of line. */ line_length = 0; l = attr_length; p = attr; /* Record start of line. */ while (l>0) { if (*p == ' ') { p++; l--; break; } if (*p < '0' || *p > '9') { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Ignoring malformed pax extended attributes"); return (ARCHIVE_WARN); } line_length *= 10; line_length += *p - '0'; if (line_length > 999999) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Rejecting pax extended attribute > 1MB"); return (ARCHIVE_WARN); } p++; l--; } /* * Parsed length must be no bigger than available data, * at least 1, and the last character of the line must * be '\n'. */ if (line_length > attr_length || line_length < 1 || attr[line_length - 1] != '\n') { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Ignoring malformed pax extended attribute"); return (ARCHIVE_WARN); } /* Null-terminate the line. */ attr[line_length - 1] = '\0'; /* Find end of key and null terminate it. */ key = p; if (key[0] == '=') return (-1); while (*p && *p != '=') ++p; if (*p == '\0') { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Invalid pax extended attributes"); return (ARCHIVE_WARN); } *p = '\0'; value = p + 1; /* Some values may be binary data */ value_length = attr + line_length - 1 - value; /* Identify this attribute and set it in the entry. */ err2 = pax_attribute(a, tar, entry, key, value, value_length); if (err2 == ARCHIVE_FATAL) return (err2); err = err_combine(err, err2); /* Skip to next line */ attr += line_length; attr_length -= line_length; } /* * PAX format uses UTF-8 as default charset for its metadata * unless hdrcharset=BINARY is present in its header. * We apply the charset specified by the hdrcharset option only * when the hdrcharset attribute(in PAX header) is BINARY because * we respect the charset described in PAX header and BINARY also * means that metadata(filename,uname and gname) character-set * is unknown. */ if (tar->pax_hdrcharset_binary) sconv = tar->opt_sconv; else { sconv = archive_string_conversion_from_charset( &(a->archive), "UTF-8", 1); if (sconv == NULL) return (ARCHIVE_FATAL); if (tar->compat_2x) archive_string_conversion_set_opt(sconv, SCONV_SET_OPT_UTF8_LIBARCHIVE2X); } if (archive_strlen(&(tar->entry_gname)) > 0) { if (archive_entry_copy_gname_l(entry, tar->entry_gname.s, archive_strlen(&(tar->entry_gname)), sconv) != 0) { err = set_conversion_failed_error(a, sconv, "Gname"); if (err == ARCHIVE_FATAL) return (err); /* Use a converted an original name. */ archive_entry_copy_gname(entry, tar->entry_gname.s); } } if (archive_strlen(&(tar->entry_linkpath)) > 0) { if (archive_entry_copy_link_l(entry, tar->entry_linkpath.s, archive_strlen(&(tar->entry_linkpath)), sconv) != 0) { err = set_conversion_failed_error(a, sconv, "Linkname"); if (err == ARCHIVE_FATAL) return (err); /* Use a converted an original name. */ archive_entry_copy_link(entry, tar->entry_linkpath.s); } } /* * Some extensions (such as the GNU sparse file extensions) * deliberately store a synthetic name under the regular 'path' * attribute and the real file name under a different attribute. * Since we're supposed to not care about the order, we * have no choice but to store all of the various filenames * we find and figure it all out afterwards. This is the * figuring out part. */ as = NULL; if (archive_strlen(&(tar->entry_pathname_override)) > 0) as = &(tar->entry_pathname_override); else if (archive_strlen(&(tar->entry_pathname)) > 0) as = &(tar->entry_pathname); if (as != NULL) { if (archive_entry_copy_pathname_l(entry, as->s, archive_strlen(as), sconv) != 0) { err = set_conversion_failed_error(a, sconv, "Pathname"); if (err == ARCHIVE_FATAL) return (err); /* Use a converted an original name. */ archive_entry_copy_pathname(entry, as->s); } } if (archive_strlen(&(tar->entry_uname)) > 0) { if (archive_entry_copy_uname_l(entry, tar->entry_uname.s, archive_strlen(&(tar->entry_uname)), sconv) != 0) { err = set_conversion_failed_error(a, sconv, "Uname"); if (err == ARCHIVE_FATAL) return (err); /* Use a converted an original name. */ archive_entry_copy_uname(entry, tar->entry_uname.s); } } return (err); } static int pax_attribute_xattr(struct archive_entry *entry, const char *name, const char *value) { char *name_decoded; void *value_decoded; size_t value_len; if (strlen(name) < 18 || (memcmp(name, "LIBARCHIVE.xattr.", 17)) != 0) return 3; name += 17; /* URL-decode name */ name_decoded = url_decode(name); if (name_decoded == NULL) return 2; /* Base-64 decode value */ value_decoded = base64_decode(value, strlen(value), &value_len); if (value_decoded == NULL) { free(name_decoded); return 1; } archive_entry_xattr_add_entry(entry, name_decoded, value_decoded, value_len); free(name_decoded); free(value_decoded); return 0; } static int pax_attribute_schily_xattr(struct archive_entry *entry, const char *name, const char *value, size_t value_length) { if (strlen(name) < 14 || (memcmp(name, "SCHILY.xattr.", 13)) != 0) return 1; name += 13; archive_entry_xattr_add_entry(entry, name, value, value_length); return 0; } static int pax_attribute_acl(struct archive_read *a, struct tar *tar, struct archive_entry *entry, const char *value, int type) { int r; const char* errstr; switch (type) { case ARCHIVE_ENTRY_ACL_TYPE_ACCESS: errstr = "SCHILY.acl.access"; break; case ARCHIVE_ENTRY_ACL_TYPE_DEFAULT: errstr = "SCHILY.acl.default"; break; case ARCHIVE_ENTRY_ACL_TYPE_NFS4: errstr = "SCHILY.acl.ace"; break; default: archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Unknown ACL type: %d", type); return(ARCHIVE_FATAL); } if (tar->sconv_acl == NULL) { tar->sconv_acl = archive_string_conversion_from_charset( &(a->archive), "UTF-8", 1); if (tar->sconv_acl == NULL) return (ARCHIVE_FATAL); } r = archive_acl_from_text_l(archive_entry_acl(entry), value, type, tar->sconv_acl); if (r != ARCHIVE_OK) { if (r == ARCHIVE_FATAL) { archive_set_error(&a->archive, ENOMEM, "%s %s", "Can't allocate memory for ", errstr); return (r); } archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "%s %s", "Parse error: ", errstr); } return (r); } /* * Parse a single key=value attribute. key/value pointers are * assumed to point into reasonably long-lived storage. * * Note that POSIX reserves all-lowercase keywords. Vendor-specific * extensions should always have keywords of the form "VENDOR.attribute" * In particular, it's quite feasible to support many different * vendor extensions here. I'm using "LIBARCHIVE" for extensions * unique to this library. * * Investigate other vendor-specific extensions and see if * any of them look useful. */ static int pax_attribute(struct archive_read *a, struct tar *tar, struct archive_entry *entry, const char *key, const char *value, size_t value_length) { int64_t s; long n; int err = ARCHIVE_OK, r; #ifndef __FreeBSD__ if (value == NULL) value = ""; /* Disable compiler warning; do not pass * NULL pointer to strlen(). */ #endif switch (key[0]) { case 'G': /* Reject GNU.sparse.* headers on non-regular files. */ if (strncmp(key, "GNU.sparse", 10) == 0 && !tar->sparse_allowed) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Non-regular file cannot be sparse"); return (ARCHIVE_FATAL); } /* GNU "0.0" sparse pax format. */ if (strcmp(key, "GNU.sparse.numblocks") == 0) { tar->sparse_offset = -1; tar->sparse_numbytes = -1; tar->sparse_gnu_major = 0; tar->sparse_gnu_minor = 0; } if (strcmp(key, "GNU.sparse.offset") == 0) { tar->sparse_offset = tar_atol10(value, strlen(value)); if (tar->sparse_numbytes != -1) { if (gnu_add_sparse_entry(a, tar, tar->sparse_offset, tar->sparse_numbytes) != ARCHIVE_OK) return (ARCHIVE_FATAL); tar->sparse_offset = -1; tar->sparse_numbytes = -1; } } if (strcmp(key, "GNU.sparse.numbytes") == 0) { tar->sparse_numbytes = tar_atol10(value, strlen(value)); if (tar->sparse_numbytes != -1) { if (gnu_add_sparse_entry(a, tar, tar->sparse_offset, tar->sparse_numbytes) != ARCHIVE_OK) return (ARCHIVE_FATAL); tar->sparse_offset = -1; tar->sparse_numbytes = -1; } } if (strcmp(key, "GNU.sparse.size") == 0) { tar->realsize = tar_atol10(value, strlen(value)); archive_entry_set_size(entry, tar->realsize); tar->realsize_override = 1; } /* GNU "0.1" sparse pax format. */ if (strcmp(key, "GNU.sparse.map") == 0) { tar->sparse_gnu_major = 0; tar->sparse_gnu_minor = 1; if (gnu_sparse_01_parse(a, tar, value) != ARCHIVE_OK) return (ARCHIVE_WARN); } /* GNU "1.0" sparse pax format */ if (strcmp(key, "GNU.sparse.major") == 0) { tar->sparse_gnu_major = (int)tar_atol10(value, strlen(value)); tar->sparse_gnu_pending = 1; } if (strcmp(key, "GNU.sparse.minor") == 0) { tar->sparse_gnu_minor = (int)tar_atol10(value, strlen(value)); tar->sparse_gnu_pending = 1; } if (strcmp(key, "GNU.sparse.name") == 0) { /* * The real filename; when storing sparse * files, GNU tar puts a synthesized name into * the regular 'path' attribute in an attempt * to limit confusion. ;-) */ archive_strcpy(&(tar->entry_pathname_override), value); } if (strcmp(key, "GNU.sparse.realsize") == 0) { tar->realsize = tar_atol10(value, strlen(value)); archive_entry_set_size(entry, tar->realsize); tar->realsize_override = 1; } break; case 'L': /* Our extensions */ /* TODO: Handle arbitrary extended attributes... */ /* if (strcmp(key, "LIBARCHIVE.xxxxxxx") == 0) archive_entry_set_xxxxxx(entry, value); */ if (strcmp(key, "LIBARCHIVE.creationtime") == 0) { pax_time(value, &s, &n); archive_entry_set_birthtime(entry, s, n); } if (memcmp(key, "LIBARCHIVE.xattr.", 17) == 0) pax_attribute_xattr(entry, key, value); break; case 'S': /* We support some keys used by the "star" archiver */ if (strcmp(key, "SCHILY.acl.access") == 0) { r = pax_attribute_acl(a, tar, entry, value, ARCHIVE_ENTRY_ACL_TYPE_ACCESS); if (r == ARCHIVE_FATAL) return (r); } else if (strcmp(key, "SCHILY.acl.default") == 0) { r = pax_attribute_acl(a, tar, entry, value, ARCHIVE_ENTRY_ACL_TYPE_DEFAULT); if (r == ARCHIVE_FATAL) return (r); } else if (strcmp(key, "SCHILY.acl.ace") == 0) { r = pax_attribute_acl(a, tar, entry, value, ARCHIVE_ENTRY_ACL_TYPE_NFS4); if (r == ARCHIVE_FATAL) return (r); } else if (strcmp(key, "SCHILY.devmajor") == 0) { archive_entry_set_rdevmajor(entry, (dev_t)tar_atol10(value, strlen(value))); } else if (strcmp(key, "SCHILY.devminor") == 0) { archive_entry_set_rdevminor(entry, (dev_t)tar_atol10(value, strlen(value))); } else if (strcmp(key, "SCHILY.fflags") == 0) { archive_entry_copy_fflags_text(entry, value); } else if (strcmp(key, "SCHILY.dev") == 0) { archive_entry_set_dev(entry, (dev_t)tar_atol10(value, strlen(value))); } else if (strcmp(key, "SCHILY.ino") == 0) { archive_entry_set_ino(entry, tar_atol10(value, strlen(value))); } else if (strcmp(key, "SCHILY.nlink") == 0) { archive_entry_set_nlink(entry, (unsigned) tar_atol10(value, strlen(value))); } else if (strcmp(key, "SCHILY.realsize") == 0) { tar->realsize = tar_atol10(value, strlen(value)); tar->realsize_override = 1; archive_entry_set_size(entry, tar->realsize); } else if (strncmp(key, "SCHILY.xattr.", 13) == 0) { pax_attribute_schily_xattr(entry, key, value, value_length); } else if (strcmp(key, "SUN.holesdata") == 0) { /* A Solaris extension for sparse. */ r = solaris_sparse_parse(a, tar, entry, value); if (r < err) { if (r == ARCHIVE_FATAL) return (r); err = r; archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Parse error: SUN.holesdata"); } } break; case 'a': if (strcmp(key, "atime") == 0) { pax_time(value, &s, &n); archive_entry_set_atime(entry, s, n); } break; case 'c': if (strcmp(key, "ctime") == 0) { pax_time(value, &s, &n); archive_entry_set_ctime(entry, s, n); } else if (strcmp(key, "charset") == 0) { /* TODO: Publish charset information in entry. */ } else if (strcmp(key, "comment") == 0) { /* TODO: Publish comment in entry. */ } break; case 'g': if (strcmp(key, "gid") == 0) { archive_entry_set_gid(entry, tar_atol10(value, strlen(value))); } else if (strcmp(key, "gname") == 0) { archive_strcpy(&(tar->entry_gname), value); } break; case 'h': if (strcmp(key, "hdrcharset") == 0) { if (strcmp(value, "BINARY") == 0) /* Binary mode. */ tar->pax_hdrcharset_binary = 1; else if (strcmp(value, "ISO-IR 10646 2000 UTF-8") == 0) tar->pax_hdrcharset_binary = 0; } break; case 'l': /* pax interchange doesn't distinguish hardlink vs. symlink. */ if (strcmp(key, "linkpath") == 0) { archive_strcpy(&(tar->entry_linkpath), value); } break; case 'm': if (strcmp(key, "mtime") == 0) { pax_time(value, &s, &n); archive_entry_set_mtime(entry, s, n); } break; case 'p': if (strcmp(key, "path") == 0) { archive_strcpy(&(tar->entry_pathname), value); } break; case 'r': /* POSIX has reserved 'realtime.*' */ break; case 's': /* POSIX has reserved 'security.*' */ /* Someday: if (strcmp(key, "security.acl") == 0) { ... } */ if (strcmp(key, "size") == 0) { /* "size" is the size of the data in the entry. */ tar->entry_bytes_remaining = tar_atol10(value, strlen(value)); /* * The "size" pax header keyword always overrides the * "size" field in the tar header. * GNU.sparse.realsize, GNU.sparse.size and * SCHILY.realsize override this value. */ if (!tar->realsize_override) { archive_entry_set_size(entry, tar->entry_bytes_remaining); tar->realsize = tar->entry_bytes_remaining; } } break; case 'u': if (strcmp(key, "uid") == 0) { archive_entry_set_uid(entry, tar_atol10(value, strlen(value))); } else if (strcmp(key, "uname") == 0) { archive_strcpy(&(tar->entry_uname), value); } break; } return (err); } /* * parse a decimal time value, which may include a fractional portion */ static void pax_time(const char *p, int64_t *ps, long *pn) { char digit; int64_t s; unsigned long l; int sign; int64_t limit, last_digit_limit; limit = INT64_MAX / 10; last_digit_limit = INT64_MAX % 10; s = 0; sign = 1; if (*p == '-') { sign = -1; p++; } while (*p >= '0' && *p <= '9') { digit = *p - '0'; if (s > limit || (s == limit && digit > last_digit_limit)) { s = INT64_MAX; break; } s = (s * 10) + digit; ++p; } *ps = s * sign; /* Calculate nanoseconds. */ *pn = 0; if (*p != '.') return; l = 100000000UL; do { ++p; if (*p >= '0' && *p <= '9') *pn += (*p - '0') * l; else break; } while (l /= 10); } /* * Parse GNU tar header */ static int header_gnutar(struct archive_read *a, struct tar *tar, struct archive_entry *entry, const void *h, size_t *unconsumed) { const struct archive_entry_header_gnutar *header; int64_t t; int err = ARCHIVE_OK; /* * GNU header is like POSIX ustar, except 'prefix' is * replaced with some other fields. This also means the * filename is stored as in old-style archives. */ /* Grab fields common to all tar variants. */ err = header_common(a, tar, entry, h); if (err == ARCHIVE_FATAL) return (err); /* Copy filename over (to ensure null termination). */ header = (const struct archive_entry_header_gnutar *)h; if (archive_entry_copy_pathname_l(entry, header->name, sizeof(header->name), tar->sconv) != 0) { err = set_conversion_failed_error(a, tar->sconv, "Pathname"); if (err == ARCHIVE_FATAL) return (err); } /* Fields common to ustar and GNU */ /* XXX Can the following be factored out since it's common * to ustar and gnu tar? Is it okay to move it down into * header_common, perhaps? */ if (archive_entry_copy_uname_l(entry, header->uname, sizeof(header->uname), tar->sconv) != 0) { err = set_conversion_failed_error(a, tar->sconv, "Uname"); if (err == ARCHIVE_FATAL) return (err); } if (archive_entry_copy_gname_l(entry, header->gname, sizeof(header->gname), tar->sconv) != 0) { err = set_conversion_failed_error(a, tar->sconv, "Gname"); if (err == ARCHIVE_FATAL) return (err); } /* Parse out device numbers only for char and block specials */ if (header->typeflag[0] == '3' || header->typeflag[0] == '4') { archive_entry_set_rdevmajor(entry, (dev_t) tar_atol(header->rdevmajor, sizeof(header->rdevmajor))); archive_entry_set_rdevminor(entry, (dev_t) tar_atol(header->rdevminor, sizeof(header->rdevminor))); } else archive_entry_set_rdev(entry, 0); tar->entry_padding = 0x1ff & (-tar->entry_bytes_remaining); /* Grab GNU-specific fields. */ t = tar_atol(header->atime, sizeof(header->atime)); if (t > 0) archive_entry_set_atime(entry, t, 0); t = tar_atol(header->ctime, sizeof(header->ctime)); if (t > 0) archive_entry_set_ctime(entry, t, 0); if (header->realsize[0] != 0) { tar->realsize = tar_atol(header->realsize, sizeof(header->realsize)); archive_entry_set_size(entry, tar->realsize); tar->realsize_override = 1; } if (header->sparse[0].offset[0] != 0) { if (gnu_sparse_old_read(a, tar, header, unconsumed) != ARCHIVE_OK) return (ARCHIVE_FATAL); } else { if (header->isextended[0] != 0) { /* XXX WTF? XXX */ } } return (err); } static int gnu_add_sparse_entry(struct archive_read *a, struct tar *tar, int64_t offset, int64_t remaining) { struct sparse_block *p; p = (struct sparse_block *)calloc(1, sizeof(*p)); if (p == NULL) { archive_set_error(&a->archive, ENOMEM, "Out of memory"); return (ARCHIVE_FATAL); } if (tar->sparse_last != NULL) tar->sparse_last->next = p; else tar->sparse_list = p; tar->sparse_last = p; - if (remaining < 0 || offset < 0) { + if (remaining < 0 || offset < 0 || offset > INT64_MAX - remaining) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Malformed sparse map data"); return (ARCHIVE_FATAL); } p->offset = offset; p->remaining = remaining; return (ARCHIVE_OK); } static void gnu_clear_sparse_list(struct tar *tar) { struct sparse_block *p; while (tar->sparse_list != NULL) { p = tar->sparse_list; tar->sparse_list = p->next; free(p); } tar->sparse_last = NULL; } /* * GNU tar old-format sparse data. * * GNU old-format sparse data is stored in a fixed-field * format. Offset/size values are 11-byte octal fields (same * format as 'size' field in ustart header). These are * stored in the header, allocating subsequent header blocks * as needed. Extending the header in this way is a pretty * severe POSIX violation; this design has earned GNU tar a * lot of criticism. */ static int gnu_sparse_old_read(struct archive_read *a, struct tar *tar, const struct archive_entry_header_gnutar *header, size_t *unconsumed) { ssize_t bytes_read; const void *data; struct extended { struct gnu_sparse sparse[21]; char isextended[1]; char padding[7]; }; const struct extended *ext; if (gnu_sparse_old_parse(a, tar, header->sparse, 4) != ARCHIVE_OK) return (ARCHIVE_FATAL); if (header->isextended[0] == 0) return (ARCHIVE_OK); do { tar_flush_unconsumed(a, unconsumed); data = __archive_read_ahead(a, 512, &bytes_read); if (bytes_read < 0) return (ARCHIVE_FATAL); if (bytes_read < 512) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated tar archive " "detected while reading sparse file data"); return (ARCHIVE_FATAL); } *unconsumed = 512; ext = (const struct extended *)data; if (gnu_sparse_old_parse(a, tar, ext->sparse, 21) != ARCHIVE_OK) return (ARCHIVE_FATAL); } while (ext->isextended[0] != 0); if (tar->sparse_list != NULL) tar->entry_offset = tar->sparse_list->offset; return (ARCHIVE_OK); } static int gnu_sparse_old_parse(struct archive_read *a, struct tar *tar, const struct gnu_sparse *sparse, int length) { while (length > 0 && sparse->offset[0] != 0) { if (gnu_add_sparse_entry(a, tar, tar_atol(sparse->offset, sizeof(sparse->offset)), tar_atol(sparse->numbytes, sizeof(sparse->numbytes))) != ARCHIVE_OK) return (ARCHIVE_FATAL); sparse++; length--; } return (ARCHIVE_OK); } /* * GNU tar sparse format 0.0 * * Beginning with GNU tar 1.15, sparse files are stored using * information in the pax extended header. The GNU tar maintainers * have gone through a number of variations in the process of working * out this scheme; fortunately, they're all numbered. * * Sparse format 0.0 uses attribute GNU.sparse.numblocks to store the * number of blocks, and GNU.sparse.offset/GNU.sparse.numbytes to * store offset/size for each block. The repeated instances of these * latter fields violate the pax specification (which frowns on * duplicate keys), so this format was quickly replaced. */ /* * GNU tar sparse format 0.1 * * This version replaced the offset/numbytes attributes with * a single "map" attribute that stored a list of integers. This * format had two problems: First, the "map" attribute could be very * long, which caused problems for some implementations. More * importantly, the sparse data was lost when extracted by archivers * that didn't recognize this extension. */ static int gnu_sparse_01_parse(struct archive_read *a, struct tar *tar, const char *p) { const char *e; int64_t offset = -1, size = -1; for (;;) { e = p; while (*e != '\0' && *e != ',') { if (*e < '0' || *e > '9') return (ARCHIVE_WARN); e++; } if (offset < 0) { offset = tar_atol10(p, e - p); if (offset < 0) return (ARCHIVE_WARN); } else { size = tar_atol10(p, e - p); if (size < 0) return (ARCHIVE_WARN); if (gnu_add_sparse_entry(a, tar, offset, size) != ARCHIVE_OK) return (ARCHIVE_FATAL); offset = -1; } if (*e == '\0') return (ARCHIVE_OK); p = e + 1; } } /* * GNU tar sparse format 1.0 * * The idea: The offset/size data is stored as a series of base-10 * ASCII numbers prepended to the file data, so that dearchivers that * don't support this format will extract the block map along with the * data and a separate post-process can restore the sparseness. * * Unfortunately, GNU tar 1.16 had a bug that added unnecessary * padding to the body of the file when using this format. GNU tar * 1.17 corrected this bug without bumping the version number, so * it's not possible to support both variants. This code supports * the later variant at the expense of not supporting the former. * * This variant also replaced GNU.sparse.size with GNU.sparse.realsize * and introduced the GNU.sparse.major/GNU.sparse.minor attributes. */ /* * Read the next line from the input, and parse it as a decimal * integer followed by '\n'. Returns positive integer value or * negative on error. */ static int64_t gnu_sparse_10_atol(struct archive_read *a, struct tar *tar, int64_t *remaining, size_t *unconsumed) { int64_t l, limit, last_digit_limit; const char *p; ssize_t bytes_read; int base, digit; base = 10; limit = INT64_MAX / base; last_digit_limit = INT64_MAX % base; /* * Skip any lines starting with '#'; GNU tar specs * don't require this, but they should. */ do { bytes_read = readline(a, tar, &p, (ssize_t)tar_min(*remaining, 100), unconsumed); if (bytes_read <= 0) return (ARCHIVE_FATAL); *remaining -= bytes_read; } while (p[0] == '#'); l = 0; while (bytes_read > 0) { if (*p == '\n') return (l); if (*p < '0' || *p >= '0' + base) return (ARCHIVE_WARN); digit = *p - '0'; if (l > limit || (l == limit && digit > last_digit_limit)) l = INT64_MAX; /* Truncate on overflow. */ else l = (l * base) + digit; p++; bytes_read--; } /* TODO: Error message. */ return (ARCHIVE_WARN); } /* * Returns length (in bytes) of the sparse data description * that was read. */ static ssize_t gnu_sparse_10_read(struct archive_read *a, struct tar *tar, size_t *unconsumed) { ssize_t bytes_read; int entries; int64_t offset, size, to_skip, remaining; /* Clear out the existing sparse list. */ gnu_clear_sparse_list(tar); remaining = tar->entry_bytes_remaining; /* Parse entries. */ entries = (int)gnu_sparse_10_atol(a, tar, &remaining, unconsumed); if (entries < 0) return (ARCHIVE_FATAL); /* Parse the individual entries. */ while (entries-- > 0) { /* Parse offset/size */ offset = gnu_sparse_10_atol(a, tar, &remaining, unconsumed); if (offset < 0) return (ARCHIVE_FATAL); size = gnu_sparse_10_atol(a, tar, &remaining, unconsumed); if (size < 0) return (ARCHIVE_FATAL); /* Add a new sparse entry. */ if (gnu_add_sparse_entry(a, tar, offset, size) != ARCHIVE_OK) return (ARCHIVE_FATAL); } /* Skip rest of block... */ tar_flush_unconsumed(a, unconsumed); bytes_read = (ssize_t)(tar->entry_bytes_remaining - remaining); to_skip = 0x1ff & -bytes_read; /* Fail if tar->entry_bytes_remaing would get negative */ if (to_skip > remaining) return (ARCHIVE_FATAL); if (to_skip != __archive_read_consume(a, to_skip)) return (ARCHIVE_FATAL); return ((ssize_t)(bytes_read + to_skip)); } /* * Solaris pax extension for a sparse file. This is recorded with the * data and hole pairs. The way recording sparse information by Solaris' * pax simply indicates where data and sparse are, so the stored contents * consist of both data and hole. */ static int solaris_sparse_parse(struct archive_read *a, struct tar *tar, struct archive_entry *entry, const char *p) { const char *e; int64_t start, end; int hole = 1; (void)entry; /* UNUSED */ end = 0; if (*p == ' ') p++; else return (ARCHIVE_WARN); for (;;) { e = p; while (*e != '\0' && *e != ' ') { if (*e < '0' || *e > '9') return (ARCHIVE_WARN); e++; } start = end; end = tar_atol10(p, e - p); if (end < 0) return (ARCHIVE_WARN); if (start < end) { if (gnu_add_sparse_entry(a, tar, start, end - start) != ARCHIVE_OK) return (ARCHIVE_FATAL); tar->sparse_last->hole = hole; } if (*e == '\0') return (ARCHIVE_OK); p = e + 1; hole = hole == 0; } } /*- * Convert text->integer. * * Traditional tar formats (including POSIX) specify base-8 for * all of the standard numeric fields. This is a significant limitation * in practice: * = file size is limited to 8GB * = rdevmajor and rdevminor are limited to 21 bits * = uid/gid are limited to 21 bits * * There are two workarounds for this: * = pax extended headers, which use variable-length string fields * = GNU tar and STAR both allow either base-8 or base-256 in * most fields. The high bit is set to indicate base-256. * * On read, this implementation supports both extensions. */ static int64_t tar_atol(const char *p, size_t char_cnt) { /* * Technically, GNU tar considers a field to be in base-256 * only if the first byte is 0xff or 0x80. */ if (*p & 0x80) return (tar_atol256(p, char_cnt)); return (tar_atol8(p, char_cnt)); } /* * Note that this implementation does not (and should not!) obey * locale settings; you cannot simply substitute strtol here, since * it does obey locale. */ static int64_t tar_atol_base_n(const char *p, size_t char_cnt, int base) { int64_t l, maxval, limit, last_digit_limit; int digit, sign; maxval = INT64_MAX; limit = INT64_MAX / base; last_digit_limit = INT64_MAX % base; /* the pointer will not be dereferenced if char_cnt is zero * due to the way the && operator is evaluated. */ while (char_cnt != 0 && (*p == ' ' || *p == '\t')) { p++; char_cnt--; } sign = 1; if (char_cnt != 0 && *p == '-') { sign = -1; p++; char_cnt--; maxval = INT64_MIN; limit = -(INT64_MIN / base); last_digit_limit = INT64_MIN % base; } l = 0; if (char_cnt != 0) { digit = *p - '0'; while (digit >= 0 && digit < base && char_cnt != 0) { if (l>limit || (l == limit && digit > last_digit_limit)) { return maxval; /* Truncate on overflow. */ } l = (l * base) + digit; digit = *++p - '0'; char_cnt--; } } return (sign < 0) ? -l : l; } static int64_t tar_atol8(const char *p, size_t char_cnt) { return tar_atol_base_n(p, char_cnt, 8); } static int64_t tar_atol10(const char *p, size_t char_cnt) { return tar_atol_base_n(p, char_cnt, 10); } /* * Parse a base-256 integer. This is just a variable-length * twos-complement signed binary value in big-endian order, except * that the high-order bit is ignored. The values here can be up to * 12 bytes, so we need to be careful about overflowing 64-bit * (8-byte) integers. * * This code unashamedly assumes that the local machine uses 8-bit * bytes and twos-complement arithmetic. */ static int64_t tar_atol256(const char *_p, size_t char_cnt) { uint64_t l; const unsigned char *p = (const unsigned char *)_p; unsigned char c, neg; /* Extend 7-bit 2s-comp to 8-bit 2s-comp, decide sign. */ c = *p; if (c & 0x40) { neg = 0xff; c |= 0x80; l = ~ARCHIVE_LITERAL_ULL(0); } else { neg = 0; c &= 0x7f; l = 0; } /* If more than 8 bytes, check that we can ignore * high-order bits without overflow. */ while (char_cnt > sizeof(int64_t)) { --char_cnt; if (c != neg) return neg ? INT64_MIN : INT64_MAX; c = *++p; } /* c is first byte that fits; if sign mismatch, return overflow */ if ((c ^ neg) & 0x80) { return neg ? INT64_MIN : INT64_MAX; } /* Accumulate remaining bytes. */ while (--char_cnt > 0) { l = (l << 8) | c; c = *++p; } l = (l << 8) | c; /* Return signed twos-complement value. */ return (int64_t)(l); } /* * Returns length of line (including trailing newline) * or negative on error. 'start' argument is updated to * point to first character of line. This avoids copying * when possible. */ static ssize_t readline(struct archive_read *a, struct tar *tar, const char **start, ssize_t limit, size_t *unconsumed) { ssize_t bytes_read; ssize_t total_size = 0; const void *t; const char *s; void *p; tar_flush_unconsumed(a, unconsumed); t = __archive_read_ahead(a, 1, &bytes_read); if (bytes_read <= 0) return (ARCHIVE_FATAL); s = t; /* Start of line? */ p = memchr(t, '\n', bytes_read); /* If we found '\n' in the read buffer, return pointer to that. */ if (p != NULL) { bytes_read = 1 + ((const char *)p) - s; if (bytes_read > limit) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Line too long"); return (ARCHIVE_FATAL); } *unconsumed = bytes_read; *start = s; return (bytes_read); } *unconsumed = bytes_read; /* Otherwise, we need to accumulate in a line buffer. */ for (;;) { if (total_size + bytes_read > limit) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Line too long"); return (ARCHIVE_FATAL); } if (archive_string_ensure(&tar->line, total_size + bytes_read) == NULL) { archive_set_error(&a->archive, ENOMEM, "Can't allocate working buffer"); return (ARCHIVE_FATAL); } memcpy(tar->line.s + total_size, t, bytes_read); tar_flush_unconsumed(a, unconsumed); total_size += bytes_read; /* If we found '\n', clean up and return. */ if (p != NULL) { *start = tar->line.s; return (total_size); } /* Read some more. */ t = __archive_read_ahead(a, 1, &bytes_read); if (bytes_read <= 0) return (ARCHIVE_FATAL); s = t; /* Start of line? */ p = memchr(t, '\n', bytes_read); /* If we found '\n', trim the read. */ if (p != NULL) { bytes_read = 1 + ((const char *)p) - s; } *unconsumed = bytes_read; } } /* * base64_decode - Base64 decode * * This accepts most variations of base-64 encoding, including: * * with or without line breaks * * with or without the final group padded with '=' or '_' characters * (The most economical Base-64 variant does not pad the last group and * omits line breaks; RFC1341 used for MIME requires both.) */ static char * base64_decode(const char *s, size_t len, size_t *out_len) { static const unsigned char digits[64] = { 'A','B','C','D','E','F','G','H','I','J','K','L','M','N', 'O','P','Q','R','S','T','U','V','W','X','Y','Z','a','b', 'c','d','e','f','g','h','i','j','k','l','m','n','o','p', 'q','r','s','t','u','v','w','x','y','z','0','1','2','3', '4','5','6','7','8','9','+','/' }; static unsigned char decode_table[128]; char *out, *d; const unsigned char *src = (const unsigned char *)s; /* If the decode table is not yet initialized, prepare it. */ if (decode_table[digits[1]] != 1) { unsigned i; memset(decode_table, 0xff, sizeof(decode_table)); for (i = 0; i < sizeof(digits); i++) decode_table[digits[i]] = i; } /* Allocate enough space to hold the entire output. */ /* Note that we may not use all of this... */ out = (char *)malloc(len - len / 4 + 1); if (out == NULL) { *out_len = 0; return (NULL); } d = out; while (len > 0) { /* Collect the next group of (up to) four characters. */ int v = 0; int group_size = 0; while (group_size < 4 && len > 0) { /* '=' or '_' padding indicates final group. */ if (*src == '=' || *src == '_') { len = 0; break; } /* Skip illegal characters (including line breaks) */ if (*src > 127 || *src < 32 || decode_table[*src] == 0xff) { len--; src++; continue; } v <<= 6; v |= decode_table[*src++]; len --; group_size++; } /* Align a short group properly. */ v <<= 6 * (4 - group_size); /* Unpack the group we just collected. */ switch (group_size) { case 4: d[2] = v & 0xff; /* FALLTHROUGH */ case 3: d[1] = (v >> 8) & 0xff; /* FALLTHROUGH */ case 2: d[0] = (v >> 16) & 0xff; break; case 1: /* this is invalid! */ break; } d += group_size * 3 / 4; } *out_len = d - out; return (out); } static char * url_decode(const char *in) { char *out, *d; const char *s; out = (char *)malloc(strlen(in) + 1); if (out == NULL) return (NULL); for (s = in, d = out; *s != '\0'; ) { if (s[0] == '%' && s[1] != '\0' && s[2] != '\0') { /* Try to convert % escape */ int digit1 = tohex(s[1]); int digit2 = tohex(s[2]); if (digit1 >= 0 && digit2 >= 0) { /* Looks good, consume three chars */ s += 3; /* Convert output */ *d++ = ((digit1 << 4) | digit2); continue; } /* Else fall through and treat '%' as normal char */ } *d++ = *s++; } *d = '\0'; return (out); } static int tohex(int c) { if (c >= '0' && c <= '9') return (c - '0'); else if (c >= 'A' && c <= 'F') return (c - 'A' + 10); else if (c >= 'a' && c <= 'f') return (c - 'a' + 10); else return (-1); } Index: stable/10/contrib/libarchive/libarchive/archive_read_support_format_xar.c =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_read_support_format_xar.c (revision 324417) +++ stable/10/contrib/libarchive/libarchive/archive_read_support_format_xar.c (revision 324418) @@ -1,3287 +1,3301 @@ /*- * Copyright (c) 2009 Michihiro NAKAJIMA * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "archive_platform.h" __FBSDID("$FreeBSD$"); #ifdef HAVE_ERRNO_H #include #endif #ifdef HAVE_STDLIB_H #include #endif #if HAVE_LIBXML_XMLREADER_H #include #elif HAVE_BSDXML_H #include #elif HAVE_EXPAT_H #include #endif #ifdef HAVE_BZLIB_H #include #endif #if HAVE_LZMA_H #include #endif #ifdef HAVE_ZLIB_H #include #endif #include "archive.h" #include "archive_digest_private.h" #include "archive_endian.h" #include "archive_entry.h" #include "archive_entry_locale.h" #include "archive_private.h" #include "archive_read_private.h" #if (!defined(HAVE_LIBXML_XMLREADER_H) && \ !defined(HAVE_BSDXML_H) && !defined(HAVE_EXPAT_H)) ||\ !defined(HAVE_ZLIB_H) || \ !defined(ARCHIVE_HAS_MD5) || !defined(ARCHIVE_HAS_SHA1) /* * xar needs several external libraries. * o libxml2 or expat --- XML parser * o openssl or MD5/SHA1 hash function * o zlib * o bzlib2 (option) * o liblzma (option) */ int archive_read_support_format_xar(struct archive *_a) { struct archive_read *a = (struct archive_read *)_a; archive_check_magic(_a, ARCHIVE_READ_MAGIC, ARCHIVE_STATE_NEW, "archive_read_support_format_xar"); archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Xar not supported on this platform"); return (ARCHIVE_WARN); } #else /* Support xar format */ /* #define DEBUG 1 */ /* #define DEBUG_PRINT_TOC 1 */ #if DEBUG_PRINT_TOC #define PRINT_TOC(d, outbytes) do { \ unsigned char *x = (unsigned char *)(uintptr_t)d; \ unsigned char c = x[outbytes-1]; \ x[outbytes - 1] = 0; \ fprintf(stderr, "%s", x); \ fprintf(stderr, "%c", c); \ x[outbytes - 1] = c; \ } while (0) #else #define PRINT_TOC(d, outbytes) #endif #define HEADER_MAGIC 0x78617221 #define HEADER_SIZE 28 #define HEADER_VERSION 1 #define CKSUM_NONE 0 #define CKSUM_SHA1 1 #define CKSUM_MD5 2 #define MD5_SIZE 16 #define SHA1_SIZE 20 #define MAX_SUM_SIZE 20 enum enctype { NONE, GZIP, BZIP2, LZMA, XZ, }; struct chksumval { int alg; size_t len; unsigned char val[MAX_SUM_SIZE]; }; struct chksumwork { int alg; #ifdef ARCHIVE_HAS_MD5 archive_md5_ctx md5ctx; #endif #ifdef ARCHIVE_HAS_SHA1 archive_sha1_ctx sha1ctx; #endif }; struct xattr { struct xattr *next; struct archive_string name; uint64_t id; uint64_t length; uint64_t offset; uint64_t size; enum enctype encoding; struct chksumval a_sum; struct chksumval e_sum; struct archive_string fstype; }; struct xar_file { struct xar_file *next; struct xar_file *hdnext; struct xar_file *parent; int subdirs; unsigned int has; #define HAS_DATA 0x00001 #define HAS_PATHNAME 0x00002 #define HAS_SYMLINK 0x00004 #define HAS_TIME 0x00008 #define HAS_UID 0x00010 #define HAS_GID 0x00020 #define HAS_MODE 0x00040 #define HAS_TYPE 0x00080 #define HAS_DEV 0x00100 #define HAS_DEVMAJOR 0x00200 #define HAS_DEVMINOR 0x00400 #define HAS_INO 0x00800 #define HAS_FFLAGS 0x01000 #define HAS_XATTR 0x02000 #define HAS_ACL 0x04000 uint64_t id; uint64_t length; uint64_t offset; uint64_t size; enum enctype encoding; struct chksumval a_sum; struct chksumval e_sum; struct archive_string pathname; struct archive_string symlink; time_t ctime; time_t mtime; time_t atime; struct archive_string uname; int64_t uid; struct archive_string gname; int64_t gid; mode_t mode; dev_t dev; dev_t devmajor; dev_t devminor; int64_t ino64; struct archive_string fflags_text; unsigned int link; unsigned int nlink; struct archive_string hardlink; struct xattr *xattr_list; }; struct hdlink { struct hdlink *next; unsigned int id; int cnt; struct xar_file *files; }; struct heap_queue { struct xar_file **files; int allocated; int used; }; enum xmlstatus { INIT, XAR, TOC, TOC_CREATION_TIME, TOC_CHECKSUM, TOC_CHECKSUM_OFFSET, TOC_CHECKSUM_SIZE, TOC_FILE, FILE_DATA, FILE_DATA_LENGTH, FILE_DATA_OFFSET, FILE_DATA_SIZE, FILE_DATA_ENCODING, FILE_DATA_A_CHECKSUM, FILE_DATA_E_CHECKSUM, FILE_DATA_CONTENT, FILE_EA, FILE_EA_LENGTH, FILE_EA_OFFSET, FILE_EA_SIZE, FILE_EA_ENCODING, FILE_EA_A_CHECKSUM, FILE_EA_E_CHECKSUM, FILE_EA_NAME, FILE_EA_FSTYPE, FILE_CTIME, FILE_MTIME, FILE_ATIME, FILE_GROUP, FILE_GID, FILE_USER, FILE_UID, FILE_MODE, FILE_DEVICE, FILE_DEVICE_MAJOR, FILE_DEVICE_MINOR, FILE_DEVICENO, FILE_INODE, FILE_LINK, FILE_TYPE, FILE_NAME, FILE_ACL, FILE_ACL_DEFAULT, FILE_ACL_ACCESS, FILE_ACL_APPLEEXTENDED, /* BSD file flags. */ FILE_FLAGS, FILE_FLAGS_USER_NODUMP, FILE_FLAGS_USER_IMMUTABLE, FILE_FLAGS_USER_APPEND, FILE_FLAGS_USER_OPAQUE, FILE_FLAGS_USER_NOUNLINK, FILE_FLAGS_SYS_ARCHIVED, FILE_FLAGS_SYS_IMMUTABLE, FILE_FLAGS_SYS_APPEND, FILE_FLAGS_SYS_NOUNLINK, FILE_FLAGS_SYS_SNAPSHOT, /* Linux file flags. */ FILE_EXT2, FILE_EXT2_SecureDeletion, FILE_EXT2_Undelete, FILE_EXT2_Compress, FILE_EXT2_Synchronous, FILE_EXT2_Immutable, FILE_EXT2_AppendOnly, FILE_EXT2_NoDump, FILE_EXT2_NoAtime, FILE_EXT2_CompDirty, FILE_EXT2_CompBlock, FILE_EXT2_NoCompBlock, FILE_EXT2_CompError, FILE_EXT2_BTree, FILE_EXT2_HashIndexed, FILE_EXT2_iMagic, FILE_EXT2_Journaled, FILE_EXT2_NoTail, FILE_EXT2_DirSync, FILE_EXT2_TopDir, FILE_EXT2_Reserved, UNKNOWN, }; struct unknown_tag { struct unknown_tag *next; struct archive_string name; }; struct xar { uint64_t offset; /* Current position in the file. */ int64_t total; uint64_t h_base; int end_of_file; #define OUTBUFF_SIZE (1024 * 64) unsigned char *outbuff; enum xmlstatus xmlsts; enum xmlstatus xmlsts_unknown; struct unknown_tag *unknowntags; int base64text; /* * TOC */ uint64_t toc_remaining; uint64_t toc_total; uint64_t toc_chksum_offset; uint64_t toc_chksum_size; /* * For Decoding data. */ enum enctype rd_encoding; z_stream stream; int stream_valid; #if defined(HAVE_BZLIB_H) && defined(BZ_CONFIG_ERROR) bz_stream bzstream; int bzstream_valid; #endif #if HAVE_LZMA_H && HAVE_LIBLZMA lzma_stream lzstream; int lzstream_valid; #endif /* * For Checksum data. */ struct chksumwork a_sumwrk; struct chksumwork e_sumwrk; struct xar_file *file; /* current reading file. */ struct xattr *xattr; /* current reading extended attribute. */ struct heap_queue file_queue; struct xar_file *hdlink_orgs; struct hdlink *hdlink_list; int entry_init; uint64_t entry_total; uint64_t entry_remaining; size_t entry_unconsumed; uint64_t entry_size; enum enctype entry_encoding; struct chksumval entry_a_sum; struct chksumval entry_e_sum; struct archive_string_conv *sconv; }; struct xmlattr { struct xmlattr *next; char *name; char *value; }; struct xmlattr_list { struct xmlattr *first; struct xmlattr **last; }; static int xar_bid(struct archive_read *, int); static int xar_read_header(struct archive_read *, struct archive_entry *); static int xar_read_data(struct archive_read *, const void **, size_t *, int64_t *); static int xar_read_data_skip(struct archive_read *); static int xar_cleanup(struct archive_read *); static int move_reading_point(struct archive_read *, uint64_t); static int rd_contents_init(struct archive_read *, enum enctype, int, int); static int rd_contents(struct archive_read *, const void **, size_t *, size_t *, uint64_t); static uint64_t atol10(const char *, size_t); static int64_t atol8(const char *, size_t); static size_t atohex(unsigned char *, size_t, const char *, size_t); static time_t parse_time(const char *p, size_t n); static int heap_add_entry(struct archive_read *a, struct heap_queue *, struct xar_file *); static struct xar_file *heap_get_entry(struct heap_queue *); static int add_link(struct archive_read *, struct xar *, struct xar_file *); static void checksum_init(struct archive_read *, int, int); static void checksum_update(struct archive_read *, const void *, size_t, const void *, size_t); static int checksum_final(struct archive_read *, const void *, size_t, const void *, size_t); static void checksum_cleanup(struct archive_read *); static int decompression_init(struct archive_read *, enum enctype); static int decompress(struct archive_read *, const void **, size_t *, const void *, size_t *); static int decompression_cleanup(struct archive_read *); static void xmlattr_cleanup(struct xmlattr_list *); static int file_new(struct archive_read *, struct xar *, struct xmlattr_list *); static void file_free(struct xar_file *); static int xattr_new(struct archive_read *, struct xar *, struct xmlattr_list *); static void xattr_free(struct xattr *); static int getencoding(struct xmlattr_list *); static int getsumalgorithm(struct xmlattr_list *); static int unknowntag_start(struct archive_read *, struct xar *, const char *); static void unknowntag_end(struct xar *, const char *); static int xml_start(struct archive_read *, const char *, struct xmlattr_list *); static void xml_end(void *, const char *); static void xml_data(void *, const char *, int); static int xml_parse_file_flags(struct xar *, const char *); static int xml_parse_file_ext2(struct xar *, const char *); #if defined(HAVE_LIBXML_XMLREADER_H) static int xml2_xmlattr_setup(struct archive_read *, struct xmlattr_list *, xmlTextReaderPtr); static int xml2_read_cb(void *, char *, int); static int xml2_close_cb(void *); static void xml2_error_hdr(void *, const char *, xmlParserSeverities, xmlTextReaderLocatorPtr); static int xml2_read_toc(struct archive_read *); #elif defined(HAVE_BSDXML_H) || defined(HAVE_EXPAT_H) struct expat_userData { int state; struct archive_read *archive; }; static int expat_xmlattr_setup(struct archive_read *, struct xmlattr_list *, const XML_Char **); static void expat_start_cb(void *, const XML_Char *, const XML_Char **); static void expat_end_cb(void *, const XML_Char *); static void expat_data_cb(void *, const XML_Char *, int); static int expat_read_toc(struct archive_read *); #endif int archive_read_support_format_xar(struct archive *_a) { struct xar *xar; struct archive_read *a = (struct archive_read *)_a; int r; archive_check_magic(_a, ARCHIVE_READ_MAGIC, ARCHIVE_STATE_NEW, "archive_read_support_format_xar"); xar = (struct xar *)calloc(1, sizeof(*xar)); if (xar == NULL) { archive_set_error(&a->archive, ENOMEM, "Can't allocate xar data"); return (ARCHIVE_FATAL); } r = __archive_read_register_format(a, xar, "xar", xar_bid, NULL, xar_read_header, xar_read_data, xar_read_data_skip, NULL, xar_cleanup, NULL, NULL); if (r != ARCHIVE_OK) free(xar); return (r); } static int xar_bid(struct archive_read *a, int best_bid) { const unsigned char *b; int bid; (void)best_bid; /* UNUSED */ b = __archive_read_ahead(a, HEADER_SIZE, NULL); if (b == NULL) return (-1); bid = 0; /* * Verify magic code */ if (archive_be32dec(b) != HEADER_MAGIC) return (0); bid += 32; /* * Verify header size */ if (archive_be16dec(b+4) != HEADER_SIZE) return (0); bid += 16; /* * Verify header version */ if (archive_be16dec(b+6) != HEADER_VERSION) return (0); bid += 16; /* * Verify type of checksum */ switch (archive_be32dec(b+24)) { case CKSUM_NONE: case CKSUM_SHA1: case CKSUM_MD5: bid += 32; break; default: return (0); } return (bid); } static int read_toc(struct archive_read *a) { struct xar *xar; struct xar_file *file; const unsigned char *b; uint64_t toc_compressed_size; uint64_t toc_uncompressed_size; uint32_t toc_chksum_alg; ssize_t bytes; int r; xar = (struct xar *)(a->format->data); /* * Read xar header. */ b = __archive_read_ahead(a, HEADER_SIZE, &bytes); if (bytes < 0) return ((int)bytes); if (bytes < HEADER_SIZE) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated archive header"); return (ARCHIVE_FATAL); } if (archive_be32dec(b) != HEADER_MAGIC) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Invalid header magic"); return (ARCHIVE_FATAL); } if (archive_be16dec(b+6) != HEADER_VERSION) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Unsupported header version(%d)", archive_be16dec(b+6)); return (ARCHIVE_FATAL); } toc_compressed_size = archive_be64dec(b+8); xar->toc_remaining = toc_compressed_size; toc_uncompressed_size = archive_be64dec(b+16); toc_chksum_alg = archive_be32dec(b+24); __archive_read_consume(a, HEADER_SIZE); xar->offset += HEADER_SIZE; xar->toc_total = 0; /* * Read TOC(Table of Contents). */ /* Initialize reading contents. */ r = move_reading_point(a, HEADER_SIZE); if (r != ARCHIVE_OK) return (r); r = rd_contents_init(a, GZIP, toc_chksum_alg, CKSUM_NONE); if (r != ARCHIVE_OK) return (r); #ifdef HAVE_LIBXML_XMLREADER_H r = xml2_read_toc(a); #elif defined(HAVE_BSDXML_H) || defined(HAVE_EXPAT_H) r = expat_read_toc(a); #endif if (r != ARCHIVE_OK) return (r); /* Set 'The HEAP' base. */ xar->h_base = xar->offset; if (xar->toc_total != toc_uncompressed_size) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "TOC uncompressed size error"); return (ARCHIVE_FATAL); } /* * Checksum TOC */ if (toc_chksum_alg != CKSUM_NONE) { r = move_reading_point(a, xar->toc_chksum_offset); if (r != ARCHIVE_OK) return (r); b = __archive_read_ahead(a, (size_t)xar->toc_chksum_size, &bytes); if (bytes < 0) return ((int)bytes); if ((uint64_t)bytes < xar->toc_chksum_size) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated archive file"); return (ARCHIVE_FATAL); } r = checksum_final(a, b, (size_t)xar->toc_chksum_size, NULL, 0); __archive_read_consume(a, xar->toc_chksum_size); xar->offset += xar->toc_chksum_size; if (r != ARCHIVE_OK) return (ARCHIVE_FATAL); } /* * Connect hardlinked files. */ for (file = xar->hdlink_orgs; file != NULL; file = file->hdnext) { struct hdlink **hdlink; for (hdlink = &(xar->hdlink_list); *hdlink != NULL; hdlink = &((*hdlink)->next)) { if ((*hdlink)->id == file->id) { struct hdlink *hltmp; struct xar_file *f2; int nlink = (*hdlink)->cnt + 1; file->nlink = nlink; for (f2 = (*hdlink)->files; f2 != NULL; f2 = f2->hdnext) { f2->nlink = nlink; archive_string_copy( &(f2->hardlink), &(file->pathname)); } /* Remove resolved files from hdlist_list. */ hltmp = *hdlink; *hdlink = hltmp->next; free(hltmp); break; } } } a->archive.archive_format = ARCHIVE_FORMAT_XAR; a->archive.archive_format_name = "xar"; return (ARCHIVE_OK); } static int xar_read_header(struct archive_read *a, struct archive_entry *entry) { struct xar *xar; struct xar_file *file; struct xattr *xattr; int r; xar = (struct xar *)(a->format->data); r = ARCHIVE_OK; if (xar->offset == 0) { /* Create a character conversion object. */ if (xar->sconv == NULL) { xar->sconv = archive_string_conversion_from_charset( &(a->archive), "UTF-8", 1); if (xar->sconv == NULL) return (ARCHIVE_FATAL); } /* Read TOC. */ r = read_toc(a); if (r != ARCHIVE_OK) return (r); } for (;;) { file = xar->file = heap_get_entry(&(xar->file_queue)); if (file == NULL) { xar->end_of_file = 1; return (ARCHIVE_EOF); } if ((file->mode & AE_IFMT) != AE_IFDIR) break; if (file->has != (HAS_PATHNAME | HAS_TYPE)) break; /* * If a file type is a directory and it does not have * any metadata, do not export. */ file_free(file); } archive_entry_set_atime(entry, file->atime, 0); archive_entry_set_ctime(entry, file->ctime, 0); archive_entry_set_mtime(entry, file->mtime, 0); archive_entry_set_gid(entry, file->gid); if (file->gname.length > 0 && archive_entry_copy_gname_l(entry, file->gname.s, archive_strlen(&(file->gname)), xar->sconv) != 0) { if (errno == ENOMEM) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory for Gname"); return (ARCHIVE_FATAL); } archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Gname cannot be converted from %s to current locale.", archive_string_conversion_charset_name(xar->sconv)); r = ARCHIVE_WARN; } archive_entry_set_uid(entry, file->uid); if (file->uname.length > 0 && archive_entry_copy_uname_l(entry, file->uname.s, archive_strlen(&(file->uname)), xar->sconv) != 0) { if (errno == ENOMEM) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory for Uname"); return (ARCHIVE_FATAL); } archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Uname cannot be converted from %s to current locale.", archive_string_conversion_charset_name(xar->sconv)); r = ARCHIVE_WARN; } archive_entry_set_mode(entry, file->mode); if (archive_entry_copy_pathname_l(entry, file->pathname.s, archive_strlen(&(file->pathname)), xar->sconv) != 0) { if (errno == ENOMEM) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory for Pathname"); return (ARCHIVE_FATAL); } archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Pathname cannot be converted from %s to current locale.", archive_string_conversion_charset_name(xar->sconv)); r = ARCHIVE_WARN; } if (file->symlink.length > 0 && archive_entry_copy_symlink_l(entry, file->symlink.s, archive_strlen(&(file->symlink)), xar->sconv) != 0) { if (errno == ENOMEM) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory for Linkname"); return (ARCHIVE_FATAL); } archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Linkname cannot be converted from %s to current locale.", archive_string_conversion_charset_name(xar->sconv)); r = ARCHIVE_WARN; } /* Set proper nlink. */ if ((file->mode & AE_IFMT) == AE_IFDIR) archive_entry_set_nlink(entry, file->subdirs + 2); else archive_entry_set_nlink(entry, file->nlink); archive_entry_set_size(entry, file->size); if (archive_strlen(&(file->hardlink)) > 0) archive_entry_set_hardlink(entry, file->hardlink.s); archive_entry_set_ino64(entry, file->ino64); if (file->has & HAS_DEV) archive_entry_set_dev(entry, file->dev); if (file->has & HAS_DEVMAJOR) archive_entry_set_devmajor(entry, file->devmajor); if (file->has & HAS_DEVMINOR) archive_entry_set_devminor(entry, file->devminor); if (archive_strlen(&(file->fflags_text)) > 0) archive_entry_copy_fflags_text(entry, file->fflags_text.s); xar->entry_init = 1; xar->entry_total = 0; xar->entry_remaining = file->length; xar->entry_size = file->size; xar->entry_encoding = file->encoding; xar->entry_a_sum = file->a_sum; xar->entry_e_sum = file->e_sum; /* * Read extended attributes. */ xattr = file->xattr_list; while (xattr != NULL) { const void *d; size_t outbytes, used; r = move_reading_point(a, xattr->offset); if (r != ARCHIVE_OK) break; r = rd_contents_init(a, xattr->encoding, xattr->a_sum.alg, xattr->e_sum.alg); if (r != ARCHIVE_OK) break; d = NULL; r = rd_contents(a, &d, &outbytes, &used, xattr->length); if (r != ARCHIVE_OK) break; if (outbytes != xattr->size) { archive_set_error(&(a->archive), ARCHIVE_ERRNO_MISC, "Decompressed size error"); r = ARCHIVE_FATAL; break; } r = checksum_final(a, xattr->a_sum.val, xattr->a_sum.len, xattr->e_sum.val, xattr->e_sum.len); if (r != ARCHIVE_OK) break; archive_entry_xattr_add_entry(entry, xattr->name.s, d, outbytes); xattr = xattr->next; } if (r != ARCHIVE_OK) { file_free(file); return (r); } if (xar->entry_remaining > 0) /* Move reading point to the beginning of current * file contents. */ r = move_reading_point(a, file->offset); else r = ARCHIVE_OK; file_free(file); return (r); } static int xar_read_data(struct archive_read *a, const void **buff, size_t *size, int64_t *offset) { struct xar *xar; size_t used; int r; xar = (struct xar *)(a->format->data); if (xar->entry_unconsumed) { __archive_read_consume(a, xar->entry_unconsumed); xar->entry_unconsumed = 0; } if (xar->end_of_file || xar->entry_remaining <= 0) { r = ARCHIVE_EOF; goto abort_read_data; } if (xar->entry_init) { r = rd_contents_init(a, xar->entry_encoding, xar->entry_a_sum.alg, xar->entry_e_sum.alg); if (r != ARCHIVE_OK) { xar->entry_remaining = 0; return (r); } xar->entry_init = 0; } *buff = NULL; r = rd_contents(a, buff, size, &used, xar->entry_remaining); if (r != ARCHIVE_OK) goto abort_read_data; *offset = xar->entry_total; xar->entry_total += *size; xar->total += *size; xar->offset += used; xar->entry_remaining -= used; xar->entry_unconsumed = used; if (xar->entry_remaining == 0) { if (xar->entry_total != xar->entry_size) { archive_set_error(&(a->archive), ARCHIVE_ERRNO_MISC, "Decompressed size error"); r = ARCHIVE_FATAL; goto abort_read_data; } r = checksum_final(a, xar->entry_a_sum.val, xar->entry_a_sum.len, xar->entry_e_sum.val, xar->entry_e_sum.len); if (r != ARCHIVE_OK) goto abort_read_data; } return (ARCHIVE_OK); abort_read_data: *buff = NULL; *size = 0; *offset = xar->total; return (r); } static int xar_read_data_skip(struct archive_read *a) { struct xar *xar; int64_t bytes_skipped; xar = (struct xar *)(a->format->data); if (xar->end_of_file) return (ARCHIVE_EOF); bytes_skipped = __archive_read_consume(a, xar->entry_remaining + xar->entry_unconsumed); if (bytes_skipped < 0) return (ARCHIVE_FATAL); xar->offset += bytes_skipped; xar->entry_unconsumed = 0; return (ARCHIVE_OK); } static int xar_cleanup(struct archive_read *a) { struct xar *xar; struct hdlink *hdlink; int i; int r; xar = (struct xar *)(a->format->data); checksum_cleanup(a); r = decompression_cleanup(a); hdlink = xar->hdlink_list; while (hdlink != NULL) { struct hdlink *next = hdlink->next; free(hdlink); hdlink = next; } for (i = 0; i < xar->file_queue.used; i++) file_free(xar->file_queue.files[i]); free(xar->file_queue.files); while (xar->unknowntags != NULL) { struct unknown_tag *tag; tag = xar->unknowntags; xar->unknowntags = tag->next; archive_string_free(&(tag->name)); free(tag); } free(xar->outbuff); free(xar); a->format->data = NULL; return (r); } static int move_reading_point(struct archive_read *a, uint64_t offset) { struct xar *xar; xar = (struct xar *)(a->format->data); if (xar->offset - xar->h_base != offset) { /* Seek forward to the start of file contents. */ int64_t step; step = offset - (xar->offset - xar->h_base); if (step > 0) { step = __archive_read_consume(a, step); if (step < 0) return ((int)step); xar->offset += step; } else { int64_t pos = __archive_read_seek(a, offset, SEEK_SET); if (pos == ARCHIVE_FAILED) { archive_set_error(&(a->archive), ARCHIVE_ERRNO_MISC, "Cannot seek."); return (ARCHIVE_FAILED); } xar->offset = pos; } } return (ARCHIVE_OK); } static int rd_contents_init(struct archive_read *a, enum enctype encoding, int a_sum_alg, int e_sum_alg) { int r; /* Init decompress library. */ if ((r = decompression_init(a, encoding)) != ARCHIVE_OK) return (r); /* Init checksum library. */ checksum_init(a, a_sum_alg, e_sum_alg); return (ARCHIVE_OK); } static int rd_contents(struct archive_read *a, const void **buff, size_t *size, size_t *used, uint64_t remaining) { const unsigned char *b; ssize_t bytes; /* Get whatever bytes are immediately available. */ b = __archive_read_ahead(a, 1, &bytes); if (bytes < 0) return ((int)bytes); if (bytes == 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Truncated archive file"); return (ARCHIVE_FATAL); } if ((uint64_t)bytes > remaining) bytes = (ssize_t)remaining; /* * Decompress contents of file. */ *used = bytes; if (decompress(a, buff, size, b, used) != ARCHIVE_OK) return (ARCHIVE_FATAL); /* * Update checksum of a compressed data and a extracted data. */ checksum_update(a, b, *used, *buff, *size); return (ARCHIVE_OK); } /* * Note that this implementation does not (and should not!) obey * locale settings; you cannot simply substitute strtol here, since * it does obey locale. */ static uint64_t atol10(const char *p, size_t char_cnt) { uint64_t l; int digit; + if (char_cnt == 0) + return (0); + l = 0; digit = *p - '0'; while (digit >= 0 && digit < 10 && char_cnt-- > 0) { l = (l * 10) + digit; digit = *++p - '0'; } return (l); } static int64_t atol8(const char *p, size_t char_cnt) { int64_t l; int digit; - + + if (char_cnt == 0) + return (0); + l = 0; while (char_cnt-- > 0) { if (*p >= '0' && *p <= '7') digit = *p - '0'; else break; p++; l <<= 3; l |= digit; } return (l); } static size_t atohex(unsigned char *b, size_t bsize, const char *p, size_t psize) { size_t fbsize = bsize; while (bsize && psize > 1) { unsigned char x; if (p[0] >= 'a' && p[0] <= 'z') x = (p[0] - 'a' + 0x0a) << 4; else if (p[0] >= 'A' && p[0] <= 'Z') x = (p[0] - 'A' + 0x0a) << 4; else if (p[0] >= '0' && p[0] <= '9') x = (p[0] - '0') << 4; else return (-1); if (p[1] >= 'a' && p[1] <= 'z') x |= p[1] - 'a' + 0x0a; else if (p[1] >= 'A' && p[1] <= 'Z') x |= p[1] - 'A' + 0x0a; else if (p[1] >= '0' && p[1] <= '9') x |= p[1] - '0'; else return (-1); *b++ = x; bsize--; p += 2; psize -= 2; } return (fbsize - bsize); } static time_t time_from_tm(struct tm *t) { #if HAVE_TIMEGM /* Use platform timegm() if available. */ return (timegm(t)); #elif HAVE__MKGMTIME64 return (_mkgmtime64(t)); #else /* Else use direct calculation using POSIX assumptions. */ /* First, fix up tm_yday based on the year/month/day. */ mktime(t); /* Then we can compute timegm() from first principles. */ return (t->tm_sec + t->tm_min * 60 + t->tm_hour * 3600 + t->tm_yday * 86400 + (t->tm_year - 70) * 31536000 + ((t->tm_year - 69) / 4) * 86400 - ((t->tm_year - 1) / 100) * 86400 + ((t->tm_year + 299) / 400) * 86400); #endif } static time_t parse_time(const char *p, size_t n) { struct tm tm; time_t t = 0; int64_t data; memset(&tm, 0, sizeof(tm)); if (n != 20) return (t); data = atol10(p, 4); if (data < 1900) return (t); tm.tm_year = (int)data - 1900; p += 4; if (*p++ != '-') return (t); data = atol10(p, 2); if (data < 1 || data > 12) return (t); tm.tm_mon = (int)data -1; p += 2; if (*p++ != '-') return (t); data = atol10(p, 2); if (data < 1 || data > 31) return (t); tm.tm_mday = (int)data; p += 2; if (*p++ != 'T') return (t); data = atol10(p, 2); if (data < 0 || data > 23) return (t); tm.tm_hour = (int)data; p += 2; if (*p++ != ':') return (t); data = atol10(p, 2); if (data < 0 || data > 59) return (t); tm.tm_min = (int)data; p += 2; if (*p++ != ':') return (t); data = atol10(p, 2); if (data < 0 || data > 60) return (t); tm.tm_sec = (int)data; #if 0 p += 2; if (*p != 'Z') return (t); #endif t = time_from_tm(&tm); return (t); } static int heap_add_entry(struct archive_read *a, struct heap_queue *heap, struct xar_file *file) { uint64_t file_id, parent_id; int hole, parent; /* Expand our pending files list as necessary. */ if (heap->used >= heap->allocated) { struct xar_file **new_pending_files; int new_size = heap->allocated * 2; if (heap->allocated < 1024) new_size = 1024; /* Overflow might keep us from growing the list. */ if (new_size <= heap->allocated) { archive_set_error(&a->archive, ENOMEM, "Out of memory"); return (ARCHIVE_FATAL); } new_pending_files = (struct xar_file **) malloc(new_size * sizeof(new_pending_files[0])); if (new_pending_files == NULL) { archive_set_error(&a->archive, ENOMEM, "Out of memory"); return (ARCHIVE_FATAL); } memcpy(new_pending_files, heap->files, heap->allocated * sizeof(new_pending_files[0])); if (heap->files != NULL) free(heap->files); heap->files = new_pending_files; heap->allocated = new_size; } file_id = file->id; /* * Start with hole at end, walk it up tree to find insertion point. */ hole = heap->used++; while (hole > 0) { parent = (hole - 1)/2; parent_id = heap->files[parent]->id; if (file_id >= parent_id) { heap->files[hole] = file; return (ARCHIVE_OK); } /* Move parent into hole <==> move hole up tree. */ heap->files[hole] = heap->files[parent]; hole = parent; } heap->files[0] = file; return (ARCHIVE_OK); } static struct xar_file * heap_get_entry(struct heap_queue *heap) { uint64_t a_id, b_id, c_id; int a, b, c; struct xar_file *r, *tmp; if (heap->used < 1) return (NULL); /* * The first file in the list is the earliest; we'll return this. */ r = heap->files[0]; /* * Move the last item in the heap to the root of the tree */ heap->files[0] = heap->files[--(heap->used)]; /* * Rebalance the heap. */ a = 0; /* Starting element and its heap key */ a_id = heap->files[a]->id; for (;;) { b = a + a + 1; /* First child */ if (b >= heap->used) return (r); b_id = heap->files[b]->id; c = b + 1; /* Use second child if it is smaller. */ if (c < heap->used) { c_id = heap->files[c]->id; if (c_id < b_id) { b = c; b_id = c_id; } } if (a_id <= b_id) return (r); tmp = heap->files[a]; heap->files[a] = heap->files[b]; heap->files[b] = tmp; a = b; } } static int add_link(struct archive_read *a, struct xar *xar, struct xar_file *file) { struct hdlink *hdlink; for (hdlink = xar->hdlink_list; hdlink != NULL; hdlink = hdlink->next) { if (hdlink->id == file->link) { file->hdnext = hdlink->files; hdlink->cnt++; hdlink->files = file; return (ARCHIVE_OK); } } hdlink = malloc(sizeof(*hdlink)); if (hdlink == NULL) { archive_set_error(&a->archive, ENOMEM, "Out of memory"); return (ARCHIVE_FATAL); } file->hdnext = NULL; hdlink->id = file->link; hdlink->cnt = 1; hdlink->files = file; hdlink->next = xar->hdlink_list; xar->hdlink_list = hdlink; return (ARCHIVE_OK); } static void _checksum_init(struct chksumwork *sumwrk, int sum_alg) { sumwrk->alg = sum_alg; switch (sum_alg) { case CKSUM_NONE: break; case CKSUM_SHA1: archive_sha1_init(&(sumwrk->sha1ctx)); break; case CKSUM_MD5: archive_md5_init(&(sumwrk->md5ctx)); break; } } static void _checksum_update(struct chksumwork *sumwrk, const void *buff, size_t size) { switch (sumwrk->alg) { case CKSUM_NONE: break; case CKSUM_SHA1: archive_sha1_update(&(sumwrk->sha1ctx), buff, size); break; case CKSUM_MD5: archive_md5_update(&(sumwrk->md5ctx), buff, size); break; } } static int _checksum_final(struct chksumwork *sumwrk, const void *val, size_t len) { unsigned char sum[MAX_SUM_SIZE]; int r = ARCHIVE_OK; switch (sumwrk->alg) { case CKSUM_NONE: break; case CKSUM_SHA1: archive_sha1_final(&(sumwrk->sha1ctx), sum); if (len != SHA1_SIZE || memcmp(val, sum, SHA1_SIZE) != 0) r = ARCHIVE_FAILED; break; case CKSUM_MD5: archive_md5_final(&(sumwrk->md5ctx), sum); if (len != MD5_SIZE || memcmp(val, sum, MD5_SIZE) != 0) r = ARCHIVE_FAILED; break; } return (r); } static void checksum_init(struct archive_read *a, int a_sum_alg, int e_sum_alg) { struct xar *xar; xar = (struct xar *)(a->format->data); _checksum_init(&(xar->a_sumwrk), a_sum_alg); _checksum_init(&(xar->e_sumwrk), e_sum_alg); } static void checksum_update(struct archive_read *a, const void *abuff, size_t asize, const void *ebuff, size_t esize) { struct xar *xar; xar = (struct xar *)(a->format->data); _checksum_update(&(xar->a_sumwrk), abuff, asize); _checksum_update(&(xar->e_sumwrk), ebuff, esize); } static int checksum_final(struct archive_read *a, const void *a_sum_val, size_t a_sum_len, const void *e_sum_val, size_t e_sum_len) { struct xar *xar; int r; xar = (struct xar *)(a->format->data); r = _checksum_final(&(xar->a_sumwrk), a_sum_val, a_sum_len); if (r == ARCHIVE_OK) r = _checksum_final(&(xar->e_sumwrk), e_sum_val, e_sum_len); if (r != ARCHIVE_OK) archive_set_error(&(a->archive), ARCHIVE_ERRNO_MISC, "Sumcheck error"); return (r); } static int decompression_init(struct archive_read *a, enum enctype encoding) { struct xar *xar; const char *detail; int r; xar = (struct xar *)(a->format->data); xar->rd_encoding = encoding; switch (encoding) { case NONE: break; case GZIP: if (xar->stream_valid) r = inflateReset(&(xar->stream)); else r = inflateInit(&(xar->stream)); if (r != Z_OK) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Couldn't initialize zlib stream."); return (ARCHIVE_FATAL); } xar->stream_valid = 1; xar->stream.total_in = 0; xar->stream.total_out = 0; break; #if defined(HAVE_BZLIB_H) && defined(BZ_CONFIG_ERROR) case BZIP2: if (xar->bzstream_valid) { BZ2_bzDecompressEnd(&(xar->bzstream)); xar->bzstream_valid = 0; } r = BZ2_bzDecompressInit(&(xar->bzstream), 0, 0); if (r == BZ_MEM_ERROR) r = BZ2_bzDecompressInit(&(xar->bzstream), 0, 1); if (r != BZ_OK) { int err = ARCHIVE_ERRNO_MISC; detail = NULL; switch (r) { case BZ_PARAM_ERROR: detail = "invalid setup parameter"; break; case BZ_MEM_ERROR: err = ENOMEM; detail = "out of memory"; break; case BZ_CONFIG_ERROR: detail = "mis-compiled library"; break; } archive_set_error(&a->archive, err, "Internal error initializing decompressor: %s", detail == NULL ? "??" : detail); xar->bzstream_valid = 0; return (ARCHIVE_FATAL); } xar->bzstream_valid = 1; xar->bzstream.total_in_lo32 = 0; xar->bzstream.total_in_hi32 = 0; xar->bzstream.total_out_lo32 = 0; xar->bzstream.total_out_hi32 = 0; break; #endif #if defined(HAVE_LZMA_H) && defined(HAVE_LIBLZMA) #if LZMA_VERSION_MAJOR >= 5 /* Effectively disable the limiter. */ #define LZMA_MEMLIMIT UINT64_MAX #else /* NOTE: This needs to check memory size which running system has. */ #define LZMA_MEMLIMIT (1U << 30) #endif case XZ: case LZMA: if (xar->lzstream_valid) { lzma_end(&(xar->lzstream)); xar->lzstream_valid = 0; } if (xar->entry_encoding == XZ) r = lzma_stream_decoder(&(xar->lzstream), LZMA_MEMLIMIT,/* memlimit */ LZMA_CONCATENATED); else r = lzma_alone_decoder(&(xar->lzstream), LZMA_MEMLIMIT);/* memlimit */ if (r != LZMA_OK) { switch (r) { case LZMA_MEM_ERROR: archive_set_error(&a->archive, ENOMEM, "Internal error initializing " "compression library: " "Cannot allocate memory"); break; case LZMA_OPTIONS_ERROR: archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Internal error initializing " "compression library: " "Invalid or unsupported options"); break; default: archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Internal error initializing " "lzma library"); break; } return (ARCHIVE_FATAL); } xar->lzstream_valid = 1; xar->lzstream.total_in = 0; xar->lzstream.total_out = 0; break; #endif /* * Unsupported compression. */ default: #if !defined(HAVE_BZLIB_H) || !defined(BZ_CONFIG_ERROR) case BZIP2: #endif #if !defined(HAVE_LZMA_H) || !defined(HAVE_LIBLZMA) case LZMA: case XZ: #endif switch (xar->entry_encoding) { case BZIP2: detail = "bzip2"; break; case LZMA: detail = "lzma"; break; case XZ: detail = "xz"; break; default: detail = "??"; break; } archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "%s compression not supported on this platform", detail); return (ARCHIVE_FAILED); } return (ARCHIVE_OK); } static int decompress(struct archive_read *a, const void **buff, size_t *outbytes, const void *b, size_t *used) { struct xar *xar; void *outbuff; size_t avail_in, avail_out; int r; xar = (struct xar *)(a->format->data); avail_in = *used; outbuff = (void *)(uintptr_t)*buff; if (outbuff == NULL) { if (xar->outbuff == NULL) { xar->outbuff = malloc(OUTBUFF_SIZE); if (xar->outbuff == NULL) { archive_set_error(&a->archive, ENOMEM, "Couldn't allocate memory for out buffer"); return (ARCHIVE_FATAL); } } outbuff = xar->outbuff; *buff = outbuff; avail_out = OUTBUFF_SIZE; } else avail_out = *outbytes; switch (xar->rd_encoding) { case GZIP: xar->stream.next_in = (Bytef *)(uintptr_t)b; xar->stream.avail_in = avail_in; xar->stream.next_out = (unsigned char *)outbuff; xar->stream.avail_out = avail_out; r = inflate(&(xar->stream), 0); switch (r) { case Z_OK: /* Decompressor made some progress.*/ case Z_STREAM_END: /* Found end of stream. */ break; default: archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "File decompression failed (%d)", r); return (ARCHIVE_FATAL); } *used = avail_in - xar->stream.avail_in; *outbytes = avail_out - xar->stream.avail_out; break; #if defined(HAVE_BZLIB_H) && defined(BZ_CONFIG_ERROR) case BZIP2: xar->bzstream.next_in = (char *)(uintptr_t)b; xar->bzstream.avail_in = avail_in; xar->bzstream.next_out = (char *)outbuff; xar->bzstream.avail_out = avail_out; r = BZ2_bzDecompress(&(xar->bzstream)); switch (r) { case BZ_STREAM_END: /* Found end of stream. */ switch (BZ2_bzDecompressEnd(&(xar->bzstream))) { case BZ_OK: break; default: archive_set_error(&(a->archive), ARCHIVE_ERRNO_MISC, "Failed to clean up decompressor"); return (ARCHIVE_FATAL); } xar->bzstream_valid = 0; /* FALLTHROUGH */ case BZ_OK: /* Decompressor made some progress. */ break; default: archive_set_error(&(a->archive), ARCHIVE_ERRNO_MISC, "bzip decompression failed"); return (ARCHIVE_FATAL); } *used = avail_in - xar->bzstream.avail_in; *outbytes = avail_out - xar->bzstream.avail_out; break; #endif #if defined(HAVE_LZMA_H) && defined(HAVE_LIBLZMA) case LZMA: case XZ: xar->lzstream.next_in = b; xar->lzstream.avail_in = avail_in; xar->lzstream.next_out = (unsigned char *)outbuff; xar->lzstream.avail_out = avail_out; r = lzma_code(&(xar->lzstream), LZMA_RUN); switch (r) { case LZMA_STREAM_END: /* Found end of stream. */ lzma_end(&(xar->lzstream)); xar->lzstream_valid = 0; /* FALLTHROUGH */ case LZMA_OK: /* Decompressor made some progress. */ break; default: archive_set_error(&(a->archive), ARCHIVE_ERRNO_MISC, "%s decompression failed(%d)", (xar->entry_encoding == XZ)?"xz":"lzma", r); return (ARCHIVE_FATAL); } *used = avail_in - xar->lzstream.avail_in; *outbytes = avail_out - xar->lzstream.avail_out; break; #endif #if !defined(HAVE_BZLIB_H) || !defined(BZ_CONFIG_ERROR) case BZIP2: #endif #if !defined(HAVE_LZMA_H) || !defined(HAVE_LIBLZMA) case LZMA: case XZ: #endif case NONE: default: if (outbuff == xar->outbuff) { *buff = b; *used = avail_in; *outbytes = avail_in; } else { if (avail_out > avail_in) avail_out = avail_in; memcpy(outbuff, b, avail_out); *used = avail_out; *outbytes = avail_out; } break; } return (ARCHIVE_OK); } static int decompression_cleanup(struct archive_read *a) { struct xar *xar; int r; xar = (struct xar *)(a->format->data); r = ARCHIVE_OK; if (xar->stream_valid) { if (inflateEnd(&(xar->stream)) != Z_OK) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Failed to clean up zlib decompressor"); r = ARCHIVE_FATAL; } } #if defined(HAVE_BZLIB_H) && defined(BZ_CONFIG_ERROR) if (xar->bzstream_valid) { if (BZ2_bzDecompressEnd(&(xar->bzstream)) != BZ_OK) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Failed to clean up bzip2 decompressor"); r = ARCHIVE_FATAL; } } #endif #if defined(HAVE_LZMA_H) && defined(HAVE_LIBLZMA) if (xar->lzstream_valid) lzma_end(&(xar->lzstream)); #elif defined(HAVE_LZMA_H) && defined(HAVE_LIBLZMA) if (xar->lzstream_valid) { if (lzmadec_end(&(xar->lzstream)) != LZMADEC_OK) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Failed to clean up lzmadec decompressor"); r = ARCHIVE_FATAL; } } #endif return (r); } static void checksum_cleanup(struct archive_read *a) { struct xar *xar; xar = (struct xar *)(a->format->data); _checksum_final(&(xar->a_sumwrk), NULL, 0); _checksum_final(&(xar->e_sumwrk), NULL, 0); } static void xmlattr_cleanup(struct xmlattr_list *list) { struct xmlattr *attr, *next; attr = list->first; while (attr != NULL) { next = attr->next; free(attr->name); free(attr->value); free(attr); attr = next; } list->first = NULL; list->last = &(list->first); } static int file_new(struct archive_read *a, struct xar *xar, struct xmlattr_list *list) { struct xar_file *file; struct xmlattr *attr; file = calloc(1, sizeof(*file)); if (file == NULL) { archive_set_error(&a->archive, ENOMEM, "Out of memory"); return (ARCHIVE_FATAL); } file->parent = xar->file; file->mode = 0777 | AE_IFREG; file->atime = time(NULL); file->mtime = time(NULL); xar->file = file; xar->xattr = NULL; for (attr = list->first; attr != NULL; attr = attr->next) { if (strcmp(attr->name, "id") == 0) file->id = atol10(attr->value, strlen(attr->value)); } file->nlink = 1; if (heap_add_entry(a, &(xar->file_queue), file) != ARCHIVE_OK) return (ARCHIVE_FATAL); return (ARCHIVE_OK); } static void file_free(struct xar_file *file) { struct xattr *xattr; archive_string_free(&(file->pathname)); archive_string_free(&(file->symlink)); archive_string_free(&(file->uname)); archive_string_free(&(file->gname)); archive_string_free(&(file->hardlink)); xattr = file->xattr_list; while (xattr != NULL) { struct xattr *next; next = xattr->next; xattr_free(xattr); xattr = next; } free(file); } static int xattr_new(struct archive_read *a, struct xar *xar, struct xmlattr_list *list) { struct xattr *xattr, **nx; struct xmlattr *attr; xattr = calloc(1, sizeof(*xattr)); if (xattr == NULL) { archive_set_error(&a->archive, ENOMEM, "Out of memory"); return (ARCHIVE_FATAL); } xar->xattr = xattr; for (attr = list->first; attr != NULL; attr = attr->next) { if (strcmp(attr->name, "id") == 0) xattr->id = atol10(attr->value, strlen(attr->value)); } /* Chain to xattr list. */ for (nx = &(xar->file->xattr_list); *nx != NULL; nx = &((*nx)->next)) { if (xattr->id < (*nx)->id) break; } xattr->next = *nx; *nx = xattr; return (ARCHIVE_OK); } static void xattr_free(struct xattr *xattr) { archive_string_free(&(xattr->name)); free(xattr); } static int getencoding(struct xmlattr_list *list) { struct xmlattr *attr; enum enctype encoding = NONE; for (attr = list->first; attr != NULL; attr = attr->next) { if (strcmp(attr->name, "style") == 0) { if (strcmp(attr->value, "application/octet-stream") == 0) encoding = NONE; else if (strcmp(attr->value, "application/x-gzip") == 0) encoding = GZIP; else if (strcmp(attr->value, "application/x-bzip2") == 0) encoding = BZIP2; else if (strcmp(attr->value, "application/x-lzma") == 0) encoding = LZMA; else if (strcmp(attr->value, "application/x-xz") == 0) encoding = XZ; } } return (encoding); } static int getsumalgorithm(struct xmlattr_list *list) { struct xmlattr *attr; int alg = CKSUM_NONE; for (attr = list->first; attr != NULL; attr = attr->next) { if (strcmp(attr->name, "style") == 0) { const char *v = attr->value; if ((v[0] == 'S' || v[0] == 's') && (v[1] == 'H' || v[1] == 'h') && (v[2] == 'A' || v[2] == 'a') && v[3] == '1' && v[4] == '\0') alg = CKSUM_SHA1; if ((v[0] == 'M' || v[0] == 'm') && (v[1] == 'D' || v[1] == 'd') && v[2] == '5' && v[3] == '\0') alg = CKSUM_MD5; } } return (alg); } static int unknowntag_start(struct archive_read *a, struct xar *xar, const char *name) { struct unknown_tag *tag; tag = malloc(sizeof(*tag)); if (tag == NULL) { archive_set_error(&a->archive, ENOMEM, "Out of memory"); return (ARCHIVE_FATAL); } tag->next = xar->unknowntags; archive_string_init(&(tag->name)); archive_strcpy(&(tag->name), name); if (xar->unknowntags == NULL) { #if DEBUG fprintf(stderr, "UNKNOWNTAG_START:%s\n", name); #endif xar->xmlsts_unknown = xar->xmlsts; xar->xmlsts = UNKNOWN; } xar->unknowntags = tag; return (ARCHIVE_OK); } static void unknowntag_end(struct xar *xar, const char *name) { struct unknown_tag *tag; tag = xar->unknowntags; if (tag == NULL || name == NULL) return; if (strcmp(tag->name.s, name) == 0) { xar->unknowntags = tag->next; archive_string_free(&(tag->name)); free(tag); if (xar->unknowntags == NULL) { #if DEBUG fprintf(stderr, "UNKNOWNTAG_END:%s\n", name); #endif xar->xmlsts = xar->xmlsts_unknown; } } } static int xml_start(struct archive_read *a, const char *name, struct xmlattr_list *list) { struct xar *xar; struct xmlattr *attr; xar = (struct xar *)(a->format->data); #if DEBUG fprintf(stderr, "xml_sta:[%s]\n", name); for (attr = list->first; attr != NULL; attr = attr->next) fprintf(stderr, " attr:\"%s\"=\"%s\"\n", attr->name, attr->value); #endif xar->base64text = 0; switch (xar->xmlsts) { case INIT: if (strcmp(name, "xar") == 0) xar->xmlsts = XAR; else if (unknowntag_start(a, xar, name) != ARCHIVE_OK) return (ARCHIVE_FATAL); break; case XAR: if (strcmp(name, "toc") == 0) xar->xmlsts = TOC; else if (unknowntag_start(a, xar, name) != ARCHIVE_OK) return (ARCHIVE_FATAL); break; case TOC: if (strcmp(name, "creation-time") == 0) xar->xmlsts = TOC_CREATION_TIME; else if (strcmp(name, "checksum") == 0) xar->xmlsts = TOC_CHECKSUM; else if (strcmp(name, "file") == 0) { if (file_new(a, xar, list) != ARCHIVE_OK) return (ARCHIVE_FATAL); xar->xmlsts = TOC_FILE; } else if (unknowntag_start(a, xar, name) != ARCHIVE_OK) return (ARCHIVE_FATAL); break; case TOC_CHECKSUM: if (strcmp(name, "offset") == 0) xar->xmlsts = TOC_CHECKSUM_OFFSET; else if (strcmp(name, "size") == 0) xar->xmlsts = TOC_CHECKSUM_SIZE; else if (unknowntag_start(a, xar, name) != ARCHIVE_OK) return (ARCHIVE_FATAL); break; case TOC_FILE: if (strcmp(name, "file") == 0) { if (file_new(a, xar, list) != ARCHIVE_OK) return (ARCHIVE_FATAL); } else if (strcmp(name, "data") == 0) xar->xmlsts = FILE_DATA; else if (strcmp(name, "ea") == 0) { if (xattr_new(a, xar, list) != ARCHIVE_OK) return (ARCHIVE_FATAL); xar->xmlsts = FILE_EA; } else if (strcmp(name, "ctime") == 0) xar->xmlsts = FILE_CTIME; else if (strcmp(name, "mtime") == 0) xar->xmlsts = FILE_MTIME; else if (strcmp(name, "atime") == 0) xar->xmlsts = FILE_ATIME; else if (strcmp(name, "group") == 0) xar->xmlsts = FILE_GROUP; else if (strcmp(name, "gid") == 0) xar->xmlsts = FILE_GID; else if (strcmp(name, "user") == 0) xar->xmlsts = FILE_USER; else if (strcmp(name, "uid") == 0) xar->xmlsts = FILE_UID; else if (strcmp(name, "mode") == 0) xar->xmlsts = FILE_MODE; else if (strcmp(name, "device") == 0) xar->xmlsts = FILE_DEVICE; else if (strcmp(name, "deviceno") == 0) xar->xmlsts = FILE_DEVICENO; else if (strcmp(name, "inode") == 0) xar->xmlsts = FILE_INODE; else if (strcmp(name, "link") == 0) xar->xmlsts = FILE_LINK; else if (strcmp(name, "type") == 0) { xar->xmlsts = FILE_TYPE; for (attr = list->first; attr != NULL; attr = attr->next) { if (strcmp(attr->name, "link") != 0) continue; if (strcmp(attr->value, "original") == 0) { xar->file->hdnext = xar->hdlink_orgs; xar->hdlink_orgs = xar->file; } else { xar->file->link = (unsigned)atol10(attr->value, strlen(attr->value)); if (xar->file->link > 0) if (add_link(a, xar, xar->file) != ARCHIVE_OK) { return (ARCHIVE_FATAL); }; } } } else if (strcmp(name, "name") == 0) { xar->xmlsts = FILE_NAME; for (attr = list->first; attr != NULL; attr = attr->next) { if (strcmp(attr->name, "enctype") == 0 && strcmp(attr->value, "base64") == 0) xar->base64text = 1; } } else if (strcmp(name, "acl") == 0) xar->xmlsts = FILE_ACL; else if (strcmp(name, "flags") == 0) xar->xmlsts = FILE_FLAGS; else if (strcmp(name, "ext2") == 0) xar->xmlsts = FILE_EXT2; else if (unknowntag_start(a, xar, name) != ARCHIVE_OK) return (ARCHIVE_FATAL); break; case FILE_DATA: if (strcmp(name, "length") == 0) xar->xmlsts = FILE_DATA_LENGTH; else if (strcmp(name, "offset") == 0) xar->xmlsts = FILE_DATA_OFFSET; else if (strcmp(name, "size") == 0) xar->xmlsts = FILE_DATA_SIZE; else if (strcmp(name, "encoding") == 0) { xar->xmlsts = FILE_DATA_ENCODING; xar->file->encoding = getencoding(list); } else if (strcmp(name, "archived-checksum") == 0) { xar->xmlsts = FILE_DATA_A_CHECKSUM; xar->file->a_sum.alg = getsumalgorithm(list); } else if (strcmp(name, "extracted-checksum") == 0) { xar->xmlsts = FILE_DATA_E_CHECKSUM; xar->file->e_sum.alg = getsumalgorithm(list); } else if (strcmp(name, "content") == 0) xar->xmlsts = FILE_DATA_CONTENT; else if (unknowntag_start(a, xar, name) != ARCHIVE_OK) return (ARCHIVE_FATAL); break; case FILE_DEVICE: if (strcmp(name, "major") == 0) xar->xmlsts = FILE_DEVICE_MAJOR; else if (strcmp(name, "minor") == 0) xar->xmlsts = FILE_DEVICE_MINOR; else if (unknowntag_start(a, xar, name) != ARCHIVE_OK) return (ARCHIVE_FATAL); break; case FILE_DATA_CONTENT: if (unknowntag_start(a, xar, name) != ARCHIVE_OK) return (ARCHIVE_FATAL); break; case FILE_EA: if (strcmp(name, "length") == 0) xar->xmlsts = FILE_EA_LENGTH; else if (strcmp(name, "offset") == 0) xar->xmlsts = FILE_EA_OFFSET; else if (strcmp(name, "size") == 0) xar->xmlsts = FILE_EA_SIZE; else if (strcmp(name, "encoding") == 0) { xar->xmlsts = FILE_EA_ENCODING; xar->xattr->encoding = getencoding(list); } else if (strcmp(name, "archived-checksum") == 0) xar->xmlsts = FILE_EA_A_CHECKSUM; else if (strcmp(name, "extracted-checksum") == 0) xar->xmlsts = FILE_EA_E_CHECKSUM; else if (strcmp(name, "name") == 0) xar->xmlsts = FILE_EA_NAME; else if (strcmp(name, "fstype") == 0) xar->xmlsts = FILE_EA_FSTYPE; else if (unknowntag_start(a, xar, name) != ARCHIVE_OK) return (ARCHIVE_FATAL); break; case FILE_ACL: if (strcmp(name, "appleextended") == 0) xar->xmlsts = FILE_ACL_APPLEEXTENDED; else if (strcmp(name, "default") == 0) xar->xmlsts = FILE_ACL_DEFAULT; else if (strcmp(name, "access") == 0) xar->xmlsts = FILE_ACL_ACCESS; else if (unknowntag_start(a, xar, name) != ARCHIVE_OK) return (ARCHIVE_FATAL); break; case FILE_FLAGS: if (!xml_parse_file_flags(xar, name)) if (unknowntag_start(a, xar, name) != ARCHIVE_OK) return (ARCHIVE_FATAL); break; case FILE_EXT2: if (!xml_parse_file_ext2(xar, name)) if (unknowntag_start(a, xar, name) != ARCHIVE_OK) return (ARCHIVE_FATAL); break; case TOC_CREATION_TIME: case TOC_CHECKSUM_OFFSET: case TOC_CHECKSUM_SIZE: case FILE_DATA_LENGTH: case FILE_DATA_OFFSET: case FILE_DATA_SIZE: case FILE_DATA_ENCODING: case FILE_DATA_A_CHECKSUM: case FILE_DATA_E_CHECKSUM: case FILE_EA_LENGTH: case FILE_EA_OFFSET: case FILE_EA_SIZE: case FILE_EA_ENCODING: case FILE_EA_A_CHECKSUM: case FILE_EA_E_CHECKSUM: case FILE_EA_NAME: case FILE_EA_FSTYPE: case FILE_CTIME: case FILE_MTIME: case FILE_ATIME: case FILE_GROUP: case FILE_GID: case FILE_USER: case FILE_UID: case FILE_INODE: case FILE_DEVICE_MAJOR: case FILE_DEVICE_MINOR: case FILE_DEVICENO: case FILE_MODE: case FILE_TYPE: case FILE_LINK: case FILE_NAME: case FILE_ACL_DEFAULT: case FILE_ACL_ACCESS: case FILE_ACL_APPLEEXTENDED: case FILE_FLAGS_USER_NODUMP: case FILE_FLAGS_USER_IMMUTABLE: case FILE_FLAGS_USER_APPEND: case FILE_FLAGS_USER_OPAQUE: case FILE_FLAGS_USER_NOUNLINK: case FILE_FLAGS_SYS_ARCHIVED: case FILE_FLAGS_SYS_IMMUTABLE: case FILE_FLAGS_SYS_APPEND: case FILE_FLAGS_SYS_NOUNLINK: case FILE_FLAGS_SYS_SNAPSHOT: case FILE_EXT2_SecureDeletion: case FILE_EXT2_Undelete: case FILE_EXT2_Compress: case FILE_EXT2_Synchronous: case FILE_EXT2_Immutable: case FILE_EXT2_AppendOnly: case FILE_EXT2_NoDump: case FILE_EXT2_NoAtime: case FILE_EXT2_CompDirty: case FILE_EXT2_CompBlock: case FILE_EXT2_NoCompBlock: case FILE_EXT2_CompError: case FILE_EXT2_BTree: case FILE_EXT2_HashIndexed: case FILE_EXT2_iMagic: case FILE_EXT2_Journaled: case FILE_EXT2_NoTail: case FILE_EXT2_DirSync: case FILE_EXT2_TopDir: case FILE_EXT2_Reserved: case UNKNOWN: if (unknowntag_start(a, xar, name) != ARCHIVE_OK) return (ARCHIVE_FATAL); break; } return (ARCHIVE_OK); } static void xml_end(void *userData, const char *name) { struct archive_read *a; struct xar *xar; a = (struct archive_read *)userData; xar = (struct xar *)(a->format->data); #if DEBUG fprintf(stderr, "xml_end:[%s]\n", name); #endif switch (xar->xmlsts) { case INIT: break; case XAR: if (strcmp(name, "xar") == 0) xar->xmlsts = INIT; break; case TOC: if (strcmp(name, "toc") == 0) xar->xmlsts = XAR; break; case TOC_CREATION_TIME: if (strcmp(name, "creation-time") == 0) xar->xmlsts = TOC; break; case TOC_CHECKSUM: if (strcmp(name, "checksum") == 0) xar->xmlsts = TOC; break; case TOC_CHECKSUM_OFFSET: if (strcmp(name, "offset") == 0) xar->xmlsts = TOC_CHECKSUM; break; case TOC_CHECKSUM_SIZE: if (strcmp(name, "size") == 0) xar->xmlsts = TOC_CHECKSUM; break; case TOC_FILE: if (strcmp(name, "file") == 0) { if (xar->file->parent != NULL && ((xar->file->mode & AE_IFMT) == AE_IFDIR)) xar->file->parent->subdirs++; xar->file = xar->file->parent; if (xar->file == NULL) xar->xmlsts = TOC; } break; case FILE_DATA: if (strcmp(name, "data") == 0) xar->xmlsts = TOC_FILE; break; case FILE_DATA_LENGTH: if (strcmp(name, "length") == 0) xar->xmlsts = FILE_DATA; break; case FILE_DATA_OFFSET: if (strcmp(name, "offset") == 0) xar->xmlsts = FILE_DATA; break; case FILE_DATA_SIZE: if (strcmp(name, "size") == 0) xar->xmlsts = FILE_DATA; break; case FILE_DATA_ENCODING: if (strcmp(name, "encoding") == 0) xar->xmlsts = FILE_DATA; break; case FILE_DATA_A_CHECKSUM: if (strcmp(name, "archived-checksum") == 0) xar->xmlsts = FILE_DATA; break; case FILE_DATA_E_CHECKSUM: if (strcmp(name, "extracted-checksum") == 0) xar->xmlsts = FILE_DATA; break; case FILE_DATA_CONTENT: if (strcmp(name, "content") == 0) xar->xmlsts = FILE_DATA; break; case FILE_EA: if (strcmp(name, "ea") == 0) { xar->xmlsts = TOC_FILE; xar->xattr = NULL; } break; case FILE_EA_LENGTH: if (strcmp(name, "length") == 0) xar->xmlsts = FILE_EA; break; case FILE_EA_OFFSET: if (strcmp(name, "offset") == 0) xar->xmlsts = FILE_EA; break; case FILE_EA_SIZE: if (strcmp(name, "size") == 0) xar->xmlsts = FILE_EA; break; case FILE_EA_ENCODING: if (strcmp(name, "encoding") == 0) xar->xmlsts = FILE_EA; break; case FILE_EA_A_CHECKSUM: if (strcmp(name, "archived-checksum") == 0) xar->xmlsts = FILE_EA; break; case FILE_EA_E_CHECKSUM: if (strcmp(name, "extracted-checksum") == 0) xar->xmlsts = FILE_EA; break; case FILE_EA_NAME: if (strcmp(name, "name") == 0) xar->xmlsts = FILE_EA; break; case FILE_EA_FSTYPE: if (strcmp(name, "fstype") == 0) xar->xmlsts = FILE_EA; break; case FILE_CTIME: if (strcmp(name, "ctime") == 0) xar->xmlsts = TOC_FILE; break; case FILE_MTIME: if (strcmp(name, "mtime") == 0) xar->xmlsts = TOC_FILE; break; case FILE_ATIME: if (strcmp(name, "atime") == 0) xar->xmlsts = TOC_FILE; break; case FILE_GROUP: if (strcmp(name, "group") == 0) xar->xmlsts = TOC_FILE; break; case FILE_GID: if (strcmp(name, "gid") == 0) xar->xmlsts = TOC_FILE; break; case FILE_USER: if (strcmp(name, "user") == 0) xar->xmlsts = TOC_FILE; break; case FILE_UID: if (strcmp(name, "uid") == 0) xar->xmlsts = TOC_FILE; break; case FILE_MODE: if (strcmp(name, "mode") == 0) xar->xmlsts = TOC_FILE; break; case FILE_DEVICE: if (strcmp(name, "device") == 0) xar->xmlsts = TOC_FILE; break; case FILE_DEVICE_MAJOR: if (strcmp(name, "major") == 0) xar->xmlsts = FILE_DEVICE; break; case FILE_DEVICE_MINOR: if (strcmp(name, "minor") == 0) xar->xmlsts = FILE_DEVICE; break; case FILE_DEVICENO: if (strcmp(name, "deviceno") == 0) xar->xmlsts = TOC_FILE; break; case FILE_INODE: if (strcmp(name, "inode") == 0) xar->xmlsts = TOC_FILE; break; case FILE_LINK: if (strcmp(name, "link") == 0) xar->xmlsts = TOC_FILE; break; case FILE_TYPE: if (strcmp(name, "type") == 0) xar->xmlsts = TOC_FILE; break; case FILE_NAME: if (strcmp(name, "name") == 0) xar->xmlsts = TOC_FILE; break; case FILE_ACL: if (strcmp(name, "acl") == 0) xar->xmlsts = TOC_FILE; break; case FILE_ACL_DEFAULT: if (strcmp(name, "default") == 0) xar->xmlsts = FILE_ACL; break; case FILE_ACL_ACCESS: if (strcmp(name, "access") == 0) xar->xmlsts = FILE_ACL; break; case FILE_ACL_APPLEEXTENDED: if (strcmp(name, "appleextended") == 0) xar->xmlsts = FILE_ACL; break; case FILE_FLAGS: if (strcmp(name, "flags") == 0) xar->xmlsts = TOC_FILE; break; case FILE_FLAGS_USER_NODUMP: if (strcmp(name, "UserNoDump") == 0) xar->xmlsts = FILE_FLAGS; break; case FILE_FLAGS_USER_IMMUTABLE: if (strcmp(name, "UserImmutable") == 0) xar->xmlsts = FILE_FLAGS; break; case FILE_FLAGS_USER_APPEND: if (strcmp(name, "UserAppend") == 0) xar->xmlsts = FILE_FLAGS; break; case FILE_FLAGS_USER_OPAQUE: if (strcmp(name, "UserOpaque") == 0) xar->xmlsts = FILE_FLAGS; break; case FILE_FLAGS_USER_NOUNLINK: if (strcmp(name, "UserNoUnlink") == 0) xar->xmlsts = FILE_FLAGS; break; case FILE_FLAGS_SYS_ARCHIVED: if (strcmp(name, "SystemArchived") == 0) xar->xmlsts = FILE_FLAGS; break; case FILE_FLAGS_SYS_IMMUTABLE: if (strcmp(name, "SystemImmutable") == 0) xar->xmlsts = FILE_FLAGS; break; case FILE_FLAGS_SYS_APPEND: if (strcmp(name, "SystemAppend") == 0) xar->xmlsts = FILE_FLAGS; break; case FILE_FLAGS_SYS_NOUNLINK: if (strcmp(name, "SystemNoUnlink") == 0) xar->xmlsts = FILE_FLAGS; break; case FILE_FLAGS_SYS_SNAPSHOT: if (strcmp(name, "SystemSnapshot") == 0) xar->xmlsts = FILE_FLAGS; break; case FILE_EXT2: if (strcmp(name, "ext2") == 0) xar->xmlsts = TOC_FILE; break; case FILE_EXT2_SecureDeletion: if (strcmp(name, "SecureDeletion") == 0) xar->xmlsts = FILE_EXT2; break; case FILE_EXT2_Undelete: if (strcmp(name, "Undelete") == 0) xar->xmlsts = FILE_EXT2; break; case FILE_EXT2_Compress: if (strcmp(name, "Compress") == 0) xar->xmlsts = FILE_EXT2; break; case FILE_EXT2_Synchronous: if (strcmp(name, "Synchronous") == 0) xar->xmlsts = FILE_EXT2; break; case FILE_EXT2_Immutable: if (strcmp(name, "Immutable") == 0) xar->xmlsts = FILE_EXT2; break; case FILE_EXT2_AppendOnly: if (strcmp(name, "AppendOnly") == 0) xar->xmlsts = FILE_EXT2; break; case FILE_EXT2_NoDump: if (strcmp(name, "NoDump") == 0) xar->xmlsts = FILE_EXT2; break; case FILE_EXT2_NoAtime: if (strcmp(name, "NoAtime") == 0) xar->xmlsts = FILE_EXT2; break; case FILE_EXT2_CompDirty: if (strcmp(name, "CompDirty") == 0) xar->xmlsts = FILE_EXT2; break; case FILE_EXT2_CompBlock: if (strcmp(name, "CompBlock") == 0) xar->xmlsts = FILE_EXT2; break; case FILE_EXT2_NoCompBlock: if (strcmp(name, "NoCompBlock") == 0) xar->xmlsts = FILE_EXT2; break; case FILE_EXT2_CompError: if (strcmp(name, "CompError") == 0) xar->xmlsts = FILE_EXT2; break; case FILE_EXT2_BTree: if (strcmp(name, "BTree") == 0) xar->xmlsts = FILE_EXT2; break; case FILE_EXT2_HashIndexed: if (strcmp(name, "HashIndexed") == 0) xar->xmlsts = FILE_EXT2; break; case FILE_EXT2_iMagic: if (strcmp(name, "iMagic") == 0) xar->xmlsts = FILE_EXT2; break; case FILE_EXT2_Journaled: if (strcmp(name, "Journaled") == 0) xar->xmlsts = FILE_EXT2; break; case FILE_EXT2_NoTail: if (strcmp(name, "NoTail") == 0) xar->xmlsts = FILE_EXT2; break; case FILE_EXT2_DirSync: if (strcmp(name, "DirSync") == 0) xar->xmlsts = FILE_EXT2; break; case FILE_EXT2_TopDir: if (strcmp(name, "TopDir") == 0) xar->xmlsts = FILE_EXT2; break; case FILE_EXT2_Reserved: if (strcmp(name, "Reserved") == 0) xar->xmlsts = FILE_EXT2; break; case UNKNOWN: unknowntag_end(xar, name); break; } } static const int base64[256] = { -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, /* 00 - 0F */ -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, /* 10 - 1F */ -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 62, -1, -1, -1, 63, /* 20 - 2F */ 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, -1, -1, -1, -1, -1, -1, /* 30 - 3F */ -1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, /* 40 - 4F */ 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, -1, -1, -1, -1, -1, /* 50 - 5F */ -1, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, /* 60 - 6F */ 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, -1, -1, -1, -1, -1, /* 70 - 7F */ -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, /* 80 - 8F */ -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, /* 90 - 9F */ -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, /* A0 - AF */ -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, /* B0 - BF */ -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, /* C0 - CF */ -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, /* D0 - DF */ -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, /* E0 - EF */ -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, /* F0 - FF */ }; static void strappend_base64(struct xar *xar, struct archive_string *as, const char *s, size_t l) { unsigned char buff[256]; unsigned char *out; const unsigned char *b; size_t len; (void)xar; /* UNUSED */ len = 0; out = buff; b = (const unsigned char *)s; while (l > 0) { int n = 0; if (l > 0) { if (base64[b[0]] < 0 || base64[b[1]] < 0) break; n = base64[*b++] << 18; n |= base64[*b++] << 12; *out++ = n >> 16; len++; l -= 2; } if (l > 0) { if (base64[*b] < 0) break; n |= base64[*b++] << 6; *out++ = (n >> 8) & 0xFF; len++; --l; } if (l > 0) { if (base64[*b] < 0) break; n |= base64[*b++]; *out++ = n & 0xFF; len++; --l; } if (len+3 >= sizeof(buff)) { archive_strncat(as, (const char *)buff, len); len = 0; out = buff; } } if (len > 0) archive_strncat(as, (const char *)buff, len); } +static int +is_string(const char *known, const char *data, size_t len) +{ + if (strlen(known) != len) + return -1; + return memcmp(data, known, len); +} + static void xml_data(void *userData, const char *s, int len) { struct archive_read *a; struct xar *xar; a = (struct archive_read *)userData; xar = (struct xar *)(a->format->data); #if DEBUG { char buff[1024]; if (len > (int)(sizeof(buff)-1)) len = (int)(sizeof(buff)-1); strncpy(buff, s, len); buff[len] = 0; fprintf(stderr, "\tlen=%d:\"%s\"\n", len, buff); } #endif switch (xar->xmlsts) { case TOC_CHECKSUM_OFFSET: xar->toc_chksum_offset = atol10(s, len); break; case TOC_CHECKSUM_SIZE: xar->toc_chksum_size = atol10(s, len); break; default: break; } if (xar->file == NULL) return; switch (xar->xmlsts) { case FILE_NAME: if (xar->file->parent != NULL) { archive_string_concat(&(xar->file->pathname), &(xar->file->parent->pathname)); archive_strappend_char(&(xar->file->pathname), '/'); } xar->file->has |= HAS_PATHNAME; if (xar->base64text) { strappend_base64(xar, &(xar->file->pathname), s, len); } else archive_strncat(&(xar->file->pathname), s, len); break; case FILE_LINK: xar->file->has |= HAS_SYMLINK; archive_strncpy(&(xar->file->symlink), s, len); break; case FILE_TYPE: - if (strncmp("file", s, len) == 0 || - strncmp("hardlink", s, len) == 0) + if (is_string("file", s, len) == 0 || + is_string("hardlink", s, len) == 0) xar->file->mode = (xar->file->mode & ~AE_IFMT) | AE_IFREG; - if (strncmp("directory", s, len) == 0) + if (is_string("directory", s, len) == 0) xar->file->mode = (xar->file->mode & ~AE_IFMT) | AE_IFDIR; - if (strncmp("symlink", s, len) == 0) + if (is_string("symlink", s, len) == 0) xar->file->mode = (xar->file->mode & ~AE_IFMT) | AE_IFLNK; - if (strncmp("character special", s, len) == 0) + if (is_string("character special", s, len) == 0) xar->file->mode = (xar->file->mode & ~AE_IFMT) | AE_IFCHR; - if (strncmp("block special", s, len) == 0) + if (is_string("block special", s, len) == 0) xar->file->mode = (xar->file->mode & ~AE_IFMT) | AE_IFBLK; - if (strncmp("socket", s, len) == 0) + if (is_string("socket", s, len) == 0) xar->file->mode = (xar->file->mode & ~AE_IFMT) | AE_IFSOCK; - if (strncmp("fifo", s, len) == 0) + if (is_string("fifo", s, len) == 0) xar->file->mode = (xar->file->mode & ~AE_IFMT) | AE_IFIFO; xar->file->has |= HAS_TYPE; break; case FILE_INODE: xar->file->has |= HAS_INO; xar->file->ino64 = atol10(s, len); break; case FILE_DEVICE_MAJOR: xar->file->has |= HAS_DEVMAJOR; xar->file->devmajor = (dev_t)atol10(s, len); break; case FILE_DEVICE_MINOR: xar->file->has |= HAS_DEVMINOR; xar->file->devminor = (dev_t)atol10(s, len); break; case FILE_DEVICENO: xar->file->has |= HAS_DEV; xar->file->dev = (dev_t)atol10(s, len); break; case FILE_MODE: xar->file->has |= HAS_MODE; xar->file->mode = (xar->file->mode & AE_IFMT) | ((mode_t)(atol8(s, len)) & ~AE_IFMT); break; case FILE_GROUP: xar->file->has |= HAS_GID; archive_strncpy(&(xar->file->gname), s, len); break; case FILE_GID: xar->file->has |= HAS_GID; xar->file->gid = atol10(s, len); break; case FILE_USER: xar->file->has |= HAS_UID; archive_strncpy(&(xar->file->uname), s, len); break; case FILE_UID: xar->file->has |= HAS_UID; xar->file->uid = atol10(s, len); break; case FILE_CTIME: xar->file->has |= HAS_TIME; xar->file->ctime = parse_time(s, len); break; case FILE_MTIME: xar->file->has |= HAS_TIME; xar->file->mtime = parse_time(s, len); break; case FILE_ATIME: xar->file->has |= HAS_TIME; xar->file->atime = parse_time(s, len); break; case FILE_DATA_LENGTH: xar->file->has |= HAS_DATA; xar->file->length = atol10(s, len); break; case FILE_DATA_OFFSET: xar->file->has |= HAS_DATA; xar->file->offset = atol10(s, len); break; case FILE_DATA_SIZE: xar->file->has |= HAS_DATA; xar->file->size = atol10(s, len); break; case FILE_DATA_A_CHECKSUM: xar->file->a_sum.len = atohex(xar->file->a_sum.val, sizeof(xar->file->a_sum.val), s, len); break; case FILE_DATA_E_CHECKSUM: xar->file->e_sum.len = atohex(xar->file->e_sum.val, sizeof(xar->file->e_sum.val), s, len); break; case FILE_EA_LENGTH: xar->file->has |= HAS_XATTR; xar->xattr->length = atol10(s, len); break; case FILE_EA_OFFSET: xar->file->has |= HAS_XATTR; xar->xattr->offset = atol10(s, len); break; case FILE_EA_SIZE: xar->file->has |= HAS_XATTR; xar->xattr->size = atol10(s, len); break; case FILE_EA_A_CHECKSUM: xar->file->has |= HAS_XATTR; xar->xattr->a_sum.len = atohex(xar->xattr->a_sum.val, sizeof(xar->xattr->a_sum.val), s, len); break; case FILE_EA_E_CHECKSUM: xar->file->has |= HAS_XATTR; xar->xattr->e_sum.len = atohex(xar->xattr->e_sum.val, sizeof(xar->xattr->e_sum.val), s, len); break; case FILE_EA_NAME: xar->file->has |= HAS_XATTR; archive_strncpy(&(xar->xattr->name), s, len); break; case FILE_EA_FSTYPE: xar->file->has |= HAS_XATTR; archive_strncpy(&(xar->xattr->fstype), s, len); break; break; case FILE_ACL_DEFAULT: case FILE_ACL_ACCESS: case FILE_ACL_APPLEEXTENDED: xar->file->has |= HAS_ACL; /* TODO */ break; case INIT: case XAR: case TOC: case TOC_CREATION_TIME: case TOC_CHECKSUM: case TOC_CHECKSUM_OFFSET: case TOC_CHECKSUM_SIZE: case TOC_FILE: case FILE_DATA: case FILE_DATA_ENCODING: case FILE_DATA_CONTENT: case FILE_DEVICE: case FILE_EA: case FILE_EA_ENCODING: case FILE_ACL: case FILE_FLAGS: case FILE_FLAGS_USER_NODUMP: case FILE_FLAGS_USER_IMMUTABLE: case FILE_FLAGS_USER_APPEND: case FILE_FLAGS_USER_OPAQUE: case FILE_FLAGS_USER_NOUNLINK: case FILE_FLAGS_SYS_ARCHIVED: case FILE_FLAGS_SYS_IMMUTABLE: case FILE_FLAGS_SYS_APPEND: case FILE_FLAGS_SYS_NOUNLINK: case FILE_FLAGS_SYS_SNAPSHOT: case FILE_EXT2: case FILE_EXT2_SecureDeletion: case FILE_EXT2_Undelete: case FILE_EXT2_Compress: case FILE_EXT2_Synchronous: case FILE_EXT2_Immutable: case FILE_EXT2_AppendOnly: case FILE_EXT2_NoDump: case FILE_EXT2_NoAtime: case FILE_EXT2_CompDirty: case FILE_EXT2_CompBlock: case FILE_EXT2_NoCompBlock: case FILE_EXT2_CompError: case FILE_EXT2_BTree: case FILE_EXT2_HashIndexed: case FILE_EXT2_iMagic: case FILE_EXT2_Journaled: case FILE_EXT2_NoTail: case FILE_EXT2_DirSync: case FILE_EXT2_TopDir: case FILE_EXT2_Reserved: case UNKNOWN: break; } } /* * BSD file flags. */ static int xml_parse_file_flags(struct xar *xar, const char *name) { const char *flag = NULL; if (strcmp(name, "UserNoDump") == 0) { xar->xmlsts = FILE_FLAGS_USER_NODUMP; flag = "nodump"; } else if (strcmp(name, "UserImmutable") == 0) { xar->xmlsts = FILE_FLAGS_USER_IMMUTABLE; flag = "uimmutable"; } else if (strcmp(name, "UserAppend") == 0) { xar->xmlsts = FILE_FLAGS_USER_APPEND; flag = "uappend"; } else if (strcmp(name, "UserOpaque") == 0) { xar->xmlsts = FILE_FLAGS_USER_OPAQUE; flag = "opaque"; } else if (strcmp(name, "UserNoUnlink") == 0) { xar->xmlsts = FILE_FLAGS_USER_NOUNLINK; flag = "nouunlink"; } else if (strcmp(name, "SystemArchived") == 0) { xar->xmlsts = FILE_FLAGS_SYS_ARCHIVED; flag = "archived"; } else if (strcmp(name, "SystemImmutable") == 0) { xar->xmlsts = FILE_FLAGS_SYS_IMMUTABLE; flag = "simmutable"; } else if (strcmp(name, "SystemAppend") == 0) { xar->xmlsts = FILE_FLAGS_SYS_APPEND; flag = "sappend"; } else if (strcmp(name, "SystemNoUnlink") == 0) { xar->xmlsts = FILE_FLAGS_SYS_NOUNLINK; flag = "nosunlink"; } else if (strcmp(name, "SystemSnapshot") == 0) { xar->xmlsts = FILE_FLAGS_SYS_SNAPSHOT; flag = "snapshot"; } if (flag == NULL) return (0); xar->file->has |= HAS_FFLAGS; if (archive_strlen(&(xar->file->fflags_text)) > 0) archive_strappend_char(&(xar->file->fflags_text), ','); archive_strcat(&(xar->file->fflags_text), flag); return (1); } /* * Linux file flags. */ static int xml_parse_file_ext2(struct xar *xar, const char *name) { const char *flag = NULL; if (strcmp(name, "SecureDeletion") == 0) { xar->xmlsts = FILE_EXT2_SecureDeletion; flag = "securedeletion"; } else if (strcmp(name, "Undelete") == 0) { xar->xmlsts = FILE_EXT2_Undelete; flag = "nouunlink"; } else if (strcmp(name, "Compress") == 0) { xar->xmlsts = FILE_EXT2_Compress; flag = "compress"; } else if (strcmp(name, "Synchronous") == 0) { xar->xmlsts = FILE_EXT2_Synchronous; flag = "sync"; } else if (strcmp(name, "Immutable") == 0) { xar->xmlsts = FILE_EXT2_Immutable; flag = "simmutable"; } else if (strcmp(name, "AppendOnly") == 0) { xar->xmlsts = FILE_EXT2_AppendOnly; flag = "sappend"; } else if (strcmp(name, "NoDump") == 0) { xar->xmlsts = FILE_EXT2_NoDump; flag = "nodump"; } else if (strcmp(name, "NoAtime") == 0) { xar->xmlsts = FILE_EXT2_NoAtime; flag = "noatime"; } else if (strcmp(name, "CompDirty") == 0) { xar->xmlsts = FILE_EXT2_CompDirty; flag = "compdirty"; } else if (strcmp(name, "CompBlock") == 0) { xar->xmlsts = FILE_EXT2_CompBlock; flag = "comprblk"; } else if (strcmp(name, "NoCompBlock") == 0) { xar->xmlsts = FILE_EXT2_NoCompBlock; flag = "nocomprblk"; } else if (strcmp(name, "CompError") == 0) { xar->xmlsts = FILE_EXT2_CompError; flag = "comperr"; } else if (strcmp(name, "BTree") == 0) { xar->xmlsts = FILE_EXT2_BTree; flag = "btree"; } else if (strcmp(name, "HashIndexed") == 0) { xar->xmlsts = FILE_EXT2_HashIndexed; flag = "hashidx"; } else if (strcmp(name, "iMagic") == 0) { xar->xmlsts = FILE_EXT2_iMagic; flag = "imagic"; } else if (strcmp(name, "Journaled") == 0) { xar->xmlsts = FILE_EXT2_Journaled; flag = "journal"; } else if (strcmp(name, "NoTail") == 0) { xar->xmlsts = FILE_EXT2_NoTail; flag = "notail"; } else if (strcmp(name, "DirSync") == 0) { xar->xmlsts = FILE_EXT2_DirSync; flag = "dirsync"; } else if (strcmp(name, "TopDir") == 0) { xar->xmlsts = FILE_EXT2_TopDir; flag = "topdir"; } else if (strcmp(name, "Reserved") == 0) { xar->xmlsts = FILE_EXT2_Reserved; flag = "reserved"; } if (flag == NULL) return (0); if (archive_strlen(&(xar->file->fflags_text)) > 0) archive_strappend_char(&(xar->file->fflags_text), ','); archive_strcat(&(xar->file->fflags_text), flag); return (1); } #ifdef HAVE_LIBXML_XMLREADER_H static int xml2_xmlattr_setup(struct archive_read *a, struct xmlattr_list *list, xmlTextReaderPtr reader) { struct xmlattr *attr; int r; list->first = NULL; list->last = &(list->first); r = xmlTextReaderMoveToFirstAttribute(reader); while (r == 1) { attr = malloc(sizeof*(attr)); if (attr == NULL) { archive_set_error(&a->archive, ENOMEM, "Out of memory"); return (ARCHIVE_FATAL); } attr->name = strdup( (const char *)xmlTextReaderConstLocalName(reader)); if (attr->name == NULL) { free(attr); archive_set_error(&a->archive, ENOMEM, "Out of memory"); return (ARCHIVE_FATAL); } attr->value = strdup( (const char *)xmlTextReaderConstValue(reader)); if (attr->value == NULL) { free(attr->name); free(attr); archive_set_error(&a->archive, ENOMEM, "Out of memory"); return (ARCHIVE_FATAL); } attr->next = NULL; *list->last = attr; list->last = &(attr->next); r = xmlTextReaderMoveToNextAttribute(reader); } return (r); } static int xml2_read_cb(void *context, char *buffer, int len) { struct archive_read *a; struct xar *xar; const void *d; size_t outbytes; size_t used = 0; int r; a = (struct archive_read *)context; xar = (struct xar *)(a->format->data); if (xar->toc_remaining <= 0) return (0); d = buffer; outbytes = len; r = rd_contents(a, &d, &outbytes, &used, xar->toc_remaining); if (r != ARCHIVE_OK) return (r); __archive_read_consume(a, used); xar->toc_remaining -= used; xar->offset += used; xar->toc_total += outbytes; PRINT_TOC(buffer, len); return ((int)outbytes); } static int xml2_close_cb(void *context) { (void)context; /* UNUSED */ return (0); } static void xml2_error_hdr(void *arg, const char *msg, xmlParserSeverities severity, xmlTextReaderLocatorPtr locator) { struct archive_read *a; (void)locator; /* UNUSED */ a = (struct archive_read *)arg; switch (severity) { case XML_PARSER_SEVERITY_VALIDITY_WARNING: case XML_PARSER_SEVERITY_WARNING: archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "XML Parsing error: %s", msg); break; case XML_PARSER_SEVERITY_VALIDITY_ERROR: case XML_PARSER_SEVERITY_ERROR: archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "XML Parsing error: %s", msg); break; } } static int xml2_read_toc(struct archive_read *a) { xmlTextReaderPtr reader; struct xmlattr_list list; int r; reader = xmlReaderForIO(xml2_read_cb, xml2_close_cb, a, NULL, NULL, 0); if (reader == NULL) { archive_set_error(&a->archive, ENOMEM, "Couldn't allocate memory for xml parser"); return (ARCHIVE_FATAL); } xmlTextReaderSetErrorHandler(reader, xml2_error_hdr, a); while ((r = xmlTextReaderRead(reader)) == 1) { const char *name, *value; int type, empty; type = xmlTextReaderNodeType(reader); name = (const char *)xmlTextReaderConstLocalName(reader); switch (type) { case XML_READER_TYPE_ELEMENT: empty = xmlTextReaderIsEmptyElement(reader); r = xml2_xmlattr_setup(a, &list, reader); if (r == ARCHIVE_OK) r = xml_start(a, name, &list); xmlattr_cleanup(&list); if (r != ARCHIVE_OK) return (r); if (empty) xml_end(a, name); break; case XML_READER_TYPE_END_ELEMENT: xml_end(a, name); break; case XML_READER_TYPE_TEXT: value = (const char *)xmlTextReaderConstValue(reader); xml_data(a, value, strlen(value)); break; case XML_READER_TYPE_SIGNIFICANT_WHITESPACE: default: break; } if (r < 0) break; } xmlFreeTextReader(reader); xmlCleanupParser(); return ((r == 0)?ARCHIVE_OK:ARCHIVE_FATAL); } #elif defined(HAVE_BSDXML_H) || defined(HAVE_EXPAT_H) static int expat_xmlattr_setup(struct archive_read *a, struct xmlattr_list *list, const XML_Char **atts) { struct xmlattr *attr; char *name, *value; list->first = NULL; list->last = &(list->first); if (atts == NULL) return (ARCHIVE_OK); while (atts[0] != NULL && atts[1] != NULL) { attr = malloc(sizeof*(attr)); name = strdup(atts[0]); value = strdup(atts[1]); if (attr == NULL || name == NULL || value == NULL) { archive_set_error(&a->archive, ENOMEM, "Out of memory"); free(attr); free(name); free(value); return (ARCHIVE_FATAL); } attr->name = name; attr->value = value; attr->next = NULL; *list->last = attr; list->last = &(attr->next); atts += 2; } return (ARCHIVE_OK); } static void expat_start_cb(void *userData, const XML_Char *name, const XML_Char **atts) { struct expat_userData *ud = (struct expat_userData *)userData; struct archive_read *a = ud->archive; struct xmlattr_list list; int r; r = expat_xmlattr_setup(a, &list, atts); if (r == ARCHIVE_OK) r = xml_start(a, (const char *)name, &list); xmlattr_cleanup(&list); ud->state = r; } static void expat_end_cb(void *userData, const XML_Char *name) { struct expat_userData *ud = (struct expat_userData *)userData; xml_end(ud->archive, (const char *)name); } static void expat_data_cb(void *userData, const XML_Char *s, int len) { struct expat_userData *ud = (struct expat_userData *)userData; xml_data(ud->archive, s, len); } static int expat_read_toc(struct archive_read *a) { struct xar *xar; XML_Parser parser; struct expat_userData ud; ud.state = ARCHIVE_OK; ud.archive = a; xar = (struct xar *)(a->format->data); /* Initialize XML Parser library. */ parser = XML_ParserCreate(NULL); if (parser == NULL) { archive_set_error(&a->archive, ENOMEM, "Couldn't allocate memory for xml parser"); return (ARCHIVE_FATAL); } XML_SetUserData(parser, &ud); XML_SetElementHandler(parser, expat_start_cb, expat_end_cb); XML_SetCharacterDataHandler(parser, expat_data_cb); xar->xmlsts = INIT; while (xar->toc_remaining && ud.state == ARCHIVE_OK) { enum XML_Status xr; const void *d; size_t outbytes; size_t used; int r; d = NULL; r = rd_contents(a, &d, &outbytes, &used, xar->toc_remaining); if (r != ARCHIVE_OK) return (r); xar->toc_remaining -= used; xar->offset += used; xar->toc_total += outbytes; PRINT_TOC(d, outbytes); xr = XML_Parse(parser, d, outbytes, xar->toc_remaining == 0); __archive_read_consume(a, used); if (xr == XML_STATUS_ERROR) { XML_ParserFree(parser); archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "XML Parsing failed"); return (ARCHIVE_FATAL); } } XML_ParserFree(parser); return (ud.state); } #endif /* defined(HAVE_BSDXML_H) || defined(HAVE_EXPAT_H) */ #endif /* Support xar format */ Index: stable/10/contrib/libarchive/libarchive/archive_read_support_format_zip.c =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_read_support_format_zip.c (revision 324417) +++ stable/10/contrib/libarchive/libarchive/archive_read_support_format_zip.c (revision 324418) @@ -1,3113 +1,3118 @@ /*- * Copyright (c) 2004-2013 Tim Kientzle * Copyright (c) 2011-2012,2014 Michihiro NAKAJIMA * Copyright (c) 2013 Konrad Kleine * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "archive_platform.h" __FBSDID("$FreeBSD$"); /* * The definitive documentation of the Zip file format is: * http://www.pkware.com/documents/casestudies/APPNOTE.TXT * * The Info-Zip project has pioneered various extensions to better * support Zip on Unix, including the 0x5455 "UT", 0x5855 "UX", 0x7855 * "Ux", and 0x7875 "ux" extensions for time and ownership * information. * * History of this code: The streaming Zip reader was first added to * libarchive in January 2005. Support for seekable input sources was * added in Nov 2011. Zip64 support (including a significant code * refactoring) was added in 2014. */ #ifdef HAVE_ERRNO_H #include #endif #ifdef HAVE_STDLIB_H #include #endif #ifdef HAVE_ZLIB_H #include #endif #include "archive.h" #include "archive_digest_private.h" #include "archive_cryptor_private.h" #include "archive_endian.h" #include "archive_entry.h" #include "archive_entry_locale.h" #include "archive_hmac_private.h" #include "archive_private.h" #include "archive_rb.h" #include "archive_read_private.h" #ifndef HAVE_ZLIB_H #include "archive_crc32.h" #endif struct zip_entry { struct archive_rb_node node; struct zip_entry *next; int64_t local_header_offset; int64_t compressed_size; int64_t uncompressed_size; int64_t gid; int64_t uid; struct archive_string rsrcname; time_t mtime; time_t atime; time_t ctime; uint32_t crc32; uint16_t mode; uint16_t zip_flags; /* From GP Flags Field */ unsigned char compression; unsigned char system; /* From "version written by" */ unsigned char flags; /* Our extra markers. */ unsigned char decdat;/* Used for Decryption check */ /* WinZip AES encryption extra field should be available * when compression is 99. */ struct { /* Vendor version: AE-1 - 0x0001, AE-2 - 0x0002 */ unsigned vendor; #define AES_VENDOR_AE_1 0x0001 #define AES_VENDOR_AE_2 0x0002 /* AES encryption strength: * 1 - 128 bits, 2 - 192 bits, 2 - 256 bits. */ unsigned strength; /* Actual compression method. */ unsigned char compression; } aes_extra; }; struct trad_enc_ctx { uint32_t keys[3]; }; /* Bits used in zip_flags. */ #define ZIP_ENCRYPTED (1 << 0) #define ZIP_LENGTH_AT_END (1 << 3) #define ZIP_STRONG_ENCRYPTED (1 << 6) #define ZIP_UTF8_NAME (1 << 11) /* See "7.2 Single Password Symmetric Encryption Method" in http://www.pkware.com/documents/casestudies/APPNOTE.TXT */ #define ZIP_CENTRAL_DIRECTORY_ENCRYPTED (1 << 13) /* Bits used in flags. */ #define LA_USED_ZIP64 (1 << 0) #define LA_FROM_CENTRAL_DIRECTORY (1 << 1) /* * See "WinZip - AES Encryption Information" * http://www.winzip.com/aes_info.htm */ /* Value used in compression method. */ #define WINZIP_AES_ENCRYPTION 99 /* Authentication code size. */ #define AUTH_CODE_SIZE 10 /**/ #define MAX_DERIVED_KEY_BUF_SIZE (AES_MAX_KEY_SIZE * 2 + 2) struct zip { /* Structural information about the archive. */ struct archive_string format_name; int64_t central_directory_offset; size_t central_directory_entries_total; size_t central_directory_entries_on_this_disk; int has_encrypted_entries; /* List of entries (seekable Zip only) */ struct zip_entry *zip_entries; struct archive_rb_tree tree; struct archive_rb_tree tree_rsrc; /* Bytes read but not yet consumed via __archive_read_consume() */ size_t unconsumed; /* Information about entry we're currently reading. */ struct zip_entry *entry; int64_t entry_bytes_remaining; /* These count the number of bytes actually read for the entry. */ int64_t entry_compressed_bytes_read; int64_t entry_uncompressed_bytes_read; /* Running CRC32 of the decompressed data */ unsigned long entry_crc32; unsigned long (*crc32func)(unsigned long, const void *, size_t); char ignore_crc32; /* Flags to mark progress of decompression. */ char decompress_init; char end_of_entry; #ifdef HAVE_ZLIB_H unsigned char *uncompressed_buffer; size_t uncompressed_buffer_size; z_stream stream; char stream_valid; #endif struct archive_string_conv *sconv; struct archive_string_conv *sconv_default; struct archive_string_conv *sconv_utf8; int init_default_conversion; int process_mac_extensions; char init_decryption; /* Decryption buffer. */ /* * The decrypted data starts at decrypted_ptr and * extends for decrypted_bytes_remaining. Decryption * adds new data to the end of this block, data is returned * to clients from the beginning. When the block hits the * end of decrypted_buffer, it has to be shuffled back to * the beginning of the buffer. */ unsigned char *decrypted_buffer; unsigned char *decrypted_ptr; size_t decrypted_buffer_size; size_t decrypted_bytes_remaining; size_t decrypted_unconsumed_bytes; /* Traditional PKWARE decryption. */ struct trad_enc_ctx tctx; char tctx_valid; /* WinZip AES decryption. */ /* Contexts used for AES decryption. */ archive_crypto_ctx cctx; char cctx_valid; archive_hmac_sha1_ctx hctx; char hctx_valid; /* Strong encryption's decryption header information. */ unsigned iv_size; unsigned alg_id; unsigned bit_len; unsigned flags; unsigned erd_size; unsigned v_size; unsigned v_crc32; uint8_t *iv; uint8_t *erd; uint8_t *v_data; }; /* Many systems define min or MIN, but not all. */ #define zipmin(a,b) ((a) < (b) ? (a) : (b)) /* ------------------------------------------------------------------------ */ /* Traditional PKWARE Decryption functions. */ static void trad_enc_update_keys(struct trad_enc_ctx *ctx, uint8_t c) { uint8_t t; #define CRC32(c, b) (crc32(c ^ 0xffffffffUL, &b, 1) ^ 0xffffffffUL) ctx->keys[0] = CRC32(ctx->keys[0], c); ctx->keys[1] = (ctx->keys[1] + (ctx->keys[0] & 0xff)) * 134775813L + 1; t = (ctx->keys[1] >> 24) & 0xff; ctx->keys[2] = CRC32(ctx->keys[2], t); #undef CRC32 } static uint8_t trad_enc_decrypt_byte(struct trad_enc_ctx *ctx) { unsigned temp = ctx->keys[2] | 2; return (uint8_t)((temp * (temp ^ 1)) >> 8) & 0xff; } static void trad_enc_decrypt_update(struct trad_enc_ctx *ctx, const uint8_t *in, size_t in_len, uint8_t *out, size_t out_len) { unsigned i, max; max = (unsigned)((in_len < out_len)? in_len: out_len); for (i = 0; i < max; i++) { uint8_t t = in[i] ^ trad_enc_decrypt_byte(ctx); out[i] = t; trad_enc_update_keys(ctx, t); } } static int trad_enc_init(struct trad_enc_ctx *ctx, const char *pw, size_t pw_len, const uint8_t *key, size_t key_len, uint8_t *crcchk) { uint8_t header[12]; if (key_len < 12) { *crcchk = 0xff; return -1; } ctx->keys[0] = 305419896L; ctx->keys[1] = 591751049L; ctx->keys[2] = 878082192L; for (;pw_len; --pw_len) trad_enc_update_keys(ctx, *pw++); trad_enc_decrypt_update(ctx, key, 12, header, 12); /* Return the last byte for CRC check. */ *crcchk = header[11]; return 0; } #if 0 static void crypt_derive_key_sha1(const void *p, int size, unsigned char *key, int key_size) { #define MD_SIZE 20 archive_sha1_ctx ctx; unsigned char md1[MD_SIZE]; unsigned char md2[MD_SIZE * 2]; unsigned char mkb[64]; int i; archive_sha1_init(&ctx); archive_sha1_update(&ctx, p, size); archive_sha1_final(&ctx, md1); memset(mkb, 0x36, sizeof(mkb)); for (i = 0; i < MD_SIZE; i++) mkb[i] ^= md1[i]; archive_sha1_init(&ctx); archive_sha1_update(&ctx, mkb, sizeof(mkb)); archive_sha1_final(&ctx, md2); memset(mkb, 0x5C, sizeof(mkb)); for (i = 0; i < MD_SIZE; i++) mkb[i] ^= md1[i]; archive_sha1_init(&ctx); archive_sha1_update(&ctx, mkb, sizeof(mkb)); archive_sha1_final(&ctx, md2 + MD_SIZE); if (key_size > 32) key_size = 32; memcpy(key, md2, key_size); #undef MD_SIZE } #endif /* * Common code for streaming or seeking modes. * * Includes code to read local file headers, decompress data * from entry bodies, and common API. */ static unsigned long real_crc32(unsigned long crc, const void *buff, size_t len) { return crc32(crc, buff, (unsigned int)len); } /* Used by "ignorecrc32" option to speed up tests. */ static unsigned long fake_crc32(unsigned long crc, const void *buff, size_t len) { (void)crc; /* UNUSED */ (void)buff; /* UNUSED */ (void)len; /* UNUSED */ return 0; } static const struct { int id; const char * name; } compression_methods[] = { {0, "uncompressed"}, /* The file is stored (no compression) */ {1, "shrinking"}, /* The file is Shrunk */ {2, "reduced-1"}, /* The file is Reduced with compression factor 1 */ {3, "reduced-2"}, /* The file is Reduced with compression factor 2 */ {4, "reduced-3"}, /* The file is Reduced with compression factor 3 */ {5, "reduced-4"}, /* The file is Reduced with compression factor 4 */ {6, "imploded"}, /* The file is Imploded */ {7, "reserved"}, /* Reserved for Tokenizing compression algorithm */ {8, "deflation"}, /* The file is Deflated */ {9, "deflation-64-bit"}, /* Enhanced Deflating using Deflate64(tm) */ {10, "ibm-terse"},/* PKWARE Data Compression Library Imploding * (old IBM TERSE) */ {11, "reserved"}, /* Reserved by PKWARE */ {12, "bzip"}, /* File is compressed using BZIP2 algorithm */ {13, "reserved"}, /* Reserved by PKWARE */ {14, "lzma"}, /* LZMA (EFS) */ {15, "reserved"}, /* Reserved by PKWARE */ {16, "reserved"}, /* Reserved by PKWARE */ {17, "reserved"}, /* Reserved by PKWARE */ {18, "ibm-terse-new"}, /* File is compressed using IBM TERSE (new) */ {19, "ibm-lz777"},/* IBM LZ77 z Architecture (PFS) */ {97, "wav-pack"}, /* WavPack compressed data */ {98, "ppmd-1"}, /* PPMd version I, Rev 1 */ {99, "aes"} /* WinZip AES encryption */ }; static const char * compression_name(const int compression) { static const int num_compression_methods = sizeof(compression_methods)/sizeof(compression_methods[0]); int i=0; while(compression >= 0 && i < num_compression_methods) { if (compression_methods[i].id == compression) return compression_methods[i].name; i++; } return "??"; } /* Convert an MSDOS-style date/time into Unix-style time. */ static time_t zip_time(const char *p) { int msTime, msDate; struct tm ts; msTime = (0xff & (unsigned)p[0]) + 256 * (0xff & (unsigned)p[1]); msDate = (0xff & (unsigned)p[2]) + 256 * (0xff & (unsigned)p[3]); memset(&ts, 0, sizeof(ts)); ts.tm_year = ((msDate >> 9) & 0x7f) + 80; /* Years since 1900. */ ts.tm_mon = ((msDate >> 5) & 0x0f) - 1; /* Month number. */ ts.tm_mday = msDate & 0x1f; /* Day of month. */ ts.tm_hour = (msTime >> 11) & 0x1f; ts.tm_min = (msTime >> 5) & 0x3f; ts.tm_sec = (msTime << 1) & 0x3e; ts.tm_isdst = -1; return mktime(&ts); } /* * The extra data is stored as a list of * id1+size1+data1 + id2+size2+data2 ... * triplets. id and size are 2 bytes each. */ static int process_extra(struct archive_read *a, const char *p, size_t extra_length, struct zip_entry* zip_entry) { unsigned offset = 0; if (extra_length == 0) { return ARCHIVE_OK; } if (extra_length < 4) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Too-small extra data: Need at least 4 bytes, but only found %d bytes", (int)extra_length); return ARCHIVE_FAILED; } while (offset <= extra_length - 4) { unsigned short headerid = archive_le16dec(p + offset); unsigned short datasize = archive_le16dec(p + offset + 2); offset += 4; if (offset + datasize > extra_length) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Extra data overflow: Need %d bytes but only found %d bytes", (int)datasize, (int)(extra_length - offset)); return ARCHIVE_FAILED; } #ifdef DEBUG fprintf(stderr, "Header id 0x%04x, length %d\n", headerid, datasize); #endif switch (headerid) { case 0x0001: /* Zip64 extended information extra field. */ zip_entry->flags |= LA_USED_ZIP64; if (zip_entry->uncompressed_size == 0xffffffff) { uint64_t t = 0; if (datasize < 8 || (t = archive_le64dec(p + offset)) > INT64_MAX) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Malformed 64-bit uncompressed size"); return ARCHIVE_FAILED; } zip_entry->uncompressed_size = t; offset += 8; datasize -= 8; } if (zip_entry->compressed_size == 0xffffffff) { uint64_t t = 0; if (datasize < 8 || (t = archive_le64dec(p + offset)) > INT64_MAX) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Malformed 64-bit compressed size"); return ARCHIVE_FAILED; } zip_entry->compressed_size = t; offset += 8; datasize -= 8; } if (zip_entry->local_header_offset == 0xffffffff) { uint64_t t = 0; if (datasize < 8 || (t = archive_le64dec(p + offset)) > INT64_MAX) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Malformed 64-bit local header offset"); return ARCHIVE_FAILED; } zip_entry->local_header_offset = t; offset += 8; datasize -= 8; } /* archive_le32dec(p + offset) gives disk * on which file starts, but we don't handle * multi-volume Zip files. */ break; #ifdef DEBUG case 0x0017: { /* Strong encryption field. */ if (archive_le16dec(p + offset) == 2) { unsigned algId = archive_le16dec(p + offset + 2); unsigned bitLen = archive_le16dec(p + offset + 4); int flags = archive_le16dec(p + offset + 6); fprintf(stderr, "algId=0x%04x, bitLen=%u, " "flgas=%d\n", algId, bitLen,flags); } break; } #endif case 0x5455: { /* Extended time field "UT". */ int flags = p[offset]; offset++; datasize--; /* Flag bits indicate which dates are present. */ if (flags & 0x01) { #ifdef DEBUG fprintf(stderr, "mtime: %lld -> %d\n", (long long)zip_entry->mtime, archive_le32dec(p + offset)); #endif if (datasize < 4) break; zip_entry->mtime = archive_le32dec(p + offset); offset += 4; datasize -= 4; } if (flags & 0x02) { if (datasize < 4) break; zip_entry->atime = archive_le32dec(p + offset); offset += 4; datasize -= 4; } if (flags & 0x04) { if (datasize < 4) break; zip_entry->ctime = archive_le32dec(p + offset); offset += 4; datasize -= 4; } break; } case 0x5855: { /* Info-ZIP Unix Extra Field (old version) "UX". */ if (datasize >= 8) { zip_entry->atime = archive_le32dec(p + offset); zip_entry->mtime = archive_le32dec(p + offset + 4); } if (datasize >= 12) { zip_entry->uid = archive_le16dec(p + offset + 8); zip_entry->gid = archive_le16dec(p + offset + 10); } break; } case 0x6c78: { /* Experimental 'xl' field */ /* * Introduced Dec 2013 to provide a way to * include external file attributes (and other * fields that ordinarily appear only in * central directory) in local file header. * This provides file type and permission * information necessary to support full * streaming extraction. Currently being * discussed with other Zip developers * ... subject to change. * * Format: * The field starts with a bitmap that specifies * which additional fields are included. The * bitmap is variable length and can be extended in * the future. * * n bytes - feature bitmap: first byte has low-order * 7 bits. If high-order bit is set, a subsequent * byte holds the next 7 bits, etc. * * if bitmap & 1, 2 byte "version made by" * if bitmap & 2, 2 byte "internal file attributes" * if bitmap & 4, 4 byte "external file attributes" * if bitmap & 8, 2 byte comment length + n byte comment */ int bitmap, bitmap_last; if (datasize < 1) break; bitmap_last = bitmap = 0xff & p[offset]; offset += 1; datasize -= 1; /* We only support first 7 bits of bitmap; skip rest. */ while ((bitmap_last & 0x80) != 0 && datasize >= 1) { bitmap_last = p[offset]; offset += 1; datasize -= 1; } if (bitmap & 1) { /* 2 byte "version made by" */ if (datasize < 2) break; zip_entry->system = archive_le16dec(p + offset) >> 8; offset += 2; datasize -= 2; } if (bitmap & 2) { /* 2 byte "internal file attributes" */ uint32_t internal_attributes; if (datasize < 2) break; internal_attributes = archive_le16dec(p + offset); /* Not used by libarchive at present. */ (void)internal_attributes; /* UNUSED */ offset += 2; datasize -= 2; } if (bitmap & 4) { /* 4 byte "external file attributes" */ uint32_t external_attributes; if (datasize < 4) break; external_attributes = archive_le32dec(p + offset); if (zip_entry->system == 3) { zip_entry->mode = external_attributes >> 16; } else if (zip_entry->system == 0) { // Interpret MSDOS directory bit if (0x10 == (external_attributes & 0x10)) { zip_entry->mode = AE_IFDIR | 0775; } else { zip_entry->mode = AE_IFREG | 0664; } if (0x01 == (external_attributes & 0x01)) { // Read-only bit; strip write permissions zip_entry->mode &= 0555; } } else { zip_entry->mode = 0; } offset += 4; datasize -= 4; } if (bitmap & 8) { /* 2 byte comment length + comment */ uint32_t comment_length; if (datasize < 2) break; comment_length = archive_le16dec(p + offset); offset += 2; datasize -= 2; if (datasize < comment_length) break; /* Comment is not supported by libarchive */ offset += comment_length; datasize -= comment_length; } break; } case 0x7855: /* Info-ZIP Unix Extra Field (type 2) "Ux". */ #ifdef DEBUG fprintf(stderr, "uid %d gid %d\n", archive_le16dec(p + offset), archive_le16dec(p + offset + 2)); #endif if (datasize >= 2) zip_entry->uid = archive_le16dec(p + offset); if (datasize >= 4) zip_entry->gid = archive_le16dec(p + offset + 2); break; case 0x7875: { /* Info-Zip Unix Extra Field (type 3) "ux". */ int uidsize = 0, gidsize = 0; /* TODO: support arbitrary uidsize/gidsize. */ if (datasize >= 1 && p[offset] == 1) {/* version=1 */ if (datasize >= 4) { /* get a uid size. */ uidsize = 0xff & (int)p[offset+1]; if (uidsize == 2) zip_entry->uid = archive_le16dec( p + offset + 2); else if (uidsize == 4 && datasize >= 6) zip_entry->uid = archive_le32dec( p + offset + 2); } if (datasize >= (2 + uidsize + 3)) { /* get a gid size. */ gidsize = 0xff & (int)p[offset+2+uidsize]; if (gidsize == 2) zip_entry->gid = archive_le16dec( p+offset+2+uidsize+1); else if (gidsize == 4 && datasize >= (2 + uidsize + 5)) zip_entry->gid = archive_le32dec( p+offset+2+uidsize+1); } } break; } case 0x9901: /* WinZip AES extra data field. */ + if (datasize < 6) { + archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, + "Incomplete AES field"); + return ARCHIVE_FAILED; + } if (p[offset + 2] == 'A' && p[offset + 3] == 'E') { /* Vendor version. */ zip_entry->aes_extra.vendor = archive_le16dec(p + offset); /* AES encryption strength. */ zip_entry->aes_extra.strength = p[offset + 4]; /* Actual compression method. */ zip_entry->aes_extra.compression = p[offset + 5]; } break; default: break; } offset += datasize; } if (offset != extra_length) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Malformed extra data: Consumed %d bytes of %d bytes", (int)offset, (int)extra_length); return ARCHIVE_FAILED; } return ARCHIVE_OK; } /* * Assumes file pointer is at beginning of local file header. */ static int zip_read_local_file_header(struct archive_read *a, struct archive_entry *entry, struct zip *zip) { const char *p; const void *h; const wchar_t *wp; const char *cp; size_t len, filename_length, extra_length; struct archive_string_conv *sconv; struct zip_entry *zip_entry = zip->entry; struct zip_entry zip_entry_central_dir; int ret = ARCHIVE_OK; char version; /* Save a copy of the original for consistency checks. */ zip_entry_central_dir = *zip_entry; zip->decompress_init = 0; zip->end_of_entry = 0; zip->entry_uncompressed_bytes_read = 0; zip->entry_compressed_bytes_read = 0; zip->entry_crc32 = zip->crc32func(0, NULL, 0); /* Setup default conversion. */ if (zip->sconv == NULL && !zip->init_default_conversion) { zip->sconv_default = archive_string_default_conversion_for_read(&(a->archive)); zip->init_default_conversion = 1; } if ((p = __archive_read_ahead(a, 30, NULL)) == NULL) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated ZIP file header"); return (ARCHIVE_FATAL); } if (memcmp(p, "PK\003\004", 4) != 0) { archive_set_error(&a->archive, -1, "Damaged Zip archive"); return ARCHIVE_FATAL; } version = p[4]; zip_entry->system = p[5]; zip_entry->zip_flags = archive_le16dec(p + 6); if (zip_entry->zip_flags & (ZIP_ENCRYPTED | ZIP_STRONG_ENCRYPTED)) { zip->has_encrypted_entries = 1; archive_entry_set_is_data_encrypted(entry, 1); if (zip_entry->zip_flags & ZIP_CENTRAL_DIRECTORY_ENCRYPTED && zip_entry->zip_flags & ZIP_ENCRYPTED && zip_entry->zip_flags & ZIP_STRONG_ENCRYPTED) { archive_entry_set_is_metadata_encrypted(entry, 1); return ARCHIVE_FATAL; } } zip->init_decryption = (zip_entry->zip_flags & ZIP_ENCRYPTED); zip_entry->compression = (char)archive_le16dec(p + 8); zip_entry->mtime = zip_time(p + 10); zip_entry->crc32 = archive_le32dec(p + 14); if (zip_entry->zip_flags & ZIP_LENGTH_AT_END) zip_entry->decdat = p[11]; else zip_entry->decdat = p[17]; zip_entry->compressed_size = archive_le32dec(p + 18); zip_entry->uncompressed_size = archive_le32dec(p + 22); filename_length = archive_le16dec(p + 26); extra_length = archive_le16dec(p + 28); __archive_read_consume(a, 30); /* Read the filename. */ if ((h = __archive_read_ahead(a, filename_length, NULL)) == NULL) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated ZIP file header"); return (ARCHIVE_FATAL); } if (zip_entry->zip_flags & ZIP_UTF8_NAME) { /* The filename is stored to be UTF-8. */ if (zip->sconv_utf8 == NULL) { zip->sconv_utf8 = archive_string_conversion_from_charset( &a->archive, "UTF-8", 1); if (zip->sconv_utf8 == NULL) return (ARCHIVE_FATAL); } sconv = zip->sconv_utf8; } else if (zip->sconv != NULL) sconv = zip->sconv; else sconv = zip->sconv_default; if (archive_entry_copy_pathname_l(entry, h, filename_length, sconv) != 0) { if (errno == ENOMEM) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory for Pathname"); return (ARCHIVE_FATAL); } archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Pathname cannot be converted " "from %s to current locale.", archive_string_conversion_charset_name(sconv)); ret = ARCHIVE_WARN; } __archive_read_consume(a, filename_length); /* Read the extra data. */ if ((h = __archive_read_ahead(a, extra_length, NULL)) == NULL) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated ZIP file header"); return (ARCHIVE_FATAL); } if (ARCHIVE_OK != process_extra(a, h, extra_length, zip_entry)) { return ARCHIVE_FATAL; } __archive_read_consume(a, extra_length); /* Work around a bug in Info-Zip: When reading from a pipe, it * stats the pipe instead of synthesizing a file entry. */ if ((zip_entry->mode & AE_IFMT) == AE_IFIFO) { zip_entry->mode &= ~ AE_IFMT; zip_entry->mode |= AE_IFREG; } /* If the mode is totally empty, set some sane default. */ if (zip_entry->mode == 0) { zip_entry->mode |= 0664; } /* Make sure that entries with a trailing '/' are marked as directories * even if the External File Attributes contains bogus values. If this * is not a directory and there is no type, assume regularfile. */ if ((zip_entry->mode & AE_IFMT) != AE_IFDIR) { int has_slash; wp = archive_entry_pathname_w(entry); if (wp != NULL) { len = wcslen(wp); has_slash = len > 0 && wp[len - 1] == L'/'; } else { cp = archive_entry_pathname(entry); len = (cp != NULL)?strlen(cp):0; has_slash = len > 0 && cp[len - 1] == '/'; } /* Correct file type as needed. */ if (has_slash) { zip_entry->mode &= ~AE_IFMT; zip_entry->mode |= AE_IFDIR; zip_entry->mode |= 0111; } else if ((zip_entry->mode & AE_IFMT) == 0) { zip_entry->mode |= AE_IFREG; } } /* Make sure directories end in '/' */ if ((zip_entry->mode & AE_IFMT) == AE_IFDIR) { wp = archive_entry_pathname_w(entry); if (wp != NULL) { len = wcslen(wp); if (len > 0 && wp[len - 1] != L'/') { struct archive_wstring s; archive_string_init(&s); archive_wstrcat(&s, wp); archive_wstrappend_wchar(&s, L'/'); archive_entry_copy_pathname_w(entry, s.s); archive_wstring_free(&s); } } else { cp = archive_entry_pathname(entry); len = (cp != NULL)?strlen(cp):0; if (len > 0 && cp[len - 1] != '/') { struct archive_string s; archive_string_init(&s); archive_strcat(&s, cp); archive_strappend_char(&s, '/'); archive_entry_set_pathname(entry, s.s); archive_string_free(&s); } } } if (zip_entry->flags & LA_FROM_CENTRAL_DIRECTORY) { /* If this came from the central dir, it's size info * is definitive, so ignore the length-at-end flag. */ zip_entry->zip_flags &= ~ZIP_LENGTH_AT_END; /* If local header is missing a value, use the one from the central directory. If both have it, warn about mismatches. */ if (zip_entry->crc32 == 0) { zip_entry->crc32 = zip_entry_central_dir.crc32; } else if (!zip->ignore_crc32 && zip_entry->crc32 != zip_entry_central_dir.crc32) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Inconsistent CRC32 values"); ret = ARCHIVE_WARN; } if (zip_entry->compressed_size == 0) { zip_entry->compressed_size = zip_entry_central_dir.compressed_size; } else if (zip_entry->compressed_size != zip_entry_central_dir.compressed_size) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Inconsistent compressed size: " "%jd in central directory, %jd in local header", (intmax_t)zip_entry_central_dir.compressed_size, (intmax_t)zip_entry->compressed_size); ret = ARCHIVE_WARN; } if (zip_entry->uncompressed_size == 0) { zip_entry->uncompressed_size = zip_entry_central_dir.uncompressed_size; } else if (zip_entry->uncompressed_size != zip_entry_central_dir.uncompressed_size) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Inconsistent uncompressed size: " "%jd in central directory, %jd in local header", (intmax_t)zip_entry_central_dir.uncompressed_size, (intmax_t)zip_entry->uncompressed_size); ret = ARCHIVE_WARN; } } /* Populate some additional entry fields: */ archive_entry_set_mode(entry, zip_entry->mode); archive_entry_set_uid(entry, zip_entry->uid); archive_entry_set_gid(entry, zip_entry->gid); archive_entry_set_mtime(entry, zip_entry->mtime, 0); archive_entry_set_ctime(entry, zip_entry->ctime, 0); archive_entry_set_atime(entry, zip_entry->atime, 0); if ((zip->entry->mode & AE_IFMT) == AE_IFLNK) { size_t linkname_length; if (zip_entry->compressed_size > 64 * 1024) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Zip file with oversized link entry"); return ARCHIVE_FATAL; } linkname_length = (size_t)zip_entry->compressed_size; archive_entry_set_size(entry, 0); p = __archive_read_ahead(a, linkname_length, NULL); if (p == NULL) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Truncated Zip file"); return ARCHIVE_FATAL; } sconv = zip->sconv; if (sconv == NULL && (zip->entry->zip_flags & ZIP_UTF8_NAME)) sconv = zip->sconv_utf8; if (sconv == NULL) sconv = zip->sconv_default; if (archive_entry_copy_symlink_l(entry, p, linkname_length, sconv) != 0) { if (errno != ENOMEM && sconv == zip->sconv_utf8 && (zip->entry->zip_flags & ZIP_UTF8_NAME)) archive_entry_copy_symlink_l(entry, p, linkname_length, NULL); if (errno == ENOMEM) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory for Symlink"); return (ARCHIVE_FATAL); } /* * Since there is no character-set regulation for * symlink name, do not report the conversion error * in an automatic conversion. */ if (sconv != zip->sconv_utf8 || (zip->entry->zip_flags & ZIP_UTF8_NAME) == 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Symlink cannot be converted " "from %s to current locale.", archive_string_conversion_charset_name( sconv)); ret = ARCHIVE_WARN; } } zip_entry->uncompressed_size = zip_entry->compressed_size = 0; if (__archive_read_consume(a, linkname_length) < 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Read error skipping symlink target name"); return ARCHIVE_FATAL; } } else if (0 == (zip_entry->zip_flags & ZIP_LENGTH_AT_END) || zip_entry->uncompressed_size > 0) { /* Set the size only if it's meaningful. */ archive_entry_set_size(entry, zip_entry->uncompressed_size); } zip->entry_bytes_remaining = zip_entry->compressed_size; /* If there's no body, force read_data() to return EOF immediately. */ if (0 == (zip_entry->zip_flags & ZIP_LENGTH_AT_END) && zip->entry_bytes_remaining < 1) zip->end_of_entry = 1; /* Set up a more descriptive format name. */ archive_string_sprintf(&zip->format_name, "ZIP %d.%d (%s)", version / 10, version % 10, compression_name(zip->entry->compression)); a->archive.archive_format_name = zip->format_name.s; return (ret); } static int check_authentication_code(struct archive_read *a, const void *_p) { struct zip *zip = (struct zip *)(a->format->data); /* Check authentication code. */ if (zip->hctx_valid) { const void *p; uint8_t hmac[20]; size_t hmac_len = 20; int cmp; archive_hmac_sha1_final(&zip->hctx, hmac, &hmac_len); if (_p == NULL) { /* Read authentication code. */ p = __archive_read_ahead(a, AUTH_CODE_SIZE, NULL); if (p == NULL) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated ZIP file data"); return (ARCHIVE_FATAL); } } else { p = _p; } cmp = memcmp(hmac, p, AUTH_CODE_SIZE); __archive_read_consume(a, AUTH_CODE_SIZE); if (cmp != 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "ZIP bad Authentication code"); return (ARCHIVE_WARN); } } return (ARCHIVE_OK); } /* * Read "uncompressed" data. There are three cases: * 1) We know the size of the data. This is always true for the * seeking reader (we've examined the Central Directory already). * 2) ZIP_LENGTH_AT_END was set, but only the CRC was deferred. * Info-ZIP seems to do this; we know the size but have to grab * the CRC from the data descriptor afterwards. * 3) We're streaming and ZIP_LENGTH_AT_END was specified and * we have no size information. In this case, we can do pretty * well by watching for the data descriptor record. The data * descriptor is 16 bytes and includes a computed CRC that should * provide a strong check. * * TODO: Technically, the PK\007\010 signature is optional. * In the original spec, the data descriptor contained CRC * and size fields but had no leading signature. In practice, * newer writers seem to provide the signature pretty consistently. * * For uncompressed data, the PK\007\010 marker seems essential * to be sure we've actually seen the end of the entry. * * Returns ARCHIVE_OK if successful, ARCHIVE_FATAL otherwise, sets * zip->end_of_entry if it consumes all of the data. */ static int zip_read_data_none(struct archive_read *a, const void **_buff, size_t *size, int64_t *offset) { struct zip *zip; const char *buff; ssize_t bytes_avail; int r; (void)offset; /* UNUSED */ zip = (struct zip *)(a->format->data); if (zip->entry->zip_flags & ZIP_LENGTH_AT_END) { const char *p; ssize_t grabbing_bytes = 24; if (zip->hctx_valid) grabbing_bytes += AUTH_CODE_SIZE; /* Grab at least 24 bytes. */ buff = __archive_read_ahead(a, grabbing_bytes, &bytes_avail); if (bytes_avail < grabbing_bytes) { /* Zip archives have end-of-archive markers that are longer than this, so a failure to get at least 24 bytes really does indicate a truncated file. */ archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated ZIP file data"); return (ARCHIVE_FATAL); } /* Check for a complete PK\007\010 signature, followed * by the correct 4-byte CRC. */ p = buff; if (zip->hctx_valid) p += AUTH_CODE_SIZE; if (p[0] == 'P' && p[1] == 'K' && p[2] == '\007' && p[3] == '\010' && (archive_le32dec(p + 4) == zip->entry_crc32 || zip->ignore_crc32 || (zip->hctx_valid && zip->entry->aes_extra.vendor == AES_VENDOR_AE_2))) { if (zip->entry->flags & LA_USED_ZIP64) { uint64_t compressed, uncompressed; zip->entry->crc32 = archive_le32dec(p + 4); compressed = archive_le64dec(p + 8); uncompressed = archive_le64dec(p + 16); if (compressed > INT64_MAX || uncompressed > INT64_MAX) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Overflow of 64-bit file sizes"); return ARCHIVE_FAILED; } zip->entry->compressed_size = compressed; zip->entry->uncompressed_size = uncompressed; zip->unconsumed = 24; } else { zip->entry->crc32 = archive_le32dec(p + 4); zip->entry->compressed_size = archive_le32dec(p + 8); zip->entry->uncompressed_size = archive_le32dec(p + 12); zip->unconsumed = 16; } if (zip->hctx_valid) { r = check_authentication_code(a, buff); if (r != ARCHIVE_OK) return (r); } zip->end_of_entry = 1; return (ARCHIVE_OK); } /* If not at EOF, ensure we consume at least one byte. */ ++p; /* Scan forward until we see where a PK\007\010 signature * might be. */ /* Return bytes up until that point. On the next call, * the code above will verify the data descriptor. */ while (p < buff + bytes_avail - 4) { if (p[3] == 'P') { p += 3; } else if (p[3] == 'K') { p += 2; } else if (p[3] == '\007') { p += 1; } else if (p[3] == '\010' && p[2] == '\007' && p[1] == 'K' && p[0] == 'P') { if (zip->hctx_valid) p -= AUTH_CODE_SIZE; break; } else { p += 4; } } bytes_avail = p - buff; } else { if (zip->entry_bytes_remaining == 0) { zip->end_of_entry = 1; if (zip->hctx_valid) { r = check_authentication_code(a, NULL); if (r != ARCHIVE_OK) return (r); } return (ARCHIVE_OK); } /* Grab a bunch of bytes. */ buff = __archive_read_ahead(a, 1, &bytes_avail); if (bytes_avail <= 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated ZIP file data"); return (ARCHIVE_FATAL); } if (bytes_avail > zip->entry_bytes_remaining) bytes_avail = (ssize_t)zip->entry_bytes_remaining; } if (zip->tctx_valid || zip->cctx_valid) { size_t dec_size = bytes_avail; if (dec_size > zip->decrypted_buffer_size) dec_size = zip->decrypted_buffer_size; if (zip->tctx_valid) { trad_enc_decrypt_update(&zip->tctx, (const uint8_t *)buff, dec_size, zip->decrypted_buffer, dec_size); } else { size_t dsize = dec_size; archive_hmac_sha1_update(&zip->hctx, (const uint8_t *)buff, dec_size); archive_decrypto_aes_ctr_update(&zip->cctx, (const uint8_t *)buff, dec_size, zip->decrypted_buffer, &dsize); } bytes_avail = dec_size; buff = (const char *)zip->decrypted_buffer; } *size = bytes_avail; zip->entry_bytes_remaining -= bytes_avail; zip->entry_uncompressed_bytes_read += bytes_avail; zip->entry_compressed_bytes_read += bytes_avail; zip->unconsumed += bytes_avail; *_buff = buff; return (ARCHIVE_OK); } #ifdef HAVE_ZLIB_H static int zip_deflate_init(struct archive_read *a, struct zip *zip) { int r; /* If we haven't yet read any data, initialize the decompressor. */ if (!zip->decompress_init) { if (zip->stream_valid) r = inflateReset(&zip->stream); else r = inflateInit2(&zip->stream, -15 /* Don't check for zlib header */); if (r != Z_OK) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Can't initialize ZIP decompression."); return (ARCHIVE_FATAL); } /* Stream structure has been set up. */ zip->stream_valid = 1; /* We've initialized decompression for this stream. */ zip->decompress_init = 1; } return (ARCHIVE_OK); } static int zip_read_data_deflate(struct archive_read *a, const void **buff, size_t *size, int64_t *offset) { struct zip *zip; ssize_t bytes_avail; const void *compressed_buff, *sp; int r; (void)offset; /* UNUSED */ zip = (struct zip *)(a->format->data); /* If the buffer hasn't been allocated, allocate it now. */ if (zip->uncompressed_buffer == NULL) { zip->uncompressed_buffer_size = 256 * 1024; zip->uncompressed_buffer = (unsigned char *)malloc(zip->uncompressed_buffer_size); if (zip->uncompressed_buffer == NULL) { archive_set_error(&a->archive, ENOMEM, "No memory for ZIP decompression"); return (ARCHIVE_FATAL); } } r = zip_deflate_init(a, zip); if (r != ARCHIVE_OK) return (r); /* * Note: '1' here is a performance optimization. * Recall that the decompression layer returns a count of * available bytes; asking for more than that forces the * decompressor to combine reads by copying data. */ compressed_buff = sp = __archive_read_ahead(a, 1, &bytes_avail); if (0 == (zip->entry->zip_flags & ZIP_LENGTH_AT_END) && bytes_avail > zip->entry_bytes_remaining) { bytes_avail = (ssize_t)zip->entry_bytes_remaining; } if (bytes_avail < 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated ZIP file body"); return (ARCHIVE_FATAL); } if (zip->tctx_valid || zip->cctx_valid) { if (zip->decrypted_bytes_remaining < (size_t)bytes_avail) { size_t buff_remaining = (zip->decrypted_buffer + zip->decrypted_buffer_size) - (zip->decrypted_ptr + zip->decrypted_bytes_remaining); if (buff_remaining > (size_t)bytes_avail) buff_remaining = (size_t)bytes_avail; if (0 == (zip->entry->zip_flags & ZIP_LENGTH_AT_END) && zip->entry_bytes_remaining > 0) { if ((int64_t)(zip->decrypted_bytes_remaining + buff_remaining) > zip->entry_bytes_remaining) { if (zip->entry_bytes_remaining < (int64_t)zip->decrypted_bytes_remaining) buff_remaining = 0; else buff_remaining = (size_t)zip->entry_bytes_remaining - zip->decrypted_bytes_remaining; } } if (buff_remaining > 0) { if (zip->tctx_valid) { trad_enc_decrypt_update(&zip->tctx, compressed_buff, buff_remaining, zip->decrypted_ptr + zip->decrypted_bytes_remaining, buff_remaining); } else { size_t dsize = buff_remaining; archive_decrypto_aes_ctr_update( &zip->cctx, compressed_buff, buff_remaining, zip->decrypted_ptr + zip->decrypted_bytes_remaining, &dsize); } zip->decrypted_bytes_remaining += buff_remaining; } } bytes_avail = zip->decrypted_bytes_remaining; compressed_buff = (const char *)zip->decrypted_ptr; } /* * A bug in zlib.h: stream.next_in should be marked 'const' * but isn't (the library never alters data through the * next_in pointer, only reads it). The result: this ugly * cast to remove 'const'. */ zip->stream.next_in = (Bytef *)(uintptr_t)(const void *)compressed_buff; zip->stream.avail_in = (uInt)bytes_avail; zip->stream.total_in = 0; zip->stream.next_out = zip->uncompressed_buffer; zip->stream.avail_out = (uInt)zip->uncompressed_buffer_size; zip->stream.total_out = 0; r = inflate(&zip->stream, 0); switch (r) { case Z_OK: break; case Z_STREAM_END: zip->end_of_entry = 1; break; case Z_MEM_ERROR: archive_set_error(&a->archive, ENOMEM, "Out of memory for ZIP decompression"); return (ARCHIVE_FATAL); default: archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "ZIP decompression failed (%d)", r); return (ARCHIVE_FATAL); } /* Consume as much as the compressor actually used. */ bytes_avail = zip->stream.total_in; if (zip->tctx_valid || zip->cctx_valid) { zip->decrypted_bytes_remaining -= bytes_avail; if (zip->decrypted_bytes_remaining == 0) zip->decrypted_ptr = zip->decrypted_buffer; else zip->decrypted_ptr += bytes_avail; } /* Calculate compressed data as much as we used.*/ if (zip->hctx_valid) archive_hmac_sha1_update(&zip->hctx, sp, bytes_avail); __archive_read_consume(a, bytes_avail); zip->entry_bytes_remaining -= bytes_avail; zip->entry_compressed_bytes_read += bytes_avail; *size = zip->stream.total_out; zip->entry_uncompressed_bytes_read += zip->stream.total_out; *buff = zip->uncompressed_buffer; if (zip->end_of_entry && zip->hctx_valid) { r = check_authentication_code(a, NULL); if (r != ARCHIVE_OK) return (r); } if (zip->end_of_entry && (zip->entry->zip_flags & ZIP_LENGTH_AT_END)) { const char *p; if (NULL == (p = __archive_read_ahead(a, 24, NULL))) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated ZIP end-of-file record"); return (ARCHIVE_FATAL); } /* Consume the optional PK\007\010 marker. */ if (p[0] == 'P' && p[1] == 'K' && p[2] == '\007' && p[3] == '\010') { p += 4; zip->unconsumed = 4; } if (zip->entry->flags & LA_USED_ZIP64) { uint64_t compressed, uncompressed; zip->entry->crc32 = archive_le32dec(p); compressed = archive_le64dec(p + 4); uncompressed = archive_le64dec(p + 12); if (compressed > INT64_MAX || uncompressed > INT64_MAX) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Overflow of 64-bit file sizes"); return ARCHIVE_FAILED; } zip->entry->compressed_size = compressed; zip->entry->uncompressed_size = uncompressed; zip->unconsumed += 20; } else { zip->entry->crc32 = archive_le32dec(p); zip->entry->compressed_size = archive_le32dec(p + 4); zip->entry->uncompressed_size = archive_le32dec(p + 8); zip->unconsumed += 12; } } return (ARCHIVE_OK); } #endif static int read_decryption_header(struct archive_read *a) { struct zip *zip = (struct zip *)(a->format->data); const char *p; unsigned int remaining_size; unsigned int ts; /* * Read an initialization vector data field. */ p = __archive_read_ahead(a, 2, NULL); if (p == NULL) goto truncated; ts = zip->iv_size; zip->iv_size = archive_le16dec(p); __archive_read_consume(a, 2); if (ts < zip->iv_size) { free(zip->iv); zip->iv = NULL; } p = __archive_read_ahead(a, zip->iv_size, NULL); if (p == NULL) goto truncated; if (zip->iv == NULL) { zip->iv = malloc(zip->iv_size); if (zip->iv == NULL) goto nomem; } memcpy(zip->iv, p, zip->iv_size); __archive_read_consume(a, zip->iv_size); /* * Read a size of remaining decryption header field. */ p = __archive_read_ahead(a, 14, NULL); if (p == NULL) goto truncated; remaining_size = archive_le32dec(p); if (remaining_size < 16 || remaining_size > (1 << 18)) goto corrupted; /* Check if format version is supported. */ if (archive_le16dec(p+4) != 3) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Unsupported encryption format version: %u", archive_le16dec(p+4)); return (ARCHIVE_FAILED); } /* * Read an encryption algorithm field. */ zip->alg_id = archive_le16dec(p+6); switch (zip->alg_id) { case 0x6601:/* DES */ case 0x6602:/* RC2 */ case 0x6603:/* 3DES 168 */ case 0x6609:/* 3DES 112 */ case 0x660E:/* AES 128 */ case 0x660F:/* AES 192 */ case 0x6610:/* AES 256 */ case 0x6702:/* RC2 (version >= 5.2) */ case 0x6720:/* Blowfish */ case 0x6721:/* Twofish */ case 0x6801:/* RC4 */ /* Supported encryption algorithm. */ break; default: archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Unknown encryption algorithm: %u", zip->alg_id); return (ARCHIVE_FAILED); } /* * Read a bit length field. */ zip->bit_len = archive_le16dec(p+8); /* * Read a flags field. */ zip->flags = archive_le16dec(p+10); switch (zip->flags & 0xf000) { case 0x0001: /* Password is required to decrypt. */ case 0x0002: /* Certificates only. */ case 0x0003: /* Password or certificate required to decrypt. */ break; default: archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Unknown encryption flag: %u", zip->flags); return (ARCHIVE_FAILED); } if ((zip->flags & 0xf000) == 0 || (zip->flags & 0xf000) == 0x4000) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Unknown encryption flag: %u", zip->flags); return (ARCHIVE_FAILED); } /* * Read an encrypted random data field. */ ts = zip->erd_size; zip->erd_size = archive_le16dec(p+12); __archive_read_consume(a, 14); if ((zip->erd_size & 0xf) != 0 || (zip->erd_size + 16) > remaining_size || (zip->erd_size + 16) < zip->erd_size) goto corrupted; if (ts < zip->erd_size) { free(zip->erd); zip->erd = NULL; } p = __archive_read_ahead(a, zip->erd_size, NULL); if (p == NULL) goto truncated; if (zip->erd == NULL) { zip->erd = malloc(zip->erd_size); if (zip->erd == NULL) goto nomem; } memcpy(zip->erd, p, zip->erd_size); __archive_read_consume(a, zip->erd_size); /* * Read a reserved data field. */ p = __archive_read_ahead(a, 4, NULL); if (p == NULL) goto truncated; /* Reserved data size should be zero. */ if (archive_le32dec(p) != 0) goto corrupted; __archive_read_consume(a, 4); /* * Read a password validation data field. */ p = __archive_read_ahead(a, 2, NULL); if (p == NULL) goto truncated; ts = zip->v_size; zip->v_size = archive_le16dec(p); __archive_read_consume(a, 2); if ((zip->v_size & 0x0f) != 0 || (zip->erd_size + zip->v_size + 16) > remaining_size || (zip->erd_size + zip->v_size + 16) < (zip->erd_size + zip->v_size)) goto corrupted; if (ts < zip->v_size) { free(zip->v_data); zip->v_data = NULL; } p = __archive_read_ahead(a, zip->v_size, NULL); if (p == NULL) goto truncated; if (zip->v_data == NULL) { zip->v_data = malloc(zip->v_size); if (zip->v_data == NULL) goto nomem; } memcpy(zip->v_data, p, zip->v_size); __archive_read_consume(a, zip->v_size); p = __archive_read_ahead(a, 4, NULL); if (p == NULL) goto truncated; zip->v_crc32 = archive_le32dec(p); __archive_read_consume(a, 4); /*return (ARCHIVE_OK); * This is not fully implemented yet.*/ archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Encrypted file is unsupported"); return (ARCHIVE_FAILED); truncated: archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated ZIP file data"); return (ARCHIVE_FATAL); corrupted: archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Corrupted ZIP file data"); return (ARCHIVE_FATAL); nomem: archive_set_error(&a->archive, ENOMEM, "No memory for ZIP decryption"); return (ARCHIVE_FATAL); } static int zip_alloc_decryption_buffer(struct archive_read *a) { struct zip *zip = (struct zip *)(a->format->data); size_t bs = 256 * 1024; if (zip->decrypted_buffer == NULL) { zip->decrypted_buffer_size = bs; zip->decrypted_buffer = malloc(bs); if (zip->decrypted_buffer == NULL) { archive_set_error(&a->archive, ENOMEM, "No memory for ZIP decryption"); return (ARCHIVE_FATAL); } } zip->decrypted_ptr = zip->decrypted_buffer; return (ARCHIVE_OK); } static int init_traditional_PKWARE_decryption(struct archive_read *a) { struct zip *zip = (struct zip *)(a->format->data); const void *p; int retry; int r; if (zip->tctx_valid) return (ARCHIVE_OK); /* Read the 12 bytes encryption header stored at the start of the data area. */ #define ENC_HEADER_SIZE 12 if (0 == (zip->entry->zip_flags & ZIP_LENGTH_AT_END) && zip->entry_bytes_remaining < ENC_HEADER_SIZE) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated Zip encrypted body: only %jd bytes available", (intmax_t)zip->entry_bytes_remaining); return (ARCHIVE_FATAL); } p = __archive_read_ahead(a, ENC_HEADER_SIZE, NULL); if (p == NULL) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated ZIP file data"); return (ARCHIVE_FATAL); } for (retry = 0;; retry++) { const char *passphrase; uint8_t crcchk; passphrase = __archive_read_next_passphrase(a); if (passphrase == NULL) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, (retry > 0)? "Incorrect passphrase": "Passphrase required for this entry"); return (ARCHIVE_FAILED); } /* * Initialize ctx for Traditional PKWARE Decryption. */ r = trad_enc_init(&zip->tctx, passphrase, strlen(passphrase), p, ENC_HEADER_SIZE, &crcchk); if (r == 0 && crcchk == zip->entry->decdat) break;/* The passphrase is OK. */ if (retry > 10000) { /* Avoid infinity loop. */ archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Too many incorrect passphrases"); return (ARCHIVE_FAILED); } } __archive_read_consume(a, ENC_HEADER_SIZE); zip->tctx_valid = 1; if (0 == (zip->entry->zip_flags & ZIP_LENGTH_AT_END)) { zip->entry_bytes_remaining -= ENC_HEADER_SIZE; } /*zip->entry_uncompressed_bytes_read += ENC_HEADER_SIZE;*/ zip->entry_compressed_bytes_read += ENC_HEADER_SIZE; zip->decrypted_bytes_remaining = 0; return (zip_alloc_decryption_buffer(a)); #undef ENC_HEADER_SIZE } static int init_WinZip_AES_decryption(struct archive_read *a) { struct zip *zip = (struct zip *)(a->format->data); const void *p; const uint8_t *pv; size_t key_len, salt_len; uint8_t derived_key[MAX_DERIVED_KEY_BUF_SIZE]; int retry; int r; if (zip->cctx_valid || zip->hctx_valid) return (ARCHIVE_OK); switch (zip->entry->aes_extra.strength) { case 1: salt_len = 8; key_len = 16; break; case 2: salt_len = 12; key_len = 24; break; case 3: salt_len = 16; key_len = 32; break; default: goto corrupted; } p = __archive_read_ahead(a, salt_len + 2, NULL); if (p == NULL) goto truncated; for (retry = 0;; retry++) { const char *passphrase; passphrase = __archive_read_next_passphrase(a); if (passphrase == NULL) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, (retry > 0)? "Incorrect passphrase": "Passphrase required for this entry"); return (ARCHIVE_FAILED); } memset(derived_key, 0, sizeof(derived_key)); r = archive_pbkdf2_sha1(passphrase, strlen(passphrase), p, salt_len, 1000, derived_key, key_len * 2 + 2); if (r != 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Decryption is unsupported due to lack of " "crypto library"); return (ARCHIVE_FAILED); } /* Check password verification value. */ pv = ((const uint8_t *)p) + salt_len; if (derived_key[key_len * 2] == pv[0] && derived_key[key_len * 2 + 1] == pv[1]) break;/* The passphrase is OK. */ if (retry > 10000) { /* Avoid infinity loop. */ archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Too many incorrect passphrases"); return (ARCHIVE_FAILED); } } r = archive_decrypto_aes_ctr_init(&zip->cctx, derived_key, key_len); if (r != 0) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Decryption is unsupported due to lack of crypto library"); return (ARCHIVE_FAILED); } r = archive_hmac_sha1_init(&zip->hctx, derived_key + key_len, key_len); if (r != 0) { archive_decrypto_aes_ctr_release(&zip->cctx); archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Failed to initialize HMAC-SHA1"); return (ARCHIVE_FAILED); } zip->cctx_valid = zip->hctx_valid = 1; __archive_read_consume(a, salt_len + 2); zip->entry_bytes_remaining -= salt_len + 2 + AUTH_CODE_SIZE; if (0 == (zip->entry->zip_flags & ZIP_LENGTH_AT_END) && zip->entry_bytes_remaining < 0) goto corrupted; zip->entry_compressed_bytes_read += salt_len + 2 + AUTH_CODE_SIZE; zip->decrypted_bytes_remaining = 0; zip->entry->compression = zip->entry->aes_extra.compression; return (zip_alloc_decryption_buffer(a)); truncated: archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated ZIP file data"); return (ARCHIVE_FATAL); corrupted: archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Corrupted ZIP file data"); return (ARCHIVE_FATAL); } static int archive_read_format_zip_read_data(struct archive_read *a, const void **buff, size_t *size, int64_t *offset) { int r; struct zip *zip = (struct zip *)(a->format->data); if (zip->has_encrypted_entries == ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW) { zip->has_encrypted_entries = 0; } *offset = zip->entry_uncompressed_bytes_read; *size = 0; *buff = NULL; /* If we hit end-of-entry last time, return ARCHIVE_EOF. */ if (zip->end_of_entry) return (ARCHIVE_EOF); /* Return EOF immediately if this is a non-regular file. */ if (AE_IFREG != (zip->entry->mode & AE_IFMT)) return (ARCHIVE_EOF); __archive_read_consume(a, zip->unconsumed); zip->unconsumed = 0; if (zip->init_decryption) { zip->has_encrypted_entries = 1; if (zip->entry->zip_flags & ZIP_STRONG_ENCRYPTED) r = read_decryption_header(a); else if (zip->entry->compression == WINZIP_AES_ENCRYPTION) r = init_WinZip_AES_decryption(a); else r = init_traditional_PKWARE_decryption(a); if (r != ARCHIVE_OK) return (r); zip->init_decryption = 0; } switch(zip->entry->compression) { case 0: /* No compression. */ r = zip_read_data_none(a, buff, size, offset); break; #ifdef HAVE_ZLIB_H case 8: /* Deflate compression. */ r = zip_read_data_deflate(a, buff, size, offset); break; #endif default: /* Unsupported compression. */ /* Return a warning. */ archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Unsupported ZIP compression method (%s)", compression_name(zip->entry->compression)); /* We can't decompress this entry, but we will * be able to skip() it and try the next entry. */ return (ARCHIVE_FAILED); break; } if (r != ARCHIVE_OK) return (r); /* Update checksum */ if (*size) zip->entry_crc32 = zip->crc32func(zip->entry_crc32, *buff, (unsigned)*size); /* If we hit the end, swallow any end-of-data marker. */ if (zip->end_of_entry) { /* Check file size, CRC against these values. */ if (zip->entry->compressed_size != zip->entry_compressed_bytes_read) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "ZIP compressed data is wrong size " "(read %jd, expected %jd)", (intmax_t)zip->entry_compressed_bytes_read, (intmax_t)zip->entry->compressed_size); return (ARCHIVE_WARN); } /* Size field only stores the lower 32 bits of the actual * size. */ if ((zip->entry->uncompressed_size & UINT32_MAX) != (zip->entry_uncompressed_bytes_read & UINT32_MAX)) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "ZIP uncompressed data is wrong size " "(read %jd, expected %jd)\n", (intmax_t)zip->entry_uncompressed_bytes_read, (intmax_t)zip->entry->uncompressed_size); return (ARCHIVE_WARN); } /* Check computed CRC against header */ if ((!zip->hctx_valid || zip->entry->aes_extra.vendor != AES_VENDOR_AE_2) && zip->entry->crc32 != zip->entry_crc32 && !zip->ignore_crc32) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "ZIP bad CRC: 0x%lx should be 0x%lx", (unsigned long)zip->entry_crc32, (unsigned long)zip->entry->crc32); return (ARCHIVE_WARN); } } return (ARCHIVE_OK); } static int archive_read_format_zip_cleanup(struct archive_read *a) { struct zip *zip; struct zip_entry *zip_entry, *next_zip_entry; zip = (struct zip *)(a->format->data); #ifdef HAVE_ZLIB_H if (zip->stream_valid) inflateEnd(&zip->stream); free(zip->uncompressed_buffer); #endif if (zip->zip_entries) { zip_entry = zip->zip_entries; while (zip_entry != NULL) { next_zip_entry = zip_entry->next; archive_string_free(&zip_entry->rsrcname); free(zip_entry); zip_entry = next_zip_entry; } } free(zip->decrypted_buffer); if (zip->cctx_valid) archive_decrypto_aes_ctr_release(&zip->cctx); if (zip->hctx_valid) archive_hmac_sha1_cleanup(&zip->hctx); free(zip->iv); free(zip->erd); free(zip->v_data); archive_string_free(&zip->format_name); free(zip); (a->format->data) = NULL; return (ARCHIVE_OK); } static int archive_read_format_zip_has_encrypted_entries(struct archive_read *_a) { if (_a && _a->format) { struct zip * zip = (struct zip *)_a->format->data; if (zip) { return zip->has_encrypted_entries; } } return ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW; } static int archive_read_format_zip_options(struct archive_read *a, const char *key, const char *val) { struct zip *zip; int ret = ARCHIVE_FAILED; zip = (struct zip *)(a->format->data); if (strcmp(key, "compat-2x") == 0) { /* Handle filenames as libarchive 2.x */ zip->init_default_conversion = (val != NULL) ? 1 : 0; return (ARCHIVE_OK); } else if (strcmp(key, "hdrcharset") == 0) { if (val == NULL || val[0] == 0) archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "zip: hdrcharset option needs a character-set name" ); else { zip->sconv = archive_string_conversion_from_charset( &a->archive, val, 0); if (zip->sconv != NULL) { if (strcmp(val, "UTF-8") == 0) zip->sconv_utf8 = zip->sconv; ret = ARCHIVE_OK; } else ret = ARCHIVE_FATAL; } return (ret); } else if (strcmp(key, "ignorecrc32") == 0) { /* Mostly useful for testing. */ if (val == NULL || val[0] == 0) { zip->crc32func = real_crc32; zip->ignore_crc32 = 0; } else { zip->crc32func = fake_crc32; zip->ignore_crc32 = 1; } return (ARCHIVE_OK); } else if (strcmp(key, "mac-ext") == 0) { zip->process_mac_extensions = (val != NULL && val[0] != 0); return (ARCHIVE_OK); } /* Note: The "warn" return is just to inform the options * supervisor that we didn't handle it. It will generate * a suitable error if no one used this option. */ return (ARCHIVE_WARN); } int archive_read_support_format_zip(struct archive *a) { int r; r = archive_read_support_format_zip_streamable(a); if (r != ARCHIVE_OK) return r; return (archive_read_support_format_zip_seekable(a)); } /* ------------------------------------------------------------------------ */ /* * Streaming-mode support */ static int archive_read_support_format_zip_capabilities_streamable(struct archive_read * a) { (void)a; /* UNUSED */ return (ARCHIVE_READ_FORMAT_CAPS_ENCRYPT_DATA | ARCHIVE_READ_FORMAT_CAPS_ENCRYPT_METADATA); } static int archive_read_format_zip_streamable_bid(struct archive_read *a, int best_bid) { const char *p; (void)best_bid; /* UNUSED */ if ((p = __archive_read_ahead(a, 4, NULL)) == NULL) return (-1); /* * Bid of 29 here comes from: * + 16 bits for "PK", * + next 16-bit field has 6 options so contributes * about 16 - log_2(6) ~= 16 - 2.6 ~= 13 bits * * So we've effectively verified ~29 total bits of check data. */ if (p[0] == 'P' && p[1] == 'K') { if ((p[2] == '\001' && p[3] == '\002') || (p[2] == '\003' && p[3] == '\004') || (p[2] == '\005' && p[3] == '\006') || (p[2] == '\006' && p[3] == '\006') || (p[2] == '\007' && p[3] == '\010') || (p[2] == '0' && p[3] == '0')) return (29); } /* TODO: It's worth looking ahead a little bit for a valid * PK signature. In particular, that would make it possible * to read some UUEncoded SFX files or SFX files coming from * a network socket. */ return (0); } static int archive_read_format_zip_streamable_read_header(struct archive_read *a, struct archive_entry *entry) { struct zip *zip; a->archive.archive_format = ARCHIVE_FORMAT_ZIP; if (a->archive.archive_format_name == NULL) a->archive.archive_format_name = "ZIP"; zip = (struct zip *)(a->format->data); /* * It should be sufficient to call archive_read_next_header() for * a reader to determine if an entry is encrypted or not. If the * encryption of an entry is only detectable when calling * archive_read_data(), so be it. We'll do the same check there * as well. */ if (zip->has_encrypted_entries == ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW) zip->has_encrypted_entries = 0; /* Make sure we have a zip_entry structure to use. */ if (zip->zip_entries == NULL) { zip->zip_entries = malloc(sizeof(struct zip_entry)); if (zip->zip_entries == NULL) { archive_set_error(&a->archive, ENOMEM, "Out of memory"); return ARCHIVE_FATAL; } } zip->entry = zip->zip_entries; memset(zip->entry, 0, sizeof(struct zip_entry)); if (zip->cctx_valid) archive_decrypto_aes_ctr_release(&zip->cctx); if (zip->hctx_valid) archive_hmac_sha1_cleanup(&zip->hctx); zip->tctx_valid = zip->cctx_valid = zip->hctx_valid = 0; __archive_read_reset_passphrase(a); /* Search ahead for the next local file header. */ __archive_read_consume(a, zip->unconsumed); zip->unconsumed = 0; for (;;) { int64_t skipped = 0; const char *p, *end; ssize_t bytes; p = __archive_read_ahead(a, 4, &bytes); if (p == NULL) return (ARCHIVE_FATAL); end = p + bytes; while (p + 4 <= end) { if (p[0] == 'P' && p[1] == 'K') { if (p[2] == '\003' && p[3] == '\004') { /* Regular file entry. */ __archive_read_consume(a, skipped); return zip_read_local_file_header(a, entry, zip); } /* * TODO: We cannot restore permissions * based only on the local file headers. * Consider scanning the central * directory and returning additional * entries for at least directories. * This would allow us to properly set * directory permissions. * * This won't help us fix symlinks * and may not help with regular file * permissions, either. */ if (p[2] == '\001' && p[3] == '\002') { return (ARCHIVE_EOF); } /* End of central directory? Must be an * empty archive. */ if ((p[2] == '\005' && p[3] == '\006') || (p[2] == '\006' && p[3] == '\006')) return (ARCHIVE_EOF); } ++p; ++skipped; } __archive_read_consume(a, skipped); } } static int archive_read_format_zip_read_data_skip_streamable(struct archive_read *a) { struct zip *zip; int64_t bytes_skipped; zip = (struct zip *)(a->format->data); bytes_skipped = __archive_read_consume(a, zip->unconsumed); zip->unconsumed = 0; if (bytes_skipped < 0) return (ARCHIVE_FATAL); /* If we've already read to end of data, we're done. */ if (zip->end_of_entry) return (ARCHIVE_OK); /* So we know we're streaming... */ if (0 == (zip->entry->zip_flags & ZIP_LENGTH_AT_END) || zip->entry->compressed_size > 0) { /* We know the compressed length, so we can just skip. */ bytes_skipped = __archive_read_consume(a, zip->entry_bytes_remaining); if (bytes_skipped < 0) return (ARCHIVE_FATAL); return (ARCHIVE_OK); } if (zip->init_decryption) { int r; zip->has_encrypted_entries = 1; if (zip->entry->zip_flags & ZIP_STRONG_ENCRYPTED) r = read_decryption_header(a); else if (zip->entry->compression == WINZIP_AES_ENCRYPTION) r = init_WinZip_AES_decryption(a); else r = init_traditional_PKWARE_decryption(a); if (r != ARCHIVE_OK) return (r); zip->init_decryption = 0; } /* We're streaming and we don't know the length. */ /* If the body is compressed and we know the format, we can * find an exact end-of-entry by decompressing it. */ switch (zip->entry->compression) { #ifdef HAVE_ZLIB_H case 8: /* Deflate compression. */ while (!zip->end_of_entry) { int64_t offset = 0; const void *buff = NULL; size_t size = 0; int r; r = zip_read_data_deflate(a, &buff, &size, &offset); if (r != ARCHIVE_OK) return (r); } return ARCHIVE_OK; #endif default: /* Uncompressed or unknown. */ /* Scan for a PK\007\010 signature. */ for (;;) { const char *p, *buff; ssize_t bytes_avail; buff = __archive_read_ahead(a, 16, &bytes_avail); if (bytes_avail < 16) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated ZIP file data"); return (ARCHIVE_FATAL); } p = buff; while (p <= buff + bytes_avail - 16) { if (p[3] == 'P') { p += 3; } else if (p[3] == 'K') { p += 2; } else if (p[3] == '\007') { p += 1; } else if (p[3] == '\010' && p[2] == '\007' && p[1] == 'K' && p[0] == 'P') { if (zip->entry->flags & LA_USED_ZIP64) __archive_read_consume(a, p - buff + 24); else __archive_read_consume(a, p - buff + 16); return ARCHIVE_OK; } else { p += 4; } } __archive_read_consume(a, p - buff); } } } int archive_read_support_format_zip_streamable(struct archive *_a) { struct archive_read *a = (struct archive_read *)_a; struct zip *zip; int r; archive_check_magic(_a, ARCHIVE_READ_MAGIC, ARCHIVE_STATE_NEW, "archive_read_support_format_zip"); zip = (struct zip *)calloc(1, sizeof(*zip)); if (zip == NULL) { archive_set_error(&a->archive, ENOMEM, "Can't allocate zip data"); return (ARCHIVE_FATAL); } /* Streamable reader doesn't support mac extensions. */ zip->process_mac_extensions = 0; /* * Until enough data has been read, we cannot tell about * any encrypted entries yet. */ zip->has_encrypted_entries = ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW; zip->crc32func = real_crc32; r = __archive_read_register_format(a, zip, "zip", archive_read_format_zip_streamable_bid, archive_read_format_zip_options, archive_read_format_zip_streamable_read_header, archive_read_format_zip_read_data, archive_read_format_zip_read_data_skip_streamable, NULL, archive_read_format_zip_cleanup, archive_read_support_format_zip_capabilities_streamable, archive_read_format_zip_has_encrypted_entries); if (r != ARCHIVE_OK) free(zip); return (ARCHIVE_OK); } /* ------------------------------------------------------------------------ */ /* * Seeking-mode support */ static int archive_read_support_format_zip_capabilities_seekable(struct archive_read * a) { (void)a; /* UNUSED */ return (ARCHIVE_READ_FORMAT_CAPS_ENCRYPT_DATA | ARCHIVE_READ_FORMAT_CAPS_ENCRYPT_METADATA); } /* * TODO: This is a performance sink because it forces the read core to * drop buffered data from the start of file, which will then have to * be re-read again if this bidder loses. * * We workaround this a little by passing in the best bid so far so * that later bidders can do nothing if they know they'll never * outbid. But we can certainly do better... */ static int read_eocd(struct zip *zip, const char *p, int64_t current_offset) { /* Sanity-check the EOCD we've found. */ /* This must be the first volume. */ if (archive_le16dec(p + 4) != 0) return 0; /* Central directory must be on this volume. */ if (archive_le16dec(p + 4) != archive_le16dec(p + 6)) return 0; /* All central directory entries must be on this volume. */ if (archive_le16dec(p + 10) != archive_le16dec(p + 8)) return 0; /* Central directory can't extend beyond start of EOCD record. */ if (archive_le32dec(p + 16) + archive_le32dec(p + 12) > current_offset) return 0; /* Save the central directory location for later use. */ zip->central_directory_offset = archive_le32dec(p + 16); /* This is just a tiny bit higher than the maximum returned by the streaming Zip bidder. This ensures that the more accurate seeking Zip parser wins whenever seek is available. */ return 32; } /* * Examine Zip64 EOCD locator: If it's valid, store the information * from it. */ static int read_zip64_eocd(struct archive_read *a, struct zip *zip, const char *p) { int64_t eocd64_offset; int64_t eocd64_size; /* Sanity-check the locator record. */ /* Central dir must be on first volume. */ if (archive_le32dec(p + 4) != 0) return 0; /* Must be only a single volume. */ if (archive_le32dec(p + 16) != 1) return 0; /* Find the Zip64 EOCD record. */ eocd64_offset = archive_le64dec(p + 8); if (__archive_read_seek(a, eocd64_offset, SEEK_SET) < 0) return 0; if ((p = __archive_read_ahead(a, 56, NULL)) == NULL) return 0; /* Make sure we can read all of it. */ eocd64_size = archive_le64dec(p + 4) + 12; if (eocd64_size < 56 || eocd64_size > 16384) return 0; if ((p = __archive_read_ahead(a, (size_t)eocd64_size, NULL)) == NULL) return 0; /* Sanity-check the EOCD64 */ if (archive_le32dec(p + 16) != 0) /* Must be disk #0 */ return 0; if (archive_le32dec(p + 20) != 0) /* CD must be on disk #0 */ return 0; /* CD can't be split. */ if (archive_le64dec(p + 24) != archive_le64dec(p + 32)) return 0; /* Save the central directory offset for later use. */ zip->central_directory_offset = archive_le64dec(p + 48); return 32; } static int archive_read_format_zip_seekable_bid(struct archive_read *a, int best_bid) { struct zip *zip = (struct zip *)a->format->data; int64_t file_size, current_offset; const char *p; int i, tail; /* If someone has already bid more than 32, then avoid trashing the look-ahead buffers with a seek. */ if (best_bid > 32) return (-1); file_size = __archive_read_seek(a, 0, SEEK_END); if (file_size <= 0) return 0; /* Search last 16k of file for end-of-central-directory * record (which starts with PK\005\006) */ tail = (int)zipmin(1024 * 16, file_size); current_offset = __archive_read_seek(a, -tail, SEEK_END); if (current_offset < 0) return 0; if ((p = __archive_read_ahead(a, (size_t)tail, NULL)) == NULL) return 0; /* Boyer-Moore search backwards from the end, since we want * to match the last EOCD in the file (there can be more than * one if there is an uncompressed Zip archive as a member * within this Zip archive). */ for (i = tail - 22; i > 0;) { switch (p[i]) { case 'P': if (memcmp(p + i, "PK\005\006", 4) == 0) { int ret = read_eocd(zip, p + i, current_offset + i); /* Zip64 EOCD locator precedes * regular EOCD if present. */ if (i >= 20 && memcmp(p + i - 20, "PK\006\007", 4) == 0) { int ret_zip64 = read_zip64_eocd(a, zip, p + i - 20); if (ret_zip64 > ret) ret = ret_zip64; } return (ret); } i -= 4; break; case 'K': i -= 1; break; case 005: i -= 2; break; case 006: i -= 3; break; default: i -= 4; break; } } return 0; } /* The red-black trees are only used in seeking mode to manage * the in-memory copy of the central directory. */ static int cmp_node(const struct archive_rb_node *n1, const struct archive_rb_node *n2) { const struct zip_entry *e1 = (const struct zip_entry *)n1; const struct zip_entry *e2 = (const struct zip_entry *)n2; if (e1->local_header_offset > e2->local_header_offset) return -1; if (e1->local_header_offset < e2->local_header_offset) return 1; return 0; } static int cmp_key(const struct archive_rb_node *n, const void *key) { /* This function won't be called */ (void)n; /* UNUSED */ (void)key; /* UNUSED */ return 1; } static const struct archive_rb_tree_ops rb_ops = { &cmp_node, &cmp_key }; static int rsrc_cmp_node(const struct archive_rb_node *n1, const struct archive_rb_node *n2) { const struct zip_entry *e1 = (const struct zip_entry *)n1; const struct zip_entry *e2 = (const struct zip_entry *)n2; return (strcmp(e2->rsrcname.s, e1->rsrcname.s)); } static int rsrc_cmp_key(const struct archive_rb_node *n, const void *key) { const struct zip_entry *e = (const struct zip_entry *)n; return (strcmp((const char *)key, e->rsrcname.s)); } static const struct archive_rb_tree_ops rb_rsrc_ops = { &rsrc_cmp_node, &rsrc_cmp_key }; static const char * rsrc_basename(const char *name, size_t name_length) { const char *s, *r; r = s = name; for (;;) { s = memchr(s, '/', name_length - (s - name)); if (s == NULL) break; r = ++s; } return (r); } static void expose_parent_dirs(struct zip *zip, const char *name, size_t name_length) { struct archive_string str; struct zip_entry *dir; char *s; archive_string_init(&str); archive_strncpy(&str, name, name_length); for (;;) { s = strrchr(str.s, '/'); if (s == NULL) break; *s = '\0'; /* Transfer the parent directory from zip->tree_rsrc RB * tree to zip->tree RB tree to expose. */ dir = (struct zip_entry *) __archive_rb_tree_find_node(&zip->tree_rsrc, str.s); if (dir == NULL) break; __archive_rb_tree_remove_node(&zip->tree_rsrc, &dir->node); archive_string_free(&dir->rsrcname); __archive_rb_tree_insert_node(&zip->tree, &dir->node); } archive_string_free(&str); } static int slurp_central_directory(struct archive_read *a, struct zip *zip) { ssize_t i; unsigned found; int64_t correction; ssize_t bytes_avail; const char *p; /* * Find the start of the central directory. The end-of-CD * record has our starting point, but there are lots of * Zip archives which have had other data prepended to the * file, which makes the recorded offsets all too small. * So we search forward from the specified offset until we * find the real start of the central directory. Then we * know the correction we need to apply to account for leading * padding. */ if (__archive_read_seek(a, zip->central_directory_offset, SEEK_SET) < 0) return ARCHIVE_FATAL; found = 0; while (!found) { if ((p = __archive_read_ahead(a, 20, &bytes_avail)) == NULL) return ARCHIVE_FATAL; for (found = 0, i = 0; !found && i < bytes_avail - 4;) { switch (p[i + 3]) { case 'P': i += 3; break; case 'K': i += 2; break; case 001: i += 1; break; case 002: if (memcmp(p + i, "PK\001\002", 4) == 0) { p += i; found = 1; } else i += 4; break; case 005: i += 1; break; case 006: if (memcmp(p + i, "PK\005\006", 4) == 0) { p += i; found = 1; } else if (memcmp(p + i, "PK\006\006", 4) == 0) { p += i; found = 1; } else i += 1; break; default: i += 4; break; } } __archive_read_consume(a, i); } correction = archive_filter_bytes(&a->archive, 0) - zip->central_directory_offset; __archive_rb_tree_init(&zip->tree, &rb_ops); __archive_rb_tree_init(&zip->tree_rsrc, &rb_rsrc_ops); zip->central_directory_entries_total = 0; while (1) { struct zip_entry *zip_entry; size_t filename_length, extra_length, comment_length; uint32_t external_attributes; const char *name, *r; if ((p = __archive_read_ahead(a, 4, NULL)) == NULL) return ARCHIVE_FATAL; if (memcmp(p, "PK\006\006", 4) == 0 || memcmp(p, "PK\005\006", 4) == 0) { break; } else if (memcmp(p, "PK\001\002", 4) != 0) { archive_set_error(&a->archive, -1, "Invalid central directory signature"); return ARCHIVE_FATAL; } if ((p = __archive_read_ahead(a, 46, NULL)) == NULL) return ARCHIVE_FATAL; zip_entry = calloc(1, sizeof(struct zip_entry)); zip_entry->next = zip->zip_entries; zip_entry->flags |= LA_FROM_CENTRAL_DIRECTORY; zip->zip_entries = zip_entry; zip->central_directory_entries_total++; /* version = p[4]; */ zip_entry->system = p[5]; /* version_required = archive_le16dec(p + 6); */ zip_entry->zip_flags = archive_le16dec(p + 8); if (zip_entry->zip_flags & (ZIP_ENCRYPTED | ZIP_STRONG_ENCRYPTED)){ zip->has_encrypted_entries = 1; } zip_entry->compression = (char)archive_le16dec(p + 10); zip_entry->mtime = zip_time(p + 12); zip_entry->crc32 = archive_le32dec(p + 16); if (zip_entry->zip_flags & ZIP_LENGTH_AT_END) zip_entry->decdat = p[13]; else zip_entry->decdat = p[19]; zip_entry->compressed_size = archive_le32dec(p + 20); zip_entry->uncompressed_size = archive_le32dec(p + 24); filename_length = archive_le16dec(p + 28); extra_length = archive_le16dec(p + 30); comment_length = archive_le16dec(p + 32); /* disk_start = archive_le16dec(p + 34); */ /* Better be zero. */ /* internal_attributes = archive_le16dec(p + 36); */ /* text bit */ external_attributes = archive_le32dec(p + 38); zip_entry->local_header_offset = archive_le32dec(p + 42) + correction; /* If we can't guess the mode, leave it zero here; when we read the local file header we might get more information. */ if (zip_entry->system == 3) { zip_entry->mode = external_attributes >> 16; } else if (zip_entry->system == 0) { // Interpret MSDOS directory bit if (0x10 == (external_attributes & 0x10)) { zip_entry->mode = AE_IFDIR | 0775; } else { zip_entry->mode = AE_IFREG | 0664; } if (0x01 == (external_attributes & 0x01)) { // Read-only bit; strip write permissions zip_entry->mode &= 0555; } } else { zip_entry->mode = 0; } /* We're done with the regular data; get the filename and * extra data. */ __archive_read_consume(a, 46); p = __archive_read_ahead(a, filename_length + extra_length, NULL); if (p == NULL) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated ZIP file header"); return ARCHIVE_FATAL; } if (ARCHIVE_OK != process_extra(a, p + filename_length, extra_length, zip_entry)) { return ARCHIVE_FATAL; } /* * Mac resource fork files are stored under the * "__MACOSX/" directory, so we should check if * it is. */ if (!zip->process_mac_extensions) { /* Treat every entry as a regular entry. */ __archive_rb_tree_insert_node(&zip->tree, &zip_entry->node); } else { name = p; r = rsrc_basename(name, filename_length); if (filename_length >= 9 && strncmp("__MACOSX/", name, 9) == 0) { /* If this file is not a resource fork nor * a directory. We should treat it as a non * resource fork file to expose it. */ if (name[filename_length-1] != '/' && (r - name < 3 || r[0] != '.' || r[1] != '_')) { __archive_rb_tree_insert_node( &zip->tree, &zip_entry->node); /* Expose its parent directories. */ expose_parent_dirs(zip, name, filename_length); } else { /* This file is a resource fork file or * a directory. */ archive_strncpy(&(zip_entry->rsrcname), name, filename_length); __archive_rb_tree_insert_node( &zip->tree_rsrc, &zip_entry->node); } } else { /* Generate resource fork name to find its * resource file at zip->tree_rsrc. */ archive_strcpy(&(zip_entry->rsrcname), "__MACOSX/"); archive_strncat(&(zip_entry->rsrcname), name, r - name); archive_strcat(&(zip_entry->rsrcname), "._"); archive_strncat(&(zip_entry->rsrcname), name + (r - name), filename_length - (r - name)); /* Register an entry to RB tree to sort it by * file offset. */ __archive_rb_tree_insert_node(&zip->tree, &zip_entry->node); } } /* Skip the comment too ... */ __archive_read_consume(a, filename_length + extra_length + comment_length); } return ARCHIVE_OK; } static ssize_t zip_get_local_file_header_size(struct archive_read *a, size_t extra) { const char *p; ssize_t filename_length, extra_length; if ((p = __archive_read_ahead(a, extra + 30, NULL)) == NULL) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated ZIP file header"); return (ARCHIVE_WARN); } p += extra; if (memcmp(p, "PK\003\004", 4) != 0) { archive_set_error(&a->archive, -1, "Damaged Zip archive"); return ARCHIVE_WARN; } filename_length = archive_le16dec(p + 26); extra_length = archive_le16dec(p + 28); return (30 + filename_length + extra_length); } static int zip_read_mac_metadata(struct archive_read *a, struct archive_entry *entry, struct zip_entry *rsrc) { struct zip *zip = (struct zip *)a->format->data; unsigned char *metadata, *mp; int64_t offset = archive_filter_bytes(&a->archive, 0); size_t remaining_bytes, metadata_bytes; ssize_t hsize; int ret = ARCHIVE_OK, eof; switch(rsrc->compression) { case 0: /* No compression. */ if (rsrc->uncompressed_size != rsrc->compressed_size) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Malformed OS X metadata entry: inconsistent size"); return (ARCHIVE_FATAL); } #ifdef HAVE_ZLIB_H case 8: /* Deflate compression. */ #endif break; default: /* Unsupported compression. */ /* Return a warning. */ archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Unsupported ZIP compression method (%s)", compression_name(rsrc->compression)); /* We can't decompress this entry, but we will * be able to skip() it and try the next entry. */ return (ARCHIVE_WARN); } if (rsrc->uncompressed_size > (4 * 1024 * 1024)) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Mac metadata is too large: %jd > 4M bytes", (intmax_t)rsrc->uncompressed_size); return (ARCHIVE_WARN); } if (rsrc->compressed_size > (4 * 1024 * 1024)) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Mac metadata is too large: %jd > 4M bytes", (intmax_t)rsrc->compressed_size); return (ARCHIVE_WARN); } metadata = malloc((size_t)rsrc->uncompressed_size); if (metadata == NULL) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory for Mac metadata"); return (ARCHIVE_FATAL); } if (offset < rsrc->local_header_offset) __archive_read_consume(a, rsrc->local_header_offset - offset); else if (offset != rsrc->local_header_offset) { __archive_read_seek(a, rsrc->local_header_offset, SEEK_SET); } hsize = zip_get_local_file_header_size(a, 0); __archive_read_consume(a, hsize); remaining_bytes = (size_t)rsrc->compressed_size; metadata_bytes = (size_t)rsrc->uncompressed_size; mp = metadata; eof = 0; while (!eof && remaining_bytes) { const unsigned char *p; ssize_t bytes_avail; size_t bytes_used; p = __archive_read_ahead(a, 1, &bytes_avail); if (p == NULL) { archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Truncated ZIP file header"); ret = ARCHIVE_WARN; goto exit_mac_metadata; } if ((size_t)bytes_avail > remaining_bytes) bytes_avail = remaining_bytes; switch(rsrc->compression) { case 0: /* No compression. */ if ((size_t)bytes_avail > metadata_bytes) bytes_avail = metadata_bytes; memcpy(mp, p, bytes_avail); bytes_used = (size_t)bytes_avail; metadata_bytes -= bytes_used; mp += bytes_used; if (metadata_bytes == 0) eof = 1; break; #ifdef HAVE_ZLIB_H case 8: /* Deflate compression. */ { int r; ret = zip_deflate_init(a, zip); if (ret != ARCHIVE_OK) goto exit_mac_metadata; zip->stream.next_in = (Bytef *)(uintptr_t)(const void *)p; zip->stream.avail_in = (uInt)bytes_avail; zip->stream.total_in = 0; zip->stream.next_out = mp; zip->stream.avail_out = (uInt)metadata_bytes; zip->stream.total_out = 0; r = inflate(&zip->stream, 0); switch (r) { case Z_OK: break; case Z_STREAM_END: eof = 1; break; case Z_MEM_ERROR: archive_set_error(&a->archive, ENOMEM, "Out of memory for ZIP decompression"); ret = ARCHIVE_FATAL; goto exit_mac_metadata; default: archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "ZIP decompression failed (%d)", r); ret = ARCHIVE_FATAL; goto exit_mac_metadata; } bytes_used = zip->stream.total_in; metadata_bytes -= zip->stream.total_out; mp += zip->stream.total_out; break; } #endif default: bytes_used = 0; break; } __archive_read_consume(a, bytes_used); remaining_bytes -= bytes_used; } archive_entry_copy_mac_metadata(entry, metadata, (size_t)rsrc->uncompressed_size - metadata_bytes); exit_mac_metadata: __archive_read_seek(a, offset, SEEK_SET); zip->decompress_init = 0; free(metadata); return (ret); } static int archive_read_format_zip_seekable_read_header(struct archive_read *a, struct archive_entry *entry) { struct zip *zip = (struct zip *)a->format->data; struct zip_entry *rsrc; int64_t offset; int r, ret = ARCHIVE_OK; /* * It should be sufficient to call archive_read_next_header() for * a reader to determine if an entry is encrypted or not. If the * encryption of an entry is only detectable when calling * archive_read_data(), so be it. We'll do the same check there * as well. */ if (zip->has_encrypted_entries == ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW) zip->has_encrypted_entries = 0; a->archive.archive_format = ARCHIVE_FORMAT_ZIP; if (a->archive.archive_format_name == NULL) a->archive.archive_format_name = "ZIP"; if (zip->zip_entries == NULL) { r = slurp_central_directory(a, zip); if (r != ARCHIVE_OK) return r; /* Get first entry whose local header offset is lower than * other entries in the archive file. */ zip->entry = (struct zip_entry *)ARCHIVE_RB_TREE_MIN(&zip->tree); } else if (zip->entry != NULL) { /* Get next entry in local header offset order. */ zip->entry = (struct zip_entry *)__archive_rb_tree_iterate( &zip->tree, &zip->entry->node, ARCHIVE_RB_DIR_RIGHT); } if (zip->entry == NULL) return ARCHIVE_EOF; if (zip->entry->rsrcname.s) rsrc = (struct zip_entry *)__archive_rb_tree_find_node( &zip->tree_rsrc, zip->entry->rsrcname.s); else rsrc = NULL; if (zip->cctx_valid) archive_decrypto_aes_ctr_release(&zip->cctx); if (zip->hctx_valid) archive_hmac_sha1_cleanup(&zip->hctx); zip->tctx_valid = zip->cctx_valid = zip->hctx_valid = 0; __archive_read_reset_passphrase(a); /* File entries are sorted by the header offset, we should mostly * use __archive_read_consume to advance a read point to avoid redundant * data reading. */ offset = archive_filter_bytes(&a->archive, 0); if (offset < zip->entry->local_header_offset) __archive_read_consume(a, zip->entry->local_header_offset - offset); else if (offset != zip->entry->local_header_offset) { __archive_read_seek(a, zip->entry->local_header_offset, SEEK_SET); } zip->unconsumed = 0; r = zip_read_local_file_header(a, entry, zip); if (r != ARCHIVE_OK) return r; if (rsrc) { int ret2 = zip_read_mac_metadata(a, entry, rsrc); if (ret2 < ret) ret = ret2; } return (ret); } /* * We're going to seek for the next header anyway, so we don't * need to bother doing anything here. */ static int archive_read_format_zip_read_data_skip_seekable(struct archive_read *a) { struct zip *zip; zip = (struct zip *)(a->format->data); zip->unconsumed = 0; return (ARCHIVE_OK); } int archive_read_support_format_zip_seekable(struct archive *_a) { struct archive_read *a = (struct archive_read *)_a; struct zip *zip; int r; archive_check_magic(_a, ARCHIVE_READ_MAGIC, ARCHIVE_STATE_NEW, "archive_read_support_format_zip_seekable"); zip = (struct zip *)calloc(1, sizeof(*zip)); if (zip == NULL) { archive_set_error(&a->archive, ENOMEM, "Can't allocate zip data"); return (ARCHIVE_FATAL); } #ifdef HAVE_COPYFILE_H /* Set this by default on Mac OS. */ zip->process_mac_extensions = 1; #endif /* * Until enough data has been read, we cannot tell about * any encrypted entries yet. */ zip->has_encrypted_entries = ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW; zip->crc32func = real_crc32; r = __archive_read_register_format(a, zip, "zip", archive_read_format_zip_seekable_bid, archive_read_format_zip_options, archive_read_format_zip_seekable_read_header, archive_read_format_zip_read_data, archive_read_format_zip_read_data_skip_seekable, NULL, archive_read_format_zip_cleanup, archive_read_support_format_zip_capabilities_seekable, archive_read_format_zip_has_encrypted_entries); if (r != ARCHIVE_OK) free(zip); return (ARCHIVE_OK); } Index: stable/10/contrib/libarchive/libarchive/archive_string.c =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_string.c (revision 324417) +++ stable/10/contrib/libarchive/libarchive/archive_string.c (revision 324418) @@ -1,4207 +1,4208 @@ /*- * Copyright (c) 2003-2011 Tim Kientzle * Copyright (c) 2011-2012 Michihiro NAKAJIMA * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "archive_platform.h" __FBSDID("$FreeBSD$"); /* * Basic resizable string support, to simplify manipulating arbitrary-sized * strings while minimizing heap activity. * * In particular, the buffer used by a string object is only grown, it * never shrinks, so you can clear and reuse the same string object * without incurring additional memory allocations. */ #ifdef HAVE_ERRNO_H #include #endif #ifdef HAVE_ICONV_H #include #endif #ifdef HAVE_LANGINFO_H #include #endif #ifdef HAVE_LOCALCHARSET_H #include #endif #ifdef HAVE_STDLIB_H #include #endif #ifdef HAVE_STRING_H #include #endif #ifdef HAVE_WCHAR_H #include #endif #if defined(_WIN32) && !defined(__CYGWIN__) #include #include #endif #include "archive_endian.h" #include "archive_private.h" #include "archive_string.h" #include "archive_string_composition.h" #if !defined(HAVE_WMEMCPY) && !defined(wmemcpy) #define wmemcpy(a,b,i) (wchar_t *)memcpy((a), (b), (i) * sizeof(wchar_t)) #endif #if !defined(HAVE_WMEMMOVE) && !defined(wmemmove) #define wmemmove(a,b,i) (wchar_t *)memmove((a), (b), (i) * sizeof(wchar_t)) #endif struct archive_string_conv { struct archive_string_conv *next; char *from_charset; char *to_charset; unsigned from_cp; unsigned to_cp; /* Set 1 if from_charset and to_charset are the same. */ int same; int flag; #define SCONV_TO_CHARSET 1 /* MBS is being converted to specified * charset. */ #define SCONV_FROM_CHARSET (1<<1) /* MBS is being converted from * specified charset. */ #define SCONV_BEST_EFFORT (1<<2) /* Copy at least ASCII code. */ #define SCONV_WIN_CP (1<<3) /* Use Windows API for converting * MBS. */ #define SCONV_UTF8_LIBARCHIVE_2 (1<<4) /* Incorrect UTF-8 made by libarchive * 2.x in the wrong assumption. */ #define SCONV_NORMALIZATION_C (1<<6) /* Need normalization to be Form C. * Before UTF-8 characters are actually * processed. */ #define SCONV_NORMALIZATION_D (1<<7) /* Need normalization to be Form D. * Before UTF-8 characters are actually * processed. * Currently this only for MAC OS X. */ #define SCONV_TO_UTF8 (1<<8) /* "to charset" side is UTF-8. */ #define SCONV_FROM_UTF8 (1<<9) /* "from charset" side is UTF-8. */ #define SCONV_TO_UTF16BE (1<<10) /* "to charset" side is UTF-16BE. */ #define SCONV_FROM_UTF16BE (1<<11) /* "from charset" side is UTF-16BE. */ #define SCONV_TO_UTF16LE (1<<12) /* "to charset" side is UTF-16LE. */ #define SCONV_FROM_UTF16LE (1<<13) /* "from charset" side is UTF-16LE. */ #define SCONV_TO_UTF16 (SCONV_TO_UTF16BE | SCONV_TO_UTF16LE) #define SCONV_FROM_UTF16 (SCONV_FROM_UTF16BE | SCONV_FROM_UTF16LE) #if HAVE_ICONV iconv_t cd; iconv_t cd_w;/* Use at archive_mstring on * Windows. */ #endif /* A temporary buffer for normalization. */ struct archive_string utftmp; int (*converter[2])(struct archive_string *, const void *, size_t, struct archive_string_conv *); int nconverter; }; #define CP_C_LOCALE 0 /* "C" locale only for this file. */ #define CP_UTF16LE 1200 #define CP_UTF16BE 1201 #define IS_HIGH_SURROGATE_LA(uc) ((uc) >= 0xD800 && (uc) <= 0xDBFF) #define IS_LOW_SURROGATE_LA(uc) ((uc) >= 0xDC00 && (uc) <= 0xDFFF) #define IS_SURROGATE_PAIR_LA(uc) ((uc) >= 0xD800 && (uc) <= 0xDFFF) #define UNICODE_MAX 0x10FFFF #define UNICODE_R_CHAR 0xFFFD /* Replacement character. */ /* Set U+FFFD(Replacement character) in UTF-8. */ static const char utf8_replacement_char[] = {0xef, 0xbf, 0xbd}; static struct archive_string_conv *find_sconv_object(struct archive *, const char *, const char *); static void add_sconv_object(struct archive *, struct archive_string_conv *); static struct archive_string_conv *create_sconv_object(const char *, const char *, unsigned, int); static void free_sconv_object(struct archive_string_conv *); static struct archive_string_conv *get_sconv_object(struct archive *, const char *, const char *, int); static unsigned make_codepage_from_charset(const char *); static unsigned get_current_codepage(void); static unsigned get_current_oemcp(void); static size_t mbsnbytes(const void *, size_t); static size_t utf16nbytes(const void *, size_t); #if defined(_WIN32) && !defined(__CYGWIN__) static int archive_wstring_append_from_mbs_in_codepage( struct archive_wstring *, const char *, size_t, struct archive_string_conv *); static int archive_string_append_from_wcs_in_codepage(struct archive_string *, const wchar_t *, size_t, struct archive_string_conv *); static int is_big_endian(void); static int strncat_in_codepage(struct archive_string *, const void *, size_t, struct archive_string_conv *); static int win_strncat_from_utf16be(struct archive_string *, const void *, size_t, struct archive_string_conv *); static int win_strncat_from_utf16le(struct archive_string *, const void *, size_t, struct archive_string_conv *); static int win_strncat_to_utf16be(struct archive_string *, const void *, size_t, struct archive_string_conv *); static int win_strncat_to_utf16le(struct archive_string *, const void *, size_t, struct archive_string_conv *); #endif static int best_effort_strncat_from_utf16be(struct archive_string *, const void *, size_t, struct archive_string_conv *); static int best_effort_strncat_from_utf16le(struct archive_string *, const void *, size_t, struct archive_string_conv *); static int best_effort_strncat_to_utf16be(struct archive_string *, const void *, size_t, struct archive_string_conv *); static int best_effort_strncat_to_utf16le(struct archive_string *, const void *, size_t, struct archive_string_conv *); #if defined(HAVE_ICONV) static int iconv_strncat_in_locale(struct archive_string *, const void *, size_t, struct archive_string_conv *); #endif static int best_effort_strncat_in_locale(struct archive_string *, const void *, size_t, struct archive_string_conv *); static int _utf8_to_unicode(uint32_t *, const char *, size_t); static int utf8_to_unicode(uint32_t *, const char *, size_t); static inline uint32_t combine_surrogate_pair(uint32_t, uint32_t); static int cesu8_to_unicode(uint32_t *, const char *, size_t); static size_t unicode_to_utf8(char *, size_t, uint32_t); static int utf16_to_unicode(uint32_t *, const char *, size_t, int); static size_t unicode_to_utf16be(char *, size_t, uint32_t); static size_t unicode_to_utf16le(char *, size_t, uint32_t); static int strncat_from_utf8_libarchive2(struct archive_string *, const void *, size_t, struct archive_string_conv *); static int strncat_from_utf8_to_utf8(struct archive_string *, const void *, size_t, struct archive_string_conv *); static int archive_string_normalize_C(struct archive_string *, const void *, size_t, struct archive_string_conv *); static int archive_string_normalize_D(struct archive_string *, const void *, size_t, struct archive_string_conv *); static int archive_string_append_unicode(struct archive_string *, const void *, size_t, struct archive_string_conv *); static struct archive_string * archive_string_append(struct archive_string *as, const char *p, size_t s) { if (archive_string_ensure(as, as->length + s + 1) == NULL) return (NULL); if (s) memmove(as->s + as->length, p, s); as->length += s; as->s[as->length] = 0; return (as); } static struct archive_wstring * archive_wstring_append(struct archive_wstring *as, const wchar_t *p, size_t s) { if (archive_wstring_ensure(as, as->length + s + 1) == NULL) return (NULL); - wmemmove(as->s + as->length, p, s); + if (s) + wmemmove(as->s + as->length, p, s); as->length += s; as->s[as->length] = 0; return (as); } struct archive_string * archive_array_append(struct archive_string *as, const char *p, size_t s) { return archive_string_append(as, p, s); } void archive_string_concat(struct archive_string *dest, struct archive_string *src) { if (archive_string_append(dest, src->s, src->length) == NULL) __archive_errx(1, "Out of memory"); } void archive_wstring_concat(struct archive_wstring *dest, struct archive_wstring *src) { if (archive_wstring_append(dest, src->s, src->length) == NULL) __archive_errx(1, "Out of memory"); } void archive_string_free(struct archive_string *as) { as->length = 0; as->buffer_length = 0; free(as->s); as->s = NULL; } void archive_wstring_free(struct archive_wstring *as) { as->length = 0; as->buffer_length = 0; free(as->s); as->s = NULL; } struct archive_wstring * archive_wstring_ensure(struct archive_wstring *as, size_t s) { return (struct archive_wstring *) archive_string_ensure((struct archive_string *)as, s * sizeof(wchar_t)); } /* Returns NULL on any allocation failure. */ struct archive_string * archive_string_ensure(struct archive_string *as, size_t s) { char *p; size_t new_length; /* If buffer is already big enough, don't reallocate. */ if (as->s && (s <= as->buffer_length)) return (as); /* * Growing the buffer at least exponentially ensures that * append operations are always linear in the number of * characters appended. Using a smaller growth rate for * larger buffers reduces memory waste somewhat at the cost of * a larger constant factor. */ if (as->buffer_length < 32) /* Start with a minimum 32-character buffer. */ new_length = 32; else if (as->buffer_length < 8192) /* Buffers under 8k are doubled for speed. */ new_length = as->buffer_length + as->buffer_length; else { /* Buffers 8k and over grow by at least 25% each time. */ new_length = as->buffer_length + as->buffer_length / 4; /* Be safe: If size wraps, fail. */ if (new_length < as->buffer_length) { /* On failure, wipe the string and return NULL. */ archive_string_free(as); errno = ENOMEM;/* Make sure errno has ENOMEM. */ return (NULL); } } /* * The computation above is a lower limit to how much we'll * grow the buffer. In any case, we have to grow it enough to * hold the request. */ if (new_length < s) new_length = s; /* Now we can reallocate the buffer. */ p = (char *)realloc(as->s, new_length); if (p == NULL) { /* On failure, wipe the string and return NULL. */ archive_string_free(as); errno = ENOMEM;/* Make sure errno has ENOMEM. */ return (NULL); } as->s = p; as->buffer_length = new_length; return (as); } /* * TODO: See if there's a way to avoid scanning * the source string twice. Then test to see * if it actually helps (remember that we're almost * always called with pretty short arguments, so * such an optimization might not help). */ struct archive_string * archive_strncat(struct archive_string *as, const void *_p, size_t n) { size_t s; const char *p, *pp; p = (const char *)_p; /* Like strlen(p), except won't examine positions beyond p[n]. */ s = 0; pp = p; while (s < n && *pp) { pp++; s++; } if ((as = archive_string_append(as, p, s)) == NULL) __archive_errx(1, "Out of memory"); return (as); } struct archive_wstring * archive_wstrncat(struct archive_wstring *as, const wchar_t *p, size_t n) { size_t s; const wchar_t *pp; /* Like strlen(p), except won't examine positions beyond p[n]. */ s = 0; pp = p; while (s < n && *pp) { pp++; s++; } if ((as = archive_wstring_append(as, p, s)) == NULL) __archive_errx(1, "Out of memory"); return (as); } struct archive_string * archive_strcat(struct archive_string *as, const void *p) { /* strcat is just strncat without an effective limit. * Assert that we'll never get called with a source * string over 16MB. * TODO: Review all uses of strcat in the source * and try to replace them with strncat(). */ return archive_strncat(as, p, 0x1000000); } struct archive_wstring * archive_wstrcat(struct archive_wstring *as, const wchar_t *p) { /* Ditto. */ return archive_wstrncat(as, p, 0x1000000); } struct archive_string * archive_strappend_char(struct archive_string *as, char c) { if ((as = archive_string_append(as, &c, 1)) == NULL) __archive_errx(1, "Out of memory"); return (as); } struct archive_wstring * archive_wstrappend_wchar(struct archive_wstring *as, wchar_t c) { if ((as = archive_wstring_append(as, &c, 1)) == NULL) __archive_errx(1, "Out of memory"); return (as); } /* * Get the "current character set" name to use with iconv. * On FreeBSD, the empty character set name "" chooses * the correct character encoding for the current locale, * so this isn't necessary. * But iconv on Mac OS 10.6 doesn't seem to handle this correctly; * on that system, we have to explicitly call nl_langinfo() * to get the right name. Not sure about other platforms. * * NOTE: GNU libiconv does not recognize the character-set name * which some platform nl_langinfo(CODESET) returns, so we should * use locale_charset() instead of nl_langinfo(CODESET) for GNU libiconv. */ static const char * default_iconv_charset(const char *charset) { if (charset != NULL && charset[0] != '\0') return charset; #if HAVE_LOCALE_CHARSET && !defined(__APPLE__) /* locale_charset() is broken on Mac OS */ return locale_charset(); #elif HAVE_NL_LANGINFO return nl_langinfo(CODESET); #else return ""; #endif } #if defined(_WIN32) && !defined(__CYGWIN__) /* * Convert MBS to WCS. * Note: returns -1 if conversion fails. */ int archive_wstring_append_from_mbs(struct archive_wstring *dest, const char *p, size_t len) { return archive_wstring_append_from_mbs_in_codepage(dest, p, len, NULL); } static int archive_wstring_append_from_mbs_in_codepage(struct archive_wstring *dest, const char *s, size_t length, struct archive_string_conv *sc) { int count, ret = 0; UINT from_cp; if (sc != NULL) from_cp = sc->from_cp; else from_cp = get_current_codepage(); if (from_cp == CP_C_LOCALE) { /* * "C" locale special process. */ wchar_t *ws; const unsigned char *mp; if (NULL == archive_wstring_ensure(dest, dest->length + length + 1)) return (-1); ws = dest->s + dest->length; mp = (const unsigned char *)s; count = 0; while (count < (int)length && *mp) { *ws++ = (wchar_t)*mp++; count++; } } else if (sc != NULL && (sc->flag & (SCONV_NORMALIZATION_C | SCONV_NORMALIZATION_D))) { /* * Normalize UTF-8 and UTF-16BE and convert it directly * to UTF-16 as wchar_t. */ struct archive_string u16; int saved_flag = sc->flag;/* save current flag. */ if (is_big_endian()) sc->flag |= SCONV_TO_UTF16BE; else sc->flag |= SCONV_TO_UTF16LE; if (sc->flag & SCONV_FROM_UTF16) { /* * UTF-16BE/LE NFD ===> UTF-16 NFC * UTF-16BE/LE NFC ===> UTF-16 NFD */ count = (int)utf16nbytes(s, length); } else { /* * UTF-8 NFD ===> UTF-16 NFC * UTF-8 NFC ===> UTF-16 NFD */ count = (int)mbsnbytes(s, length); } u16.s = (char *)dest->s; u16.length = dest->length << 1;; u16.buffer_length = dest->buffer_length; if (sc->flag & SCONV_NORMALIZATION_C) ret = archive_string_normalize_C(&u16, s, count, sc); else ret = archive_string_normalize_D(&u16, s, count, sc); dest->s = (wchar_t *)u16.s; dest->length = u16.length >> 1; dest->buffer_length = u16.buffer_length; sc->flag = saved_flag;/* restore the saved flag. */ return (ret); } else if (sc != NULL && (sc->flag & SCONV_FROM_UTF16)) { count = (int)utf16nbytes(s, length); count >>= 1; /* to be WCS length */ /* Allocate memory for WCS. */ if (NULL == archive_wstring_ensure(dest, dest->length + count + 1)) return (-1); wmemcpy(dest->s + dest->length, (const wchar_t *)s, count); if ((sc->flag & SCONV_FROM_UTF16BE) && !is_big_endian()) { uint16_t *u16 = (uint16_t *)(dest->s + dest->length); int b; for (b = 0; b < count; b++) { uint16_t val = archive_le16dec(u16+b); archive_be16enc(u16+b, val); } } else if ((sc->flag & SCONV_FROM_UTF16LE) && is_big_endian()) { uint16_t *u16 = (uint16_t *)(dest->s + dest->length); int b; for (b = 0; b < count; b++) { uint16_t val = archive_be16dec(u16+b); archive_le16enc(u16+b, val); } } } else { DWORD mbflag; size_t buffsize; if (sc == NULL) mbflag = 0; else if (sc->flag & SCONV_FROM_CHARSET) { /* Do not trust the length which comes from * an archive file. */ length = mbsnbytes(s, length); mbflag = 0; } else mbflag = MB_PRECOMPOSED; buffsize = dest->length + length + 1; do { /* Allocate memory for WCS. */ if (NULL == archive_wstring_ensure(dest, buffsize)) return (-1); /* Convert MBS to WCS. */ count = MultiByteToWideChar(from_cp, mbflag, s, (int)length, dest->s + dest->length, (int)(dest->buffer_length >> 1) -1); if (count == 0 && GetLastError() == ERROR_INSUFFICIENT_BUFFER) { /* Expand the WCS buffer. */ buffsize = dest->buffer_length << 1; continue; } if (count == 0 && length != 0) ret = -1; break; } while (1); } dest->length += count; dest->s[dest->length] = L'\0'; return (ret); } #else /* * Convert MBS to WCS. * Note: returns -1 if conversion fails. */ int archive_wstring_append_from_mbs(struct archive_wstring *dest, const char *p, size_t len) { size_t r; int ret_val = 0; /* * No single byte will be more than one wide character, * so this length estimate will always be big enough. */ size_t wcs_length = len; size_t mbs_length = len; const char *mbs = p; wchar_t *wcs; #if HAVE_MBRTOWC mbstate_t shift_state; memset(&shift_state, 0, sizeof(shift_state)); #endif if (NULL == archive_wstring_ensure(dest, dest->length + wcs_length + 1)) return (-1); wcs = dest->s + dest->length; /* * We cannot use mbsrtowcs/mbstowcs here because those may convert * extra MBS when strlen(p) > len and one wide character consists of * multi bytes. */ while (*mbs && mbs_length > 0) { if (wcs_length == 0) { dest->length = wcs - dest->s; dest->s[dest->length] = L'\0'; wcs_length = mbs_length; if (NULL == archive_wstring_ensure(dest, dest->length + wcs_length + 1)) return (-1); wcs = dest->s + dest->length; } #if HAVE_MBRTOWC r = mbrtowc(wcs, mbs, wcs_length, &shift_state); #else r = mbtowc(wcs, mbs, wcs_length); #endif if (r == (size_t)-1 || r == (size_t)-2) { ret_val = -1; if (errno == EILSEQ) { ++mbs; --mbs_length; continue; } else break; } if (r == 0 || r > mbs_length) break; wcs++; wcs_length--; mbs += r; mbs_length -= r; } dest->length = wcs - dest->s; dest->s[dest->length] = L'\0'; return (ret_val); } #endif #if defined(_WIN32) && !defined(__CYGWIN__) /* * WCS ==> MBS. * Note: returns -1 if conversion fails. * * Win32 builds use WideCharToMultiByte from the Windows API. * (Maybe Cygwin should too? WideCharToMultiByte will know a * lot more about local character encodings than the wcrtomb() * wrapper is going to know.) */ int archive_string_append_from_wcs(struct archive_string *as, const wchar_t *w, size_t len) { return archive_string_append_from_wcs_in_codepage(as, w, len, NULL); } static int archive_string_append_from_wcs_in_codepage(struct archive_string *as, const wchar_t *ws, size_t len, struct archive_string_conv *sc) { BOOL defchar_used, *dp; int count, ret = 0; UINT to_cp; int wslen = (int)len; if (sc != NULL) to_cp = sc->to_cp; else to_cp = get_current_codepage(); if (to_cp == CP_C_LOCALE) { /* * "C" locale special process. */ const wchar_t *wp = ws; char *p; if (NULL == archive_string_ensure(as, as->length + wslen +1)) return (-1); p = as->s + as->length; count = 0; defchar_used = 0; while (count < wslen && *wp) { if (*wp > 255) { *p++ = '?'; wp++; defchar_used = 1; } else *p++ = (char)*wp++; count++; } } else if (sc != NULL && (sc->flag & SCONV_TO_UTF16)) { uint16_t *u16; if (NULL == archive_string_ensure(as, as->length + len * 2 + 2)) return (-1); u16 = (uint16_t *)(as->s + as->length); count = 0; defchar_used = 0; if (sc->flag & SCONV_TO_UTF16BE) { while (count < (int)len && *ws) { archive_be16enc(u16+count, *ws); ws++; count++; } } else { while (count < (int)len && *ws) { archive_le16enc(u16+count, *ws); ws++; count++; } } count <<= 1; /* to be byte size */ } else { /* Make sure the MBS buffer has plenty to set. */ if (NULL == archive_string_ensure(as, as->length + len * 2 + 1)) return (-1); do { defchar_used = 0; if (to_cp == CP_UTF8 || sc == NULL) dp = NULL; else dp = &defchar_used; count = WideCharToMultiByte(to_cp, 0, ws, wslen, as->s + as->length, (int)as->buffer_length-1, NULL, dp); if (count == 0 && GetLastError() == ERROR_INSUFFICIENT_BUFFER) { /* Expand the MBS buffer and retry. */ if (NULL == archive_string_ensure(as, as->buffer_length + len)) return (-1); continue; } if (count == 0) ret = -1; break; } while (1); } as->length += count; as->s[as->length] = '\0'; return (defchar_used?-1:ret); } #elif defined(HAVE_WCTOMB) || defined(HAVE_WCRTOMB) /* * Translates a wide character string into current locale character set * and appends to the archive_string. Note: returns -1 if conversion * fails. */ int archive_string_append_from_wcs(struct archive_string *as, const wchar_t *w, size_t len) { /* We cannot use the standard wcstombs() here because it * cannot tell us how big the output buffer should be. So * I've built a loop around wcrtomb() or wctomb() that * converts a character at a time and resizes the string as * needed. We prefer wcrtomb() when it's available because * it's thread-safe. */ int n, ret_val = 0; char *p; char *end; #if HAVE_WCRTOMB mbstate_t shift_state; memset(&shift_state, 0, sizeof(shift_state)); #else /* Clear the shift state before starting. */ wctomb(NULL, L'\0'); #endif /* * Allocate buffer for MBS. * We need this allocation here since it is possible that * as->s is still NULL. */ if (archive_string_ensure(as, as->length + len + 1) == NULL) return (-1); p = as->s + as->length; end = as->s + as->buffer_length - MB_CUR_MAX -1; while (*w != L'\0' && len > 0) { if (p >= end) { as->length = p - as->s; as->s[as->length] = '\0'; /* Re-allocate buffer for MBS. */ if (archive_string_ensure(as, as->length + len * 2 + 1) == NULL) return (-1); p = as->s + as->length; end = as->s + as->buffer_length - MB_CUR_MAX -1; } #if HAVE_WCRTOMB n = wcrtomb(p, *w++, &shift_state); #else n = wctomb(p, *w++); #endif if (n == -1) { if (errno == EILSEQ) { /* Skip an illegal wide char. */ *p++ = '?'; ret_val = -1; } else { ret_val = -1; break; } } else p += n; len--; } as->length = p - as->s; as->s[as->length] = '\0'; return (ret_val); } #else /* HAVE_WCTOMB || HAVE_WCRTOMB */ /* * TODO: Test if __STDC_ISO_10646__ is defined. * Non-Windows uses ISO C wcrtomb() or wctomb() to perform the conversion * one character at a time. If a non-Windows platform doesn't have * either of these, fall back to the built-in UTF8 conversion. */ int archive_string_append_from_wcs(struct archive_string *as, const wchar_t *w, size_t len) { (void)as;/* UNUSED */ (void)w;/* UNUSED */ (void)len;/* UNUSED */ errno = ENOSYS; return (-1); } #endif /* HAVE_WCTOMB || HAVE_WCRTOMB */ /* * Find a string conversion object by a pair of 'from' charset name * and 'to' charset name from an archive object. * Return NULL if not found. */ static struct archive_string_conv * find_sconv_object(struct archive *a, const char *fc, const char *tc) { struct archive_string_conv *sc; if (a == NULL) return (NULL); for (sc = a->sconv; sc != NULL; sc = sc->next) { if (strcmp(sc->from_charset, fc) == 0 && strcmp(sc->to_charset, tc) == 0) break; } return (sc); } /* * Register a string object to an archive object. */ static void add_sconv_object(struct archive *a, struct archive_string_conv *sc) { struct archive_string_conv **psc; /* Add a new sconv to sconv list. */ psc = &(a->sconv); while (*psc != NULL) psc = &((*psc)->next); *psc = sc; } static void add_converter(struct archive_string_conv *sc, int (*converter) (struct archive_string *, const void *, size_t, struct archive_string_conv *)) { if (sc == NULL || sc->nconverter >= 2) __archive_errx(1, "Programing error"); sc->converter[sc->nconverter++] = converter; } static void setup_converter(struct archive_string_conv *sc) { /* Reset. */ sc->nconverter = 0; /* * Perform special sequence for the incorrect UTF-8 filenames * made by libarchive2.x. */ if (sc->flag & SCONV_UTF8_LIBARCHIVE_2) { add_converter(sc, strncat_from_utf8_libarchive2); return; } /* * Convert a string to UTF-16BE/LE. */ if (sc->flag & SCONV_TO_UTF16) { /* * If the current locale is UTF-8, we can translate * a UTF-8 string into a UTF-16BE string. */ if (sc->flag & SCONV_FROM_UTF8) { add_converter(sc, archive_string_append_unicode); return; } #if defined(_WIN32) && !defined(__CYGWIN__) if (sc->flag & SCONV_WIN_CP) { if (sc->flag & SCONV_TO_UTF16BE) add_converter(sc, win_strncat_to_utf16be); else add_converter(sc, win_strncat_to_utf16le); return; } #endif #if defined(HAVE_ICONV) if (sc->cd != (iconv_t)-1) { add_converter(sc, iconv_strncat_in_locale); return; } #endif if (sc->flag & SCONV_BEST_EFFORT) { if (sc->flag & SCONV_TO_UTF16BE) add_converter(sc, best_effort_strncat_to_utf16be); else add_converter(sc, best_effort_strncat_to_utf16le); } else /* Make sure we have no converter. */ sc->nconverter = 0; return; } /* * Convert a string from UTF-16BE/LE. */ if (sc->flag & SCONV_FROM_UTF16) { /* * At least we should normalize a UTF-16BE string. */ if (sc->flag & SCONV_NORMALIZATION_D) add_converter(sc,archive_string_normalize_D); else if (sc->flag & SCONV_NORMALIZATION_C) add_converter(sc, archive_string_normalize_C); if (sc->flag & SCONV_TO_UTF8) { /* * If the current locale is UTF-8, we can translate * a UTF-16BE/LE string into a UTF-8 string directly. */ if (!(sc->flag & (SCONV_NORMALIZATION_D |SCONV_NORMALIZATION_C))) add_converter(sc, archive_string_append_unicode); return; } #if defined(_WIN32) && !defined(__CYGWIN__) if (sc->flag & SCONV_WIN_CP) { if (sc->flag & SCONV_FROM_UTF16BE) add_converter(sc, win_strncat_from_utf16be); else add_converter(sc, win_strncat_from_utf16le); return; } #endif #if defined(HAVE_ICONV) if (sc->cd != (iconv_t)-1) { add_converter(sc, iconv_strncat_in_locale); return; } #endif if ((sc->flag & (SCONV_BEST_EFFORT | SCONV_FROM_UTF16BE)) == (SCONV_BEST_EFFORT | SCONV_FROM_UTF16BE)) add_converter(sc, best_effort_strncat_from_utf16be); else if ((sc->flag & (SCONV_BEST_EFFORT | SCONV_FROM_UTF16LE)) == (SCONV_BEST_EFFORT | SCONV_FROM_UTF16LE)) add_converter(sc, best_effort_strncat_from_utf16le); else /* Make sure we have no converter. */ sc->nconverter = 0; return; } if (sc->flag & SCONV_FROM_UTF8) { /* * At least we should normalize a UTF-8 string. */ if (sc->flag & SCONV_NORMALIZATION_D) add_converter(sc,archive_string_normalize_D); else if (sc->flag & SCONV_NORMALIZATION_C) add_converter(sc, archive_string_normalize_C); /* * Copy UTF-8 string with a check of CESU-8. * Apparently, iconv does not check surrogate pairs in UTF-8 * when both from-charset and to-charset are UTF-8, and then * we use our UTF-8 copy code. */ if (sc->flag & SCONV_TO_UTF8) { /* * If the current locale is UTF-8, we can translate * a UTF-16BE string into a UTF-8 string directly. */ if (!(sc->flag & (SCONV_NORMALIZATION_D |SCONV_NORMALIZATION_C))) add_converter(sc, strncat_from_utf8_to_utf8); return; } } #if defined(_WIN32) && !defined(__CYGWIN__) /* * On Windows we can use Windows API for a string conversion. */ if (sc->flag & SCONV_WIN_CP) { add_converter(sc, strncat_in_codepage); return; } #endif #if HAVE_ICONV if (sc->cd != (iconv_t)-1) { add_converter(sc, iconv_strncat_in_locale); /* * iconv generally does not support UTF-8-MAC and so * we have to the output of iconv from NFC to NFD if * need. */ if ((sc->flag & SCONV_FROM_CHARSET) && (sc->flag & SCONV_TO_UTF8)) { if (sc->flag & SCONV_NORMALIZATION_D) add_converter(sc, archive_string_normalize_D); } return; } #endif /* * Try conversion in the best effort or no conversion. */ if ((sc->flag & SCONV_BEST_EFFORT) || sc->same) add_converter(sc, best_effort_strncat_in_locale); else /* Make sure we have no converter. */ sc->nconverter = 0; } /* * Return canonicalized charset-name but this supports just UTF-8, UTF-16BE * and CP932 which are referenced in create_sconv_object(). */ static const char * canonical_charset_name(const char *charset) { char cs[16]; char *p; const char *s; if (charset == NULL || charset[0] == '\0' || strlen(charset) > 15) return (charset); /* Copy name to uppercase. */ p = cs; s = charset; while (*s) { char c = *s++; if (c >= 'a' && c <= 'z') c -= 'a' - 'A'; *p++ = c; } *p++ = '\0'; if (strcmp(cs, "UTF-8") == 0 || strcmp(cs, "UTF8") == 0) return ("UTF-8"); if (strcmp(cs, "UTF-16BE") == 0 || strcmp(cs, "UTF16BE") == 0) return ("UTF-16BE"); if (strcmp(cs, "UTF-16LE") == 0 || strcmp(cs, "UTF16LE") == 0) return ("UTF-16LE"); if (strcmp(cs, "CP932") == 0) return ("CP932"); return (charset); } /* * Create a string conversion object. */ static struct archive_string_conv * create_sconv_object(const char *fc, const char *tc, unsigned current_codepage, int flag) { struct archive_string_conv *sc; sc = calloc(1, sizeof(*sc)); if (sc == NULL) return (NULL); sc->next = NULL; sc->from_charset = strdup(fc); if (sc->from_charset == NULL) { free(sc); return (NULL); } sc->to_charset = strdup(tc); if (sc->to_charset == NULL) { free(sc->from_charset); free(sc); return (NULL); } archive_string_init(&sc->utftmp); if (flag & SCONV_TO_CHARSET) { /* * Convert characters from the current locale charset to * a specified charset. */ sc->from_cp = current_codepage; sc->to_cp = make_codepage_from_charset(tc); #if defined(_WIN32) && !defined(__CYGWIN__) if (IsValidCodePage(sc->to_cp)) flag |= SCONV_WIN_CP; #endif } else if (flag & SCONV_FROM_CHARSET) { /* * Convert characters from a specified charset to * the current locale charset. */ sc->to_cp = current_codepage; sc->from_cp = make_codepage_from_charset(fc); #if defined(_WIN32) && !defined(__CYGWIN__) if (IsValidCodePage(sc->from_cp)) flag |= SCONV_WIN_CP; #endif } /* * Check if "from charset" and "to charset" are the same. */ if (strcmp(fc, tc) == 0 || (sc->from_cp != (unsigned)-1 && sc->from_cp == sc->to_cp)) sc->same = 1; else sc->same = 0; /* * Mark if "from charset" or "to charset" are UTF-8 or UTF-16BE/LE. */ if (strcmp(tc, "UTF-8") == 0) flag |= SCONV_TO_UTF8; else if (strcmp(tc, "UTF-16BE") == 0) flag |= SCONV_TO_UTF16BE; else if (strcmp(tc, "UTF-16LE") == 0) flag |= SCONV_TO_UTF16LE; if (strcmp(fc, "UTF-8") == 0) flag |= SCONV_FROM_UTF8; else if (strcmp(fc, "UTF-16BE") == 0) flag |= SCONV_FROM_UTF16BE; else if (strcmp(fc, "UTF-16LE") == 0) flag |= SCONV_FROM_UTF16LE; #if defined(_WIN32) && !defined(__CYGWIN__) if (sc->to_cp == CP_UTF8) flag |= SCONV_TO_UTF8; else if (sc->to_cp == CP_UTF16BE) flag |= SCONV_TO_UTF16BE | SCONV_WIN_CP; else if (sc->to_cp == CP_UTF16LE) flag |= SCONV_TO_UTF16LE | SCONV_WIN_CP; if (sc->from_cp == CP_UTF8) flag |= SCONV_FROM_UTF8; else if (sc->from_cp == CP_UTF16BE) flag |= SCONV_FROM_UTF16BE | SCONV_WIN_CP; else if (sc->from_cp == CP_UTF16LE) flag |= SCONV_FROM_UTF16LE | SCONV_WIN_CP; #endif /* * Set a flag for Unicode NFD. Usually iconv cannot correctly * handle it. So we have to translate NFD characters to NFC ones * ourselves before iconv handles. Another reason is to prevent * that the same sight of two filenames, one is NFC and other * is NFD, would be in its directory. * On Mac OS X, although its filesystem layer automatically * convert filenames to NFD, it would be useful for filename * comparing to find out the same filenames that we normalize * that to be NFD ourselves. */ if ((flag & SCONV_FROM_CHARSET) && (flag & (SCONV_FROM_UTF16 | SCONV_FROM_UTF8))) { #if defined(__APPLE__) if (flag & SCONV_TO_UTF8) flag |= SCONV_NORMALIZATION_D; else #endif flag |= SCONV_NORMALIZATION_C; } #if defined(__APPLE__) /* * In case writing an archive file, make sure that a filename * going to be passed to iconv is a Unicode NFC string since * a filename in HFS Plus filesystem is a Unicode NFD one and * iconv cannot handle it with "UTF-8" charset. It is simpler * than a use of "UTF-8-MAC" charset. */ if ((flag & SCONV_TO_CHARSET) && (flag & (SCONV_FROM_UTF16 | SCONV_FROM_UTF8)) && !(flag & (SCONV_TO_UTF16 | SCONV_TO_UTF8))) flag |= SCONV_NORMALIZATION_C; /* * In case reading an archive file. make sure that a filename * will be passed to users is a Unicode NFD string in order to * correctly compare the filename with other one which comes * from HFS Plus filesystem. */ if ((flag & SCONV_FROM_CHARSET) && !(flag & (SCONV_FROM_UTF16 | SCONV_FROM_UTF8)) && (flag & SCONV_TO_UTF8)) flag |= SCONV_NORMALIZATION_D; #endif #if defined(HAVE_ICONV) sc->cd_w = (iconv_t)-1; /* * Create an iconv object. */ if (((flag & (SCONV_TO_UTF8 | SCONV_TO_UTF16)) && (flag & (SCONV_FROM_UTF8 | SCONV_FROM_UTF16))) || (flag & SCONV_WIN_CP)) { /* This case we won't use iconv. */ sc->cd = (iconv_t)-1; } else { sc->cd = iconv_open(tc, fc); if (sc->cd == (iconv_t)-1 && (sc->flag & SCONV_BEST_EFFORT)) { /* * Unfortunately, all of iconv implements do support * "CP932" character-set, so we should use "SJIS" * instead if iconv_open failed. */ if (strcmp(tc, "CP932") == 0) sc->cd = iconv_open("SJIS", fc); else if (strcmp(fc, "CP932") == 0) sc->cd = iconv_open(tc, "SJIS"); } #if defined(_WIN32) && !defined(__CYGWIN__) /* * archive_mstring on Windows directly convert multi-bytes * into archive_wstring in order not to depend on locale * so that you can do a I18N programming. This will be * used only in archive_mstring_copy_mbs_len_l so far. */ if (flag & SCONV_FROM_CHARSET) { sc->cd_w = iconv_open("UTF-8", fc); if (sc->cd_w == (iconv_t)-1 && (sc->flag & SCONV_BEST_EFFORT)) { if (strcmp(fc, "CP932") == 0) sc->cd_w = iconv_open("UTF-8", "SJIS"); } } #endif /* _WIN32 && !__CYGWIN__ */ } #endif /* HAVE_ICONV */ sc->flag = flag; /* * Set up converters. */ setup_converter(sc); return (sc); } /* * Free a string conversion object. */ static void free_sconv_object(struct archive_string_conv *sc) { free(sc->from_charset); free(sc->to_charset); archive_string_free(&sc->utftmp); #if HAVE_ICONV if (sc->cd != (iconv_t)-1) iconv_close(sc->cd); if (sc->cd_w != (iconv_t)-1) iconv_close(sc->cd_w); #endif free(sc); } #if defined(_WIN32) && !defined(__CYGWIN__) static unsigned my_atoi(const char *p) { unsigned cp; cp = 0; while (*p) { if (*p >= '0' && *p <= '9') cp = cp * 10 + (*p - '0'); else return (-1); p++; } return (cp); } /* * Translate Charset name (as used by iconv) into CodePage (as used by Windows) * Return -1 if failed. * * Note: This translation code may be insufficient. */ static struct charset { const char *name; unsigned cp; } charsets[] = { /* MUST BE SORTED! */ {"ASCII", 1252}, {"ASMO-708", 708}, {"BIG5", 950}, {"CHINESE", 936}, {"CP367", 1252}, {"CP819", 1252}, {"CP1025", 21025}, {"DOS-720", 720}, {"DOS-862", 862}, {"EUC-CN", 51936}, {"EUC-JP", 51932}, {"EUC-KR", 949}, {"EUCCN", 51936}, {"EUCJP", 51932}, {"EUCKR", 949}, {"GB18030", 54936}, {"GB2312", 936}, {"HEBREW", 1255}, {"HZ-GB-2312", 52936}, {"IBM273", 20273}, {"IBM277", 20277}, {"IBM278", 20278}, {"IBM280", 20280}, {"IBM284", 20284}, {"IBM285", 20285}, {"IBM290", 20290}, {"IBM297", 20297}, {"IBM367", 1252}, {"IBM420", 20420}, {"IBM423", 20423}, {"IBM424", 20424}, {"IBM819", 1252}, {"IBM871", 20871}, {"IBM880", 20880}, {"IBM905", 20905}, {"IBM924", 20924}, {"ISO-8859-1", 28591}, {"ISO-8859-13", 28603}, {"ISO-8859-15", 28605}, {"ISO-8859-2", 28592}, {"ISO-8859-3", 28593}, {"ISO-8859-4", 28594}, {"ISO-8859-5", 28595}, {"ISO-8859-6", 28596}, {"ISO-8859-7", 28597}, {"ISO-8859-8", 28598}, {"ISO-8859-9", 28599}, {"ISO8859-1", 28591}, {"ISO8859-13", 28603}, {"ISO8859-15", 28605}, {"ISO8859-2", 28592}, {"ISO8859-3", 28593}, {"ISO8859-4", 28594}, {"ISO8859-5", 28595}, {"ISO8859-6", 28596}, {"ISO8859-7", 28597}, {"ISO8859-8", 28598}, {"ISO8859-9", 28599}, {"JOHAB", 1361}, {"KOI8-R", 20866}, {"KOI8-U", 21866}, {"KS_C_5601-1987", 949}, {"LATIN1", 1252}, {"LATIN2", 28592}, {"MACINTOSH", 10000}, {"SHIFT-JIS", 932}, {"SHIFT_JIS", 932}, {"SJIS", 932}, {"US", 1252}, {"US-ASCII", 1252}, {"UTF-16", 1200}, {"UTF-16BE", 1201}, {"UTF-16LE", 1200}, {"UTF-8", CP_UTF8}, {"X-EUROPA", 29001}, {"X-MAC-ARABIC", 10004}, {"X-MAC-CE", 10029}, {"X-MAC-CHINESEIMP", 10008}, {"X-MAC-CHINESETRAD", 10002}, {"X-MAC-CROATIAN", 10082}, {"X-MAC-CYRILLIC", 10007}, {"X-MAC-GREEK", 10006}, {"X-MAC-HEBREW", 10005}, {"X-MAC-ICELANDIC", 10079}, {"X-MAC-JAPANESE", 10001}, {"X-MAC-KOREAN", 10003}, {"X-MAC-ROMANIAN", 10010}, {"X-MAC-THAI", 10021}, {"X-MAC-TURKISH", 10081}, {"X-MAC-UKRAINIAN", 10017}, }; static unsigned make_codepage_from_charset(const char *charset) { char cs[16]; char *p; unsigned cp; int a, b; if (charset == NULL || strlen(charset) > 15) return -1; /* Copy name to uppercase. */ p = cs; while (*charset) { char c = *charset++; if (c >= 'a' && c <= 'z') c -= 'a' - 'A'; *p++ = c; } *p++ = '\0'; cp = -1; /* Look it up in the table first, so that we can easily * override CP367, which we map to 1252 instead of 367. */ a = 0; b = sizeof(charsets)/sizeof(charsets[0]); while (b > a) { int c = (b + a) / 2; int r = strcmp(charsets[c].name, cs); if (r < 0) a = c + 1; else if (r > 0) b = c; else return charsets[c].cp; } /* If it's not in the table, try to parse it. */ switch (*cs) { case 'C': if (cs[1] == 'P' && cs[2] >= '0' && cs[2] <= '9') { cp = my_atoi(cs + 2); } else if (strcmp(cs, "CP_ACP") == 0) cp = get_current_codepage(); else if (strcmp(cs, "CP_OEMCP") == 0) cp = get_current_oemcp(); break; case 'I': if (cs[1] == 'B' && cs[2] == 'M' && cs[3] >= '0' && cs[3] <= '9') { cp = my_atoi(cs + 3); } break; case 'W': if (strncmp(cs, "WINDOWS-", 8) == 0) { cp = my_atoi(cs + 8); if (cp != 874 && (cp < 1250 || cp > 1258)) cp = -1;/* This may invalid code. */ } break; } return (cp); } /* * Return ANSI Code Page of current locale set by setlocale(). */ static unsigned get_current_codepage(void) { char *locale, *p; unsigned cp; locale = setlocale(LC_CTYPE, NULL); if (locale == NULL) return (GetACP()); if (locale[0] == 'C' && locale[1] == '\0') return (CP_C_LOCALE); p = strrchr(locale, '.'); if (p == NULL) return (GetACP()); cp = my_atoi(p+1); if (cp <= 0) return (GetACP()); return (cp); } /* * Translation table between Locale Name and ACP/OEMCP. */ static struct { unsigned acp; unsigned ocp; const char *locale; } acp_ocp_map[] = { { 950, 950, "Chinese_Taiwan" }, { 936, 936, "Chinese_People's Republic of China" }, { 950, 950, "Chinese_Taiwan" }, { 1250, 852, "Czech_Czech Republic" }, { 1252, 850, "Danish_Denmark" }, { 1252, 850, "Dutch_Netherlands" }, { 1252, 850, "Dutch_Belgium" }, { 1252, 437, "English_United States" }, { 1252, 850, "English_Australia" }, { 1252, 850, "English_Canada" }, { 1252, 850, "English_New Zealand" }, { 1252, 850, "English_United Kingdom" }, { 1252, 437, "English_United States" }, { 1252, 850, "Finnish_Finland" }, { 1252, 850, "French_France" }, { 1252, 850, "French_Belgium" }, { 1252, 850, "French_Canada" }, { 1252, 850, "French_Switzerland" }, { 1252, 850, "German_Germany" }, { 1252, 850, "German_Austria" }, { 1252, 850, "German_Switzerland" }, { 1253, 737, "Greek_Greece" }, { 1250, 852, "Hungarian_Hungary" }, { 1252, 850, "Icelandic_Iceland" }, { 1252, 850, "Italian_Italy" }, { 1252, 850, "Italian_Switzerland" }, { 932, 932, "Japanese_Japan" }, { 949, 949, "Korean_Korea" }, { 1252, 850, "Norwegian (BokmOl)_Norway" }, { 1252, 850, "Norwegian (BokmOl)_Norway" }, { 1252, 850, "Norwegian-Nynorsk_Norway" }, { 1250, 852, "Polish_Poland" }, { 1252, 850, "Portuguese_Portugal" }, { 1252, 850, "Portuguese_Brazil" }, { 1251, 866, "Russian_Russia" }, { 1250, 852, "Slovak_Slovakia" }, { 1252, 850, "Spanish_Spain" }, { 1252, 850, "Spanish_Mexico" }, { 1252, 850, "Spanish_Spain" }, { 1252, 850, "Swedish_Sweden" }, { 1254, 857, "Turkish_Turkey" }, { 0, 0, NULL} }; /* * Return OEM Code Page of current locale set by setlocale(). */ static unsigned get_current_oemcp(void) { int i; char *locale, *p; size_t len; locale = setlocale(LC_CTYPE, NULL); if (locale == NULL) return (GetOEMCP()); if (locale[0] == 'C' && locale[1] == '\0') return (CP_C_LOCALE); p = strrchr(locale, '.'); if (p == NULL) return (GetOEMCP()); len = p - locale; for (i = 0; acp_ocp_map[i].acp; i++) { if (strncmp(acp_ocp_map[i].locale, locale, len) == 0) return (acp_ocp_map[i].ocp); } return (GetOEMCP()); } #else /* * POSIX platform does not use CodePage. */ static unsigned get_current_codepage(void) { return (-1);/* Unknown */ } static unsigned make_codepage_from_charset(const char *charset) { (void)charset; /* UNUSED */ return (-1);/* Unknown */ } static unsigned get_current_oemcp(void) { return (-1);/* Unknown */ } #endif /* defined(_WIN32) && !defined(__CYGWIN__) */ /* * Return a string conversion object. */ static struct archive_string_conv * get_sconv_object(struct archive *a, const char *fc, const char *tc, int flag) { struct archive_string_conv *sc; unsigned current_codepage; /* Check if we have made the sconv object. */ sc = find_sconv_object(a, fc, tc); if (sc != NULL) return (sc); if (a == NULL) current_codepage = get_current_codepage(); else current_codepage = a->current_codepage; sc = create_sconv_object(canonical_charset_name(fc), canonical_charset_name(tc), current_codepage, flag); if (sc == NULL) { if (a != NULL) archive_set_error(a, ENOMEM, "Could not allocate memory for " "a string conversion object"); return (NULL); } /* * If there is no converter for current string conversion object, * we cannot handle this conversion. */ if (sc->nconverter == 0) { if (a != NULL) { #if HAVE_ICONV archive_set_error(a, ARCHIVE_ERRNO_MISC, "iconv_open failed : Cannot handle ``%s''", (flag & SCONV_TO_CHARSET)?tc:fc); #else archive_set_error(a, ARCHIVE_ERRNO_MISC, "A character-set conversion not fully supported " "on this platform"); #endif } /* Failed; free a sconv object. */ free_sconv_object(sc); return (NULL); } /* * Success! */ if (a != NULL) add_sconv_object(a, sc); return (sc); } static const char * get_current_charset(struct archive *a) { const char *cur_charset; if (a == NULL) cur_charset = default_iconv_charset(""); else { cur_charset = default_iconv_charset(a->current_code); if (a->current_code == NULL) { a->current_code = strdup(cur_charset); a->current_codepage = get_current_codepage(); a->current_oemcp = get_current_oemcp(); } } return (cur_charset); } /* * Make and Return a string conversion object. * Return NULL if the platform does not support the specified conversion * and best_effort is 0. * If best_effort is set, A string conversion object must be returned * unless memory allocation for the object fails, but the conversion * might fail when non-ASCII code is found. */ struct archive_string_conv * archive_string_conversion_to_charset(struct archive *a, const char *charset, int best_effort) { int flag = SCONV_TO_CHARSET; if (best_effort) flag |= SCONV_BEST_EFFORT; return (get_sconv_object(a, get_current_charset(a), charset, flag)); } struct archive_string_conv * archive_string_conversion_from_charset(struct archive *a, const char *charset, int best_effort) { int flag = SCONV_FROM_CHARSET; if (best_effort) flag |= SCONV_BEST_EFFORT; return (get_sconv_object(a, charset, get_current_charset(a), flag)); } /* * archive_string_default_conversion_*_archive() are provided for Windows * platform because other archiver application use CP_OEMCP for * MultiByteToWideChar() and WideCharToMultiByte() for the filenames * in tar or zip files. But mbstowcs/wcstombs(CRT) usually use CP_ACP * unless you use setlocale(LC_ALL, ".OCP")(specify CP_OEMCP). * So we should make a string conversion between CP_ACP and CP_OEMCP * for compatibility. */ #if defined(_WIN32) && !defined(__CYGWIN__) struct archive_string_conv * archive_string_default_conversion_for_read(struct archive *a) { const char *cur_charset = get_current_charset(a); char oemcp[16]; /* NOTE: a check of cur_charset is unneeded but we need * that get_current_charset() has been surely called at * this time whatever C compiler optimized. */ if (cur_charset != NULL && (a->current_codepage == CP_C_LOCALE || a->current_codepage == a->current_oemcp)) return (NULL);/* no conversion. */ _snprintf(oemcp, sizeof(oemcp)-1, "CP%d", a->current_oemcp); /* Make sure a null termination must be set. */ oemcp[sizeof(oemcp)-1] = '\0'; return (get_sconv_object(a, oemcp, cur_charset, SCONV_FROM_CHARSET)); } struct archive_string_conv * archive_string_default_conversion_for_write(struct archive *a) { const char *cur_charset = get_current_charset(a); char oemcp[16]; /* NOTE: a check of cur_charset is unneeded but we need * that get_current_charset() has been surely called at * this time whatever C compiler optimized. */ if (cur_charset != NULL && (a->current_codepage == CP_C_LOCALE || a->current_codepage == a->current_oemcp)) return (NULL);/* no conversion. */ _snprintf(oemcp, sizeof(oemcp)-1, "CP%d", a->current_oemcp); /* Make sure a null termination must be set. */ oemcp[sizeof(oemcp)-1] = '\0'; return (get_sconv_object(a, cur_charset, oemcp, SCONV_TO_CHARSET)); } #else struct archive_string_conv * archive_string_default_conversion_for_read(struct archive *a) { (void)a; /* UNUSED */ return (NULL); } struct archive_string_conv * archive_string_default_conversion_for_write(struct archive *a) { (void)a; /* UNUSED */ return (NULL); } #endif /* * Dispose of all character conversion objects in the archive object. */ void archive_string_conversion_free(struct archive *a) { struct archive_string_conv *sc; struct archive_string_conv *sc_next; for (sc = a->sconv; sc != NULL; sc = sc_next) { sc_next = sc->next; free_sconv_object(sc); } a->sconv = NULL; free(a->current_code); a->current_code = NULL; } /* * Return a conversion charset name. */ const char * archive_string_conversion_charset_name(struct archive_string_conv *sc) { if (sc->flag & SCONV_TO_CHARSET) return (sc->to_charset); else return (sc->from_charset); } /* * Change the behavior of a string conversion. */ void archive_string_conversion_set_opt(struct archive_string_conv *sc, int opt) { switch (opt) { /* * A filename in UTF-8 was made with libarchive 2.x in a wrong * assumption that wchar_t was Unicode. * This option enables simulating the assumption in order to read * that filename correctly. */ case SCONV_SET_OPT_UTF8_LIBARCHIVE2X: #if (defined(_WIN32) && !defined(__CYGWIN__)) \ || defined(__STDC_ISO_10646__) || defined(__APPLE__) /* * Nothing to do for it since wchar_t on these platforms * is really Unicode. */ (void)sc; /* UNUSED */ #else if ((sc->flag & SCONV_UTF8_LIBARCHIVE_2) == 0) { sc->flag |= SCONV_UTF8_LIBARCHIVE_2; /* Set up string converters. */ setup_converter(sc); } #endif break; case SCONV_SET_OPT_NORMALIZATION_C: if ((sc->flag & SCONV_NORMALIZATION_C) == 0) { sc->flag |= SCONV_NORMALIZATION_C; sc->flag &= ~SCONV_NORMALIZATION_D; /* Set up string converters. */ setup_converter(sc); } break; case SCONV_SET_OPT_NORMALIZATION_D: #if defined(HAVE_ICONV) /* * If iconv will take the string, do not change the * setting of the normalization. */ if (!(sc->flag & SCONV_WIN_CP) && (sc->flag & (SCONV_FROM_UTF16 | SCONV_FROM_UTF8)) && !(sc->flag & (SCONV_TO_UTF16 | SCONV_TO_UTF8))) break; #endif if ((sc->flag & SCONV_NORMALIZATION_D) == 0) { sc->flag |= SCONV_NORMALIZATION_D; sc->flag &= ~SCONV_NORMALIZATION_C; /* Set up string converters. */ setup_converter(sc); } break; default: break; } } /* * * Copy one archive_string to another in locale conversion. * * archive_strncat_l(); * archive_strncpy_l(); * */ static size_t mbsnbytes(const void *_p, size_t n) { size_t s; const char *p, *pp; if (_p == NULL) return (0); p = (const char *)_p; /* Like strlen(p), except won't examine positions beyond p[n]. */ s = 0; pp = p; while (s < n && *pp) { pp++; s++; } return (s); } static size_t utf16nbytes(const void *_p, size_t n) { size_t s; const char *p, *pp; if (_p == NULL) return (0); p = (const char *)_p; /* Like strlen(p), except won't examine positions beyond p[n]. */ s = 0; pp = p; n >>= 1; while (s < n && (pp[0] || pp[1])) { pp += 2; s++; } return (s<<1); } int archive_strncpy_l(struct archive_string *as, const void *_p, size_t n, struct archive_string_conv *sc) { as->length = 0; return (archive_strncat_l(as, _p, n, sc)); } int archive_strncat_l(struct archive_string *as, const void *_p, size_t n, struct archive_string_conv *sc) { const void *s; size_t length = 0; int i, r = 0, r2; if (_p != NULL && n > 0) { if (sc != NULL && (sc->flag & SCONV_FROM_UTF16)) length = utf16nbytes(_p, n); else length = mbsnbytes(_p, n); } /* We must allocate memory even if there is no data for conversion * or copy. This simulates archive_string_append behavior. */ if (length == 0) { int tn = 1; if (sc != NULL && (sc->flag & SCONV_TO_UTF16)) tn = 2; if (archive_string_ensure(as, as->length + tn) == NULL) return (-1); as->s[as->length] = 0; if (tn == 2) as->s[as->length+1] = 0; return (0); } /* * If sc is NULL, we just make a copy. */ if (sc == NULL) { if (archive_string_append(as, _p, length) == NULL) return (-1);/* No memory */ return (0); } s = _p; i = 0; if (sc->nconverter > 1) { sc->utftmp.length = 0; r2 = sc->converter[0](&(sc->utftmp), s, length, sc); if (r2 != 0 && errno == ENOMEM) return (r2); if (r > r2) r = r2; s = sc->utftmp.s; length = sc->utftmp.length; ++i; } r2 = sc->converter[i](as, s, length, sc); if (r > r2) r = r2; return (r); } #if HAVE_ICONV /* * Return -1 if conversion fails. */ static int iconv_strncat_in_locale(struct archive_string *as, const void *_p, size_t length, struct archive_string_conv *sc) { ICONV_CONST char *itp; size_t remaining; iconv_t cd; char *outp; size_t avail, bs; int return_value = 0; /* success */ int to_size, from_size; if (sc->flag & SCONV_TO_UTF16) to_size = 2; else to_size = 1; if (sc->flag & SCONV_FROM_UTF16) from_size = 2; else from_size = 1; if (archive_string_ensure(as, as->length + length*2+to_size) == NULL) return (-1); cd = sc->cd; itp = (char *)(uintptr_t)_p; remaining = length; outp = as->s + as->length; avail = as->buffer_length - as->length - to_size; while (remaining >= (size_t)from_size) { size_t result = iconv(cd, &itp, &remaining, &outp, &avail); if (result != (size_t)-1) break; /* Conversion completed. */ if (errno == EILSEQ || errno == EINVAL) { /* * If an output charset is UTF-8 or UTF-16BE/LE, * unknown character should be U+FFFD * (replacement character). */ if (sc->flag & (SCONV_TO_UTF8 | SCONV_TO_UTF16)) { size_t rbytes; if (sc->flag & SCONV_TO_UTF8) rbytes = sizeof(utf8_replacement_char); else rbytes = 2; if (avail < rbytes) { as->length = outp - as->s; bs = as->buffer_length + (remaining * to_size) + rbytes; if (NULL == archive_string_ensure(as, bs)) return (-1); outp = as->s + as->length; avail = as->buffer_length - as->length - to_size; } if (sc->flag & SCONV_TO_UTF8) memcpy(outp, utf8_replacement_char, sizeof(utf8_replacement_char)); else if (sc->flag & SCONV_TO_UTF16BE) archive_be16enc(outp, UNICODE_R_CHAR); else archive_le16enc(outp, UNICODE_R_CHAR); outp += rbytes; avail -= rbytes; } else { /* Skip the illegal input bytes. */ *outp++ = '?'; avail--; } itp += from_size; remaining -= from_size; return_value = -1; /* failure */ } else { /* E2BIG no output buffer, * Increase an output buffer. */ as->length = outp - as->s; bs = as->buffer_length + remaining * 2; if (NULL == archive_string_ensure(as, bs)) return (-1); outp = as->s + as->length; avail = as->buffer_length - as->length - to_size; } } as->length = outp - as->s; as->s[as->length] = 0; if (to_size == 2) as->s[as->length+1] = 0; return (return_value); } #endif /* HAVE_ICONV */ #if defined(_WIN32) && !defined(__CYGWIN__) /* * Translate a string from a some CodePage to an another CodePage by * Windows APIs, and copy the result. Return -1 if conversion fails. */ static int strncat_in_codepage(struct archive_string *as, const void *_p, size_t length, struct archive_string_conv *sc) { const char *s = (const char *)_p; struct archive_wstring aws; size_t l; int r, saved_flag; archive_string_init(&aws); saved_flag = sc->flag; sc->flag &= ~(SCONV_NORMALIZATION_D | SCONV_NORMALIZATION_C); r = archive_wstring_append_from_mbs_in_codepage(&aws, s, length, sc); sc->flag = saved_flag; if (r != 0) { archive_wstring_free(&aws); if (errno != ENOMEM) archive_string_append(as, s, length); return (-1); } l = as->length; r = archive_string_append_from_wcs_in_codepage( as, aws.s, aws.length, sc); if (r != 0 && errno != ENOMEM && l == as->length) archive_string_append(as, s, length); archive_wstring_free(&aws); return (r); } /* * Test whether MBS ==> WCS is okay. */ static int invalid_mbs(const void *_p, size_t n, struct archive_string_conv *sc) { const char *p = (const char *)_p; unsigned codepage; DWORD mbflag = MB_ERR_INVALID_CHARS; if (sc->flag & SCONV_FROM_CHARSET) codepage = sc->to_cp; else codepage = sc->from_cp; if (codepage == CP_C_LOCALE) return (0); if (codepage != CP_UTF8) mbflag |= MB_PRECOMPOSED; if (MultiByteToWideChar(codepage, mbflag, p, (int)n, NULL, 0) == 0) return (-1); /* Invalid */ return (0); /* Okay */ } #else /* * Test whether MBS ==> WCS is okay. */ static int invalid_mbs(const void *_p, size_t n, struct archive_string_conv *sc) { const char *p = (const char *)_p; size_t r; #if HAVE_MBRTOWC mbstate_t shift_state; memset(&shift_state, 0, sizeof(shift_state)); #else /* Clear the shift state before starting. */ mbtowc(NULL, NULL, 0); #endif while (n) { wchar_t wc; #if HAVE_MBRTOWC r = mbrtowc(&wc, p, n, &shift_state); #else r = mbtowc(&wc, p, n); #endif if (r == (size_t)-1 || r == (size_t)-2) return (-1);/* Invalid. */ if (r == 0) break; p += r; n -= r; } (void)sc; /* UNUSED */ return (0); /* All Okey. */ } #endif /* defined(_WIN32) && !defined(__CYGWIN__) */ /* * Basically returns -1 because we cannot make a conversion of charset * without iconv but in some cases this would return 0. * Returns 0 if all copied characters are ASCII. * Returns 0 if both from-locale and to-locale are the same and those * can be WCS with no error. */ static int best_effort_strncat_in_locale(struct archive_string *as, const void *_p, size_t length, struct archive_string_conv *sc) { size_t remaining; const uint8_t *itp; int return_value = 0; /* success */ /* * If both from-locale and to-locale is the same, this makes a copy. * And then this checks all copied MBS can be WCS if so returns 0. */ if (sc->same) { if (archive_string_append(as, _p, length) == NULL) return (-1);/* No memory */ return (invalid_mbs(_p, length, sc)); } /* * If a character is ASCII, this just copies it. If not, this * assigns '?' character instead but in UTF-8 locale this assigns * byte sequence 0xEF 0xBD 0xBD, which are code point U+FFFD, * a Replacement Character in Unicode. */ remaining = length; itp = (const uint8_t *)_p; while (*itp && remaining > 0) { if (*itp > 127) { // Non-ASCII: Substitute with suitable replacement if (sc->flag & SCONV_TO_UTF8) { if (archive_string_append(as, utf8_replacement_char, sizeof(utf8_replacement_char)) == NULL) { __archive_errx(1, "Out of memory"); } } else { archive_strappend_char(as, '?'); } return_value = -1; } else { archive_strappend_char(as, *itp); } ++itp; } return (return_value); } /* * Unicode conversion functions. * - UTF-8 <===> UTF-8 in removing surrogate pairs. * - UTF-8 NFD ===> UTF-8 NFC in removing surrogate pairs. * - UTF-8 made by libarchive 2.x ===> UTF-8. * - UTF-16BE <===> UTF-8. * */ /* * Utility to convert a single UTF-8 sequence. * * Usually return used bytes, return used byte in negative value when * a unicode character is replaced with U+FFFD. * See also http://unicode.org/review/pr-121.html Public Review Issue #121 * Recommended Practice for Replacement Characters. */ static int _utf8_to_unicode(uint32_t *pwc, const char *s, size_t n) { static const char utf8_count[256] = { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,/* 00 - 0F */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,/* 10 - 1F */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,/* 20 - 2F */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,/* 30 - 3F */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,/* 40 - 4F */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,/* 50 - 5F */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,/* 60 - 6F */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,/* 70 - 7F */ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,/* 80 - 8F */ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,/* 90 - 9F */ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,/* A0 - AF */ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,/* B0 - BF */ 0, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,/* C0 - CF */ 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,/* D0 - DF */ 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,/* E0 - EF */ 4, 4, 4, 4, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 /* F0 - FF */ }; int ch, i; int cnt; uint32_t wc; /* Sanity check. */ if (n == 0) return (0); /* * Decode 1-4 bytes depending on the value of the first byte. */ ch = (unsigned char)*s; if (ch == 0) return (0); /* Standard: return 0 for end-of-string. */ cnt = utf8_count[ch]; /* Invalid sequence or there are not plenty bytes. */ if ((int)n < cnt) { cnt = (int)n; for (i = 1; i < cnt; i++) { if ((s[i] & 0xc0) != 0x80) { cnt = i; break; } } goto invalid_sequence; } /* Make a Unicode code point from a single UTF-8 sequence. */ switch (cnt) { case 1: /* 1 byte sequence. */ *pwc = ch & 0x7f; return (cnt); case 2: /* 2 bytes sequence. */ if ((s[1] & 0xc0) != 0x80) { cnt = 1; goto invalid_sequence; } *pwc = ((ch & 0x1f) << 6) | (s[1] & 0x3f); return (cnt); case 3: /* 3 bytes sequence. */ if ((s[1] & 0xc0) != 0x80) { cnt = 1; goto invalid_sequence; } if ((s[2] & 0xc0) != 0x80) { cnt = 2; goto invalid_sequence; } wc = ((ch & 0x0f) << 12) | ((s[1] & 0x3f) << 6) | (s[2] & 0x3f); if (wc < 0x800) goto invalid_sequence;/* Overlong sequence. */ break; case 4: /* 4 bytes sequence. */ if ((s[1] & 0xc0) != 0x80) { cnt = 1; goto invalid_sequence; } if ((s[2] & 0xc0) != 0x80) { cnt = 2; goto invalid_sequence; } if ((s[3] & 0xc0) != 0x80) { cnt = 3; goto invalid_sequence; } wc = ((ch & 0x07) << 18) | ((s[1] & 0x3f) << 12) | ((s[2] & 0x3f) << 6) | (s[3] & 0x3f); if (wc < 0x10000) goto invalid_sequence;/* Overlong sequence. */ break; default: /* Others are all invalid sequence. */ if (ch == 0xc0 || ch == 0xc1) cnt = 2; else if (ch >= 0xf5 && ch <= 0xf7) cnt = 4; else if (ch >= 0xf8 && ch <= 0xfb) cnt = 5; else if (ch == 0xfc || ch == 0xfd) cnt = 6; else cnt = 1; if ((int)n < cnt) cnt = (int)n; for (i = 1; i < cnt; i++) { if ((s[i] & 0xc0) != 0x80) { cnt = i; break; } } goto invalid_sequence; } /* The code point larger than 0x10FFFF is not legal * Unicode values. */ if (wc > UNICODE_MAX) goto invalid_sequence; /* Correctly gets a Unicode, returns used bytes. */ *pwc = wc; return (cnt); invalid_sequence: *pwc = UNICODE_R_CHAR;/* set the Replacement Character instead. */ return (cnt * -1); } static int utf8_to_unicode(uint32_t *pwc, const char *s, size_t n) { int cnt; cnt = _utf8_to_unicode(pwc, s, n); /* Any of Surrogate pair is not legal Unicode values. */ if (cnt == 3 && IS_SURROGATE_PAIR_LA(*pwc)) return (-3); return (cnt); } static inline uint32_t combine_surrogate_pair(uint32_t uc, uint32_t uc2) { uc -= 0xD800; uc *= 0x400; uc += uc2 - 0xDC00; uc += 0x10000; return (uc); } /* * Convert a single UTF-8/CESU-8 sequence to a Unicode code point in * removing surrogate pairs. * * CESU-8: The Compatibility Encoding Scheme for UTF-16. * * Usually return used bytes, return used byte in negative value when * a unicode character is replaced with U+FFFD. */ static int cesu8_to_unicode(uint32_t *pwc, const char *s, size_t n) { uint32_t wc = 0; int cnt; cnt = _utf8_to_unicode(&wc, s, n); if (cnt == 3 && IS_HIGH_SURROGATE_LA(wc)) { uint32_t wc2 = 0; if (n - 3 < 3) { /* Invalid byte sequence. */ goto invalid_sequence; } cnt = _utf8_to_unicode(&wc2, s+3, n-3); if (cnt != 3 || !IS_LOW_SURROGATE_LA(wc2)) { /* Invalid byte sequence. */ goto invalid_sequence; } wc = combine_surrogate_pair(wc, wc2); cnt = 6; } else if (cnt == 3 && IS_LOW_SURROGATE_LA(wc)) { /* Invalid byte sequence. */ goto invalid_sequence; } *pwc = wc; return (cnt); invalid_sequence: *pwc = UNICODE_R_CHAR;/* set the Replacement Character instead. */ if (cnt > 0) cnt *= -1; return (cnt); } /* * Convert a Unicode code point to a single UTF-8 sequence. * * NOTE:This function does not check if the Unicode is legal or not. * Please you definitely check it before calling this. */ static size_t unicode_to_utf8(char *p, size_t remaining, uint32_t uc) { char *_p = p; /* Invalid Unicode char maps to Replacement character */ if (uc > UNICODE_MAX) uc = UNICODE_R_CHAR; /* Translate code point to UTF8 */ if (uc <= 0x7f) { if (remaining == 0) return (0); *p++ = (char)uc; } else if (uc <= 0x7ff) { if (remaining < 2) return (0); *p++ = 0xc0 | ((uc >> 6) & 0x1f); *p++ = 0x80 | (uc & 0x3f); } else if (uc <= 0xffff) { if (remaining < 3) return (0); *p++ = 0xe0 | ((uc >> 12) & 0x0f); *p++ = 0x80 | ((uc >> 6) & 0x3f); *p++ = 0x80 | (uc & 0x3f); } else { if (remaining < 4) return (0); *p++ = 0xf0 | ((uc >> 18) & 0x07); *p++ = 0x80 | ((uc >> 12) & 0x3f); *p++ = 0x80 | ((uc >> 6) & 0x3f); *p++ = 0x80 | (uc & 0x3f); } return (p - _p); } static int utf16be_to_unicode(uint32_t *pwc, const char *s, size_t n) { return (utf16_to_unicode(pwc, s, n, 1)); } static int utf16le_to_unicode(uint32_t *pwc, const char *s, size_t n) { return (utf16_to_unicode(pwc, s, n, 0)); } static int utf16_to_unicode(uint32_t *pwc, const char *s, size_t n, int be) { const char *utf16 = s; unsigned uc; if (n == 0) return (0); if (n == 1) { /* set the Replacement Character instead. */ *pwc = UNICODE_R_CHAR; return (-1); } if (be) uc = archive_be16dec(utf16); else uc = archive_le16dec(utf16); utf16 += 2; /* If this is a surrogate pair, assemble the full code point.*/ if (IS_HIGH_SURROGATE_LA(uc)) { unsigned uc2; if (n >= 4) { if (be) uc2 = archive_be16dec(utf16); else uc2 = archive_le16dec(utf16); } else uc2 = 0; if (IS_LOW_SURROGATE_LA(uc2)) { uc = combine_surrogate_pair(uc, uc2); utf16 += 2; } else { /* Undescribed code point should be U+FFFD * (replacement character). */ *pwc = UNICODE_R_CHAR; return (-2); } } /* * Surrogate pair values(0xd800 through 0xdfff) are only * used by UTF-16, so, after above calculation, the code * must not be surrogate values, and Unicode has no codes * larger than 0x10ffff. Thus, those are not legal Unicode * values. */ if (IS_SURROGATE_PAIR_LA(uc) || uc > UNICODE_MAX) { /* Undescribed code point should be U+FFFD * (replacement character). */ *pwc = UNICODE_R_CHAR; return (((int)(utf16 - s)) * -1); } *pwc = uc; return ((int)(utf16 - s)); } static size_t unicode_to_utf16be(char *p, size_t remaining, uint32_t uc) { char *utf16 = p; if (uc > 0xffff) { /* We have a code point that won't fit into a * wchar_t; convert it to a surrogate pair. */ if (remaining < 4) return (0); uc -= 0x10000; archive_be16enc(utf16, ((uc >> 10) & 0x3ff) + 0xD800); archive_be16enc(utf16+2, (uc & 0x3ff) + 0xDC00); return (4); } else { if (remaining < 2) return (0); archive_be16enc(utf16, uc); return (2); } } static size_t unicode_to_utf16le(char *p, size_t remaining, uint32_t uc) { char *utf16 = p; if (uc > 0xffff) { /* We have a code point that won't fit into a * wchar_t; convert it to a surrogate pair. */ if (remaining < 4) return (0); uc -= 0x10000; archive_le16enc(utf16, ((uc >> 10) & 0x3ff) + 0xD800); archive_le16enc(utf16+2, (uc & 0x3ff) + 0xDC00); return (4); } else { if (remaining < 2) return (0); archive_le16enc(utf16, uc); return (2); } } /* * Copy UTF-8 string in checking surrogate pair. * If any surrogate pair are found, it would be canonicalized. */ static int strncat_from_utf8_to_utf8(struct archive_string *as, const void *_p, size_t len, struct archive_string_conv *sc) { const char *s; char *p, *endp; int n, ret = 0; (void)sc; /* UNUSED */ if (archive_string_ensure(as, as->length + len + 1) == NULL) return (-1); s = (const char *)_p; p = as->s + as->length; endp = as->s + as->buffer_length -1; do { uint32_t uc; const char *ss = s; size_t w; /* * Forward byte sequence until a conversion of that is needed. */ while ((n = utf8_to_unicode(&uc, s, len)) > 0) { s += n; len -= n; } if (ss < s) { if (p + (s - ss) > endp) { as->length = p - as->s; if (archive_string_ensure(as, as->buffer_length + len + 1) == NULL) return (-1); p = as->s + as->length; endp = as->s + as->buffer_length -1; } memcpy(p, ss, s - ss); p += s - ss; } /* * If n is negative, current byte sequence needs a replacement. */ if (n < 0) { if (n == -3 && IS_SURROGATE_PAIR_LA(uc)) { /* Current byte sequence may be CESU-8. */ n = cesu8_to_unicode(&uc, s, len); } if (n < 0) { ret = -1; n *= -1;/* Use a replaced unicode character. */ } /* Rebuild UTF-8 byte sequence. */ while ((w = unicode_to_utf8(p, endp - p, uc)) == 0) { as->length = p - as->s; if (archive_string_ensure(as, as->buffer_length + len + 1) == NULL) return (-1); p = as->s + as->length; endp = as->s + as->buffer_length -1; } p += w; s += n; len -= n; } } while (n > 0); as->length = p - as->s; as->s[as->length] = '\0'; return (ret); } static int archive_string_append_unicode(struct archive_string *as, const void *_p, size_t len, struct archive_string_conv *sc) { const char *s; char *p, *endp; uint32_t uc; size_t w; int n, ret = 0, ts, tm; int (*parse)(uint32_t *, const char *, size_t); size_t (*unparse)(char *, size_t, uint32_t); if (sc->flag & SCONV_TO_UTF16BE) { unparse = unicode_to_utf16be; ts = 2; } else if (sc->flag & SCONV_TO_UTF16LE) { unparse = unicode_to_utf16le; ts = 2; } else if (sc->flag & SCONV_TO_UTF8) { unparse = unicode_to_utf8; ts = 1; } else { /* * This case is going to be converted to another * character-set through iconv. */ if (sc->flag & SCONV_FROM_UTF16BE) { unparse = unicode_to_utf16be; ts = 2; } else if (sc->flag & SCONV_FROM_UTF16LE) { unparse = unicode_to_utf16le; ts = 2; } else { unparse = unicode_to_utf8; ts = 1; } } if (sc->flag & SCONV_FROM_UTF16BE) { parse = utf16be_to_unicode; tm = 1; } else if (sc->flag & SCONV_FROM_UTF16LE) { parse = utf16le_to_unicode; tm = 1; } else { parse = cesu8_to_unicode; tm = ts; } if (archive_string_ensure(as, as->length + len * tm + ts) == NULL) return (-1); s = (const char *)_p; p = as->s + as->length; endp = as->s + as->buffer_length - ts; while ((n = parse(&uc, s, len)) != 0) { if (n < 0) { /* Use a replaced unicode character. */ n *= -1; ret = -1; } s += n; len -= n; while ((w = unparse(p, endp - p, uc)) == 0) { /* There is not enough output buffer so * we have to expand it. */ as->length = p - as->s; if (archive_string_ensure(as, as->buffer_length + len * tm + ts) == NULL) return (-1); p = as->s + as->length; endp = as->s + as->buffer_length - ts; } p += w; } as->length = p - as->s; as->s[as->length] = '\0'; if (ts == 2) as->s[as->length+1] = '\0'; return (ret); } /* * Following Constants for Hangul compositions this information comes from * Unicode Standard Annex #15 http://unicode.org/reports/tr15/ */ #define HC_SBASE 0xAC00 #define HC_LBASE 0x1100 #define HC_VBASE 0x1161 #define HC_TBASE 0x11A7 #define HC_LCOUNT 19 #define HC_VCOUNT 21 #define HC_TCOUNT 28 #define HC_NCOUNT (HC_VCOUNT * HC_TCOUNT) #define HC_SCOUNT (HC_LCOUNT * HC_NCOUNT) static uint32_t get_nfc(uint32_t uc, uint32_t uc2) { int t, b; t = 0; b = sizeof(u_composition_table)/sizeof(u_composition_table[0]) -1; while (b >= t) { int m = (t + b) / 2; if (u_composition_table[m].cp1 < uc) t = m + 1; else if (u_composition_table[m].cp1 > uc) b = m - 1; else if (u_composition_table[m].cp2 < uc2) t = m + 1; else if (u_composition_table[m].cp2 > uc2) b = m - 1; else return (u_composition_table[m].nfc); } return (0); } #define FDC_MAX 10 /* The maximum number of Following Decomposable * Characters. */ /* * Update first code point. */ #define UPDATE_UC(new_uc) do { \ uc = new_uc; \ ucptr = NULL; \ } while (0) /* * Replace first code point with second code point. */ #define REPLACE_UC_WITH_UC2() do { \ uc = uc2; \ ucptr = uc2ptr; \ n = n2; \ } while (0) #define EXPAND_BUFFER() do { \ as->length = p - as->s; \ if (archive_string_ensure(as, \ as->buffer_length + len * tm + ts) == NULL)\ return (-1); \ p = as->s + as->length; \ endp = as->s + as->buffer_length - ts; \ } while (0) #define UNPARSE(p, endp, uc) do { \ while ((w = unparse(p, (endp) - (p), uc)) == 0) {\ EXPAND_BUFFER(); \ } \ p += w; \ } while (0) /* * Write first code point. * If the code point has not be changed from its original code, * this just copies it from its original buffer pointer. * If not, this converts it to UTF-8 byte sequence and copies it. */ #define WRITE_UC() do { \ if (ucptr) { \ if (p + n > endp) \ EXPAND_BUFFER(); \ switch (n) { \ case 4: \ *p++ = *ucptr++; \ /* FALL THROUGH */ \ case 3: \ *p++ = *ucptr++; \ /* FALL THROUGH */ \ case 2: \ *p++ = *ucptr++; \ /* FALL THROUGH */ \ case 1: \ *p++ = *ucptr; \ break; \ } \ ucptr = NULL; \ } else { \ UNPARSE(p, endp, uc); \ } \ } while (0) /* * Collect following decomposable code points. */ #define COLLECT_CPS(start) do { \ int _i; \ for (_i = start; _i < FDC_MAX ; _i++) { \ nx = parse(&ucx[_i], s, len); \ if (nx <= 0) \ break; \ cx = CCC(ucx[_i]); \ if (cl >= cx && cl != 228 && cx != 228)\ break; \ s += nx; \ len -= nx; \ cl = cx; \ ccx[_i] = cx; \ } \ if (_i >= FDC_MAX) { \ ret = -1; \ ucx_size = FDC_MAX; \ } else \ ucx_size = _i; \ } while (0) /* * Normalize UTF-8/UTF-16BE characters to Form C and copy the result. * * TODO: Convert composition exclusions, which are never converted * from NFC,NFD,NFKC and NFKD, to Form C. */ static int archive_string_normalize_C(struct archive_string *as, const void *_p, size_t len, struct archive_string_conv *sc) { const char *s = (const char *)_p; char *p, *endp; uint32_t uc, uc2; size_t w; int always_replace, n, n2, ret = 0, spair, ts, tm; int (*parse)(uint32_t *, const char *, size_t); size_t (*unparse)(char *, size_t, uint32_t); always_replace = 1; ts = 1;/* text size. */ if (sc->flag & SCONV_TO_UTF16BE) { unparse = unicode_to_utf16be; ts = 2; if (sc->flag & SCONV_FROM_UTF16BE) always_replace = 0; } else if (sc->flag & SCONV_TO_UTF16LE) { unparse = unicode_to_utf16le; ts = 2; if (sc->flag & SCONV_FROM_UTF16LE) always_replace = 0; } else if (sc->flag & SCONV_TO_UTF8) { unparse = unicode_to_utf8; if (sc->flag & SCONV_FROM_UTF8) always_replace = 0; } else { /* * This case is going to be converted to another * character-set through iconv. */ always_replace = 0; if (sc->flag & SCONV_FROM_UTF16BE) { unparse = unicode_to_utf16be; ts = 2; } else if (sc->flag & SCONV_FROM_UTF16LE) { unparse = unicode_to_utf16le; ts = 2; } else { unparse = unicode_to_utf8; } } if (sc->flag & SCONV_FROM_UTF16BE) { parse = utf16be_to_unicode; tm = 1; spair = 4;/* surrogate pair size in UTF-16. */ } else if (sc->flag & SCONV_FROM_UTF16LE) { parse = utf16le_to_unicode; tm = 1; spair = 4;/* surrogate pair size in UTF-16. */ } else { parse = cesu8_to_unicode; tm = ts; spair = 6;/* surrogate pair size in UTF-8. */ } if (archive_string_ensure(as, as->length + len * tm + ts) == NULL) return (-1); p = as->s + as->length; endp = as->s + as->buffer_length - ts; while ((n = parse(&uc, s, len)) != 0) { const char *ucptr, *uc2ptr; if (n < 0) { /* Use a replaced unicode character. */ UNPARSE(p, endp, uc); s += n*-1; len -= n*-1; ret = -1; continue; } else if (n == spair || always_replace) /* uc is converted from a surrogate pair. * this should be treated as a changed code. */ ucptr = NULL; else ucptr = s; s += n; len -= n; /* Read second code point. */ while ((n2 = parse(&uc2, s, len)) > 0) { uint32_t ucx[FDC_MAX]; int ccx[FDC_MAX]; int cl, cx, i, nx, ucx_size; int LIndex,SIndex; uint32_t nfc; if (n2 == spair || always_replace) /* uc2 is converted from a surrogate pair. * this should be treated as a changed code. */ uc2ptr = NULL; else uc2ptr = s; s += n2; len -= n2; /* * If current second code point is out of decomposable * code points, finding compositions is unneeded. */ if (!IS_DECOMPOSABLE_BLOCK(uc2)) { WRITE_UC(); REPLACE_UC_WITH_UC2(); continue; } /* * Try to combine current code points. */ /* * We have to combine Hangul characters according to * http://uniicode.org/reports/tr15/#Hangul */ if (0 <= (LIndex = uc - HC_LBASE) && LIndex < HC_LCOUNT) { /* * Hangul Composition. * 1. Two current code points are L and V. */ int VIndex = uc2 - HC_VBASE; if (0 <= VIndex && VIndex < HC_VCOUNT) { /* Make syllable of form LV. */ UPDATE_UC(HC_SBASE + (LIndex * HC_VCOUNT + VIndex) * HC_TCOUNT); } else { WRITE_UC(); REPLACE_UC_WITH_UC2(); } continue; } else if (0 <= (SIndex = uc - HC_SBASE) && SIndex < HC_SCOUNT && (SIndex % HC_TCOUNT) == 0) { /* * Hangul Composition. * 2. Two current code points are LV and T. */ int TIndex = uc2 - HC_TBASE; if (0 < TIndex && TIndex < HC_TCOUNT) { /* Make syllable of form LVT. */ UPDATE_UC(uc + TIndex); } else { WRITE_UC(); REPLACE_UC_WITH_UC2(); } continue; } else if ((nfc = get_nfc(uc, uc2)) != 0) { /* A composition to current code points * is found. */ UPDATE_UC(nfc); continue; } else if ((cl = CCC(uc2)) == 0) { /* Clearly 'uc2' the second code point is not * a decomposable code. */ WRITE_UC(); REPLACE_UC_WITH_UC2(); continue; } /* * Collect following decomposable code points. */ cx = 0; ucx[0] = uc2; ccx[0] = cl; COLLECT_CPS(1); /* * Find a composed code in the collected code points. */ i = 1; while (i < ucx_size) { int j; if ((nfc = get_nfc(uc, ucx[i])) == 0) { i++; continue; } /* * nfc is composed of uc and ucx[i]. */ UPDATE_UC(nfc); /* * Remove ucx[i] by shifting * following code points. */ for (j = i; j+1 < ucx_size; j++) { ucx[j] = ucx[j+1]; ccx[j] = ccx[j+1]; } ucx_size --; /* * Collect following code points blocked * by ucx[i] the removed code point. */ if (ucx_size > 0 && i == ucx_size && nx > 0 && cx == cl) { cl = ccx[ucx_size-1]; COLLECT_CPS(ucx_size); } /* * Restart finding a composed code with * the updated uc from the top of the * collected code points. */ i = 0; } /* * Apparently the current code points are not * decomposed characters or already composed. */ WRITE_UC(); for (i = 0; i < ucx_size; i++) UNPARSE(p, endp, ucx[i]); /* * Flush out remaining canonical combining characters. */ if (nx > 0 && cx == cl && len > 0) { while ((nx = parse(&ucx[0], s, len)) > 0) { cx = CCC(ucx[0]); if (cl > cx) break; s += nx; len -= nx; cl = cx; UNPARSE(p, endp, ucx[0]); } } break; } if (n2 < 0) { WRITE_UC(); /* Use a replaced unicode character. */ UNPARSE(p, endp, uc2); s += n2*-1; len -= n2*-1; ret = -1; continue; } else if (n2 == 0) { WRITE_UC(); break; } } as->length = p - as->s; as->s[as->length] = '\0'; if (ts == 2) as->s[as->length+1] = '\0'; return (ret); } static int get_nfd(uint32_t *cp1, uint32_t *cp2, uint32_t uc) { int t, b; /* * These are not converted to NFD on Mac OS. */ if ((uc >= 0x2000 && uc <= 0x2FFF) || (uc >= 0xF900 && uc <= 0xFAFF) || (uc >= 0x2F800 && uc <= 0x2FAFF)) return (0); /* * Those code points are not converted to NFD on Mac OS. * I do not know the reason because it is undocumented. * NFC NFD * 1109A ==> 11099 110BA * 1109C ==> 1109B 110BA * 110AB ==> 110A5 110BA */ if (uc == 0x1109A || uc == 0x1109C || uc == 0x110AB) return (0); t = 0; b = sizeof(u_decomposition_table)/sizeof(u_decomposition_table[0]) -1; while (b >= t) { int m = (t + b) / 2; if (u_decomposition_table[m].nfc < uc) t = m + 1; else if (u_decomposition_table[m].nfc > uc) b = m - 1; else { *cp1 = u_decomposition_table[m].cp1; *cp2 = u_decomposition_table[m].cp2; return (1); } } return (0); } #define REPLACE_UC_WITH(cp) do { \ uc = cp; \ ucptr = NULL; \ } while (0) /* * Normalize UTF-8 characters to Form D and copy the result. */ static int archive_string_normalize_D(struct archive_string *as, const void *_p, size_t len, struct archive_string_conv *sc) { const char *s = (const char *)_p; char *p, *endp; uint32_t uc, uc2; size_t w; int always_replace, n, n2, ret = 0, spair, ts, tm; int (*parse)(uint32_t *, const char *, size_t); size_t (*unparse)(char *, size_t, uint32_t); always_replace = 1; ts = 1;/* text size. */ if (sc->flag & SCONV_TO_UTF16BE) { unparse = unicode_to_utf16be; ts = 2; if (sc->flag & SCONV_FROM_UTF16BE) always_replace = 0; } else if (sc->flag & SCONV_TO_UTF16LE) { unparse = unicode_to_utf16le; ts = 2; if (sc->flag & SCONV_FROM_UTF16LE) always_replace = 0; } else if (sc->flag & SCONV_TO_UTF8) { unparse = unicode_to_utf8; if (sc->flag & SCONV_FROM_UTF8) always_replace = 0; } else { /* * This case is going to be converted to another * character-set through iconv. */ always_replace = 0; if (sc->flag & SCONV_FROM_UTF16BE) { unparse = unicode_to_utf16be; ts = 2; } else if (sc->flag & SCONV_FROM_UTF16LE) { unparse = unicode_to_utf16le; ts = 2; } else { unparse = unicode_to_utf8; } } if (sc->flag & SCONV_FROM_UTF16BE) { parse = utf16be_to_unicode; tm = 1; spair = 4;/* surrogate pair size in UTF-16. */ } else if (sc->flag & SCONV_FROM_UTF16LE) { parse = utf16le_to_unicode; tm = 1; spair = 4;/* surrogate pair size in UTF-16. */ } else { parse = cesu8_to_unicode; tm = ts; spair = 6;/* surrogate pair size in UTF-8. */ } if (archive_string_ensure(as, as->length + len * tm + ts) == NULL) return (-1); p = as->s + as->length; endp = as->s + as->buffer_length - ts; while ((n = parse(&uc, s, len)) != 0) { const char *ucptr; uint32_t cp1, cp2; int SIndex; struct { uint32_t uc; int ccc; } fdc[FDC_MAX]; int fdi, fdj; int ccc; check_first_code: if (n < 0) { /* Use a replaced unicode character. */ UNPARSE(p, endp, uc); s += n*-1; len -= n*-1; ret = -1; continue; } else if (n == spair || always_replace) /* uc is converted from a surrogate pair. * this should be treated as a changed code. */ ucptr = NULL; else ucptr = s; s += n; len -= n; /* Hangul Decomposition. */ if ((SIndex = uc - HC_SBASE) >= 0 && SIndex < HC_SCOUNT) { int L = HC_LBASE + SIndex / HC_NCOUNT; int V = HC_VBASE + (SIndex % HC_NCOUNT) / HC_TCOUNT; int T = HC_TBASE + SIndex % HC_TCOUNT; REPLACE_UC_WITH(L); WRITE_UC(); REPLACE_UC_WITH(V); WRITE_UC(); if (T != HC_TBASE) { REPLACE_UC_WITH(T); WRITE_UC(); } continue; } if (IS_DECOMPOSABLE_BLOCK(uc) && CCC(uc) != 0) { WRITE_UC(); continue; } fdi = 0; while (get_nfd(&cp1, &cp2, uc) && fdi < FDC_MAX) { int k; for (k = fdi; k > 0; k--) fdc[k] = fdc[k-1]; fdc[0].ccc = CCC(cp2); fdc[0].uc = cp2; fdi++; REPLACE_UC_WITH(cp1); } /* Read following code points. */ while ((n2 = parse(&uc2, s, len)) > 0 && (ccc = CCC(uc2)) != 0 && fdi < FDC_MAX) { int j, k; s += n2; len -= n2; for (j = 0; j < fdi; j++) { if (fdc[j].ccc > ccc) break; } if (j < fdi) { for (k = fdi; k > j; k--) fdc[k] = fdc[k-1]; fdc[j].ccc = ccc; fdc[j].uc = uc2; } else { fdc[fdi].ccc = ccc; fdc[fdi].uc = uc2; } fdi++; } WRITE_UC(); for (fdj = 0; fdj < fdi; fdj++) { REPLACE_UC_WITH(fdc[fdj].uc); WRITE_UC(); } if (n2 == 0) break; REPLACE_UC_WITH(uc2); n = n2; goto check_first_code; } as->length = p - as->s; as->s[as->length] = '\0'; if (ts == 2) as->s[as->length+1] = '\0'; return (ret); } /* * libarchive 2.x made incorrect UTF-8 strings in the wrong assumption * that WCS is Unicode. It is true for several platforms but some are false. * And then people who did not use UTF-8 locale on the non Unicode WCS * platform and made a tar file with libarchive(mostly bsdtar) 2.x. Those * now cannot get right filename from libarchive 3.x and later since we * fixed the wrong assumption and it is incompatible to older its versions. * So we provide special option, "compat-2x.x", for resolving it. * That option enable the string conversion of libarchive 2.x. * * Translates the wrong UTF-8 string made by libarchive 2.x into current * locale character set and appends to the archive_string. * Note: returns -1 if conversion fails. */ static int strncat_from_utf8_libarchive2(struct archive_string *as, const void *_p, size_t len, struct archive_string_conv *sc) { const char *s; int n; char *p; char *end; uint32_t unicode; #if HAVE_WCRTOMB mbstate_t shift_state; memset(&shift_state, 0, sizeof(shift_state)); #else /* Clear the shift state before starting. */ wctomb(NULL, L'\0'); #endif (void)sc; /* UNUSED */ /* * Allocate buffer for MBS. * We need this allocation here since it is possible that * as->s is still NULL. */ if (archive_string_ensure(as, as->length + len + 1) == NULL) return (-1); s = (const char *)_p; p = as->s + as->length; end = as->s + as->buffer_length - MB_CUR_MAX -1; while ((n = _utf8_to_unicode(&unicode, s, len)) != 0) { wchar_t wc; if (p >= end) { as->length = p - as->s; /* Re-allocate buffer for MBS. */ if (archive_string_ensure(as, as->length + len * 2 + 1) == NULL) return (-1); p = as->s + as->length; end = as->s + as->buffer_length - MB_CUR_MAX -1; } /* * As libarchive 2.x, translates the UTF-8 characters into * wide-characters in the assumption that WCS is Unicode. */ if (n < 0) { n *= -1; wc = L'?'; } else wc = (wchar_t)unicode; s += n; len -= n; /* * Translates the wide-character into the current locale MBS. */ #if HAVE_WCRTOMB n = (int)wcrtomb(p, wc, &shift_state); #else n = (int)wctomb(p, wc); #endif if (n == -1) return (-1); p += n; } as->length = p - as->s; as->s[as->length] = '\0'; return (0); } /* * Conversion functions between current locale dependent MBS and UTF-16BE. * strncat_from_utf16be() : UTF-16BE --> MBS * strncat_to_utf16be() : MBS --> UTF16BE */ #if defined(_WIN32) && !defined(__CYGWIN__) /* * Convert a UTF-16BE/LE string to current locale and copy the result. * Return -1 if conversion fails. */ static int win_strncat_from_utf16(struct archive_string *as, const void *_p, size_t bytes, struct archive_string_conv *sc, int be) { struct archive_string tmp; const char *u16; int ll; BOOL defchar; char *mbs; size_t mbs_size, b; int ret = 0; bytes &= ~1; if (archive_string_ensure(as, as->length + bytes +1) == NULL) return (-1); mbs = as->s + as->length; mbs_size = as->buffer_length - as->length -1; if (sc->to_cp == CP_C_LOCALE) { /* * "C" locale special process. */ u16 = _p; ll = 0; for (b = 0; b < bytes; b += 2) { uint16_t val; if (be) val = archive_be16dec(u16+b); else val = archive_le16dec(u16+b); if (val > 255) { *mbs++ = '?'; ret = -1; } else *mbs++ = (char)(val&0xff); ll++; } as->length += ll; as->s[as->length] = '\0'; return (ret); } archive_string_init(&tmp); if (be) { if (is_big_endian()) { u16 = _p; } else { if (archive_string_ensure(&tmp, bytes+2) == NULL) return (-1); memcpy(tmp.s, _p, bytes); for (b = 0; b < bytes; b += 2) { uint16_t val = archive_be16dec(tmp.s+b); archive_le16enc(tmp.s+b, val); } u16 = tmp.s; } } else { if (!is_big_endian()) { u16 = _p; } else { if (archive_string_ensure(&tmp, bytes+2) == NULL) return (-1); memcpy(tmp.s, _p, bytes); for (b = 0; b < bytes; b += 2) { uint16_t val = archive_le16dec(tmp.s+b); archive_be16enc(tmp.s+b, val); } u16 = tmp.s; } } do { defchar = 0; ll = WideCharToMultiByte(sc->to_cp, 0, (LPCWSTR)u16, (int)bytes>>1, mbs, (int)mbs_size, NULL, &defchar); /* Exit loop if we succeeded */ if (ll != 0 || GetLastError() != ERROR_INSUFFICIENT_BUFFER) { break; } /* Else expand buffer and loop to try again. */ ll = WideCharToMultiByte(sc->to_cp, 0, (LPCWSTR)u16, (int)bytes, NULL, 0, NULL, NULL); if (archive_string_ensure(as, ll +1) == NULL) return (-1); mbs = as->s + as->length; mbs_size = as->buffer_length - as->length -1; } while (1); archive_string_free(&tmp); as->length += ll; as->s[as->length] = '\0'; if (ll == 0 || defchar) ret = -1; return (ret); } static int win_strncat_from_utf16be(struct archive_string *as, const void *_p, size_t bytes, struct archive_string_conv *sc) { return (win_strncat_from_utf16(as, _p, bytes, sc, 1)); } static int win_strncat_from_utf16le(struct archive_string *as, const void *_p, size_t bytes, struct archive_string_conv *sc) { return (win_strncat_from_utf16(as, _p, bytes, sc, 0)); } static int is_big_endian(void) { uint16_t d = 1; return (archive_be16dec(&d) == 1); } /* * Convert a current locale string to UTF-16BE/LE and copy the result. * Return -1 if conversion fails. */ static int win_strncat_to_utf16(struct archive_string *as16, const void *_p, size_t length, struct archive_string_conv *sc, int bigendian) { const char *s = (const char *)_p; char *u16; size_t count, avail; if (archive_string_ensure(as16, as16->length + (length + 1) * 2) == NULL) return (-1); u16 = as16->s + as16->length; avail = as16->buffer_length - 2; if (sc->from_cp == CP_C_LOCALE) { /* * "C" locale special process. */ count = 0; while (count < length && *s) { if (bigendian) archive_be16enc(u16, *s); else archive_le16enc(u16, *s); u16 += 2; s++; count++; } as16->length += count << 1; as16->s[as16->length] = 0; as16->s[as16->length+1] = 0; return (0); } do { count = MultiByteToWideChar(sc->from_cp, MB_PRECOMPOSED, s, (int)length, (LPWSTR)u16, (int)avail>>1); /* Exit loop if we succeeded */ if (count != 0 || GetLastError() != ERROR_INSUFFICIENT_BUFFER) { break; } /* Expand buffer and try again */ count = MultiByteToWideChar(sc->from_cp, MB_PRECOMPOSED, s, (int)length, NULL, 0); if (archive_string_ensure(as16, (count +1) * 2) == NULL) return (-1); u16 = as16->s + as16->length; avail = as16->buffer_length - 2; } while (1); as16->length += count * 2; as16->s[as16->length] = 0; as16->s[as16->length+1] = 0; if (count == 0) return (-1); if (is_big_endian()) { if (!bigendian) { while (count > 0) { uint16_t v = archive_be16dec(u16); archive_le16enc(u16, v); u16 += 2; count--; } } } else { if (bigendian) { while (count > 0) { uint16_t v = archive_le16dec(u16); archive_be16enc(u16, v); u16 += 2; count--; } } } return (0); } static int win_strncat_to_utf16be(struct archive_string *as16, const void *_p, size_t length, struct archive_string_conv *sc) { return (win_strncat_to_utf16(as16, _p, length, sc, 1)); } static int win_strncat_to_utf16le(struct archive_string *as16, const void *_p, size_t length, struct archive_string_conv *sc) { return (win_strncat_to_utf16(as16, _p, length, sc, 0)); } #endif /* _WIN32 && !__CYGWIN__ */ /* * Do the best effort for conversions. * We cannot handle UTF-16BE character-set without such iconv, * but there is a chance if a string consists just ASCII code or * a current locale is UTF-8. */ /* * Convert a UTF-16BE string to current locale and copy the result. * Return -1 if conversion fails. */ static int best_effort_strncat_from_utf16(struct archive_string *as, const void *_p, size_t bytes, struct archive_string_conv *sc, int be) { const char *utf16 = (const char *)_p; char *mbs; uint32_t uc; int n, ret; (void)sc; /* UNUSED */ /* * Other case, we should do the best effort. * If all character are ASCII(<0x7f), we can convert it. * if not , we set a alternative character and return -1. */ ret = 0; if (archive_string_ensure(as, as->length + bytes +1) == NULL) return (-1); mbs = as->s + as->length; while ((n = utf16_to_unicode(&uc, utf16, bytes, be)) != 0) { if (n < 0) { n *= -1; ret = -1; } bytes -= n; utf16 += n; if (uc > 127) { /* We cannot handle it. */ *mbs++ = '?'; ret = -1; } else *mbs++ = (char)uc; } as->length = mbs - as->s; as->s[as->length] = '\0'; return (ret); } static int best_effort_strncat_from_utf16be(struct archive_string *as, const void *_p, size_t bytes, struct archive_string_conv *sc) { return (best_effort_strncat_from_utf16(as, _p, bytes, sc, 1)); } static int best_effort_strncat_from_utf16le(struct archive_string *as, const void *_p, size_t bytes, struct archive_string_conv *sc) { return (best_effort_strncat_from_utf16(as, _p, bytes, sc, 0)); } /* * Convert a current locale string to UTF-16BE/LE and copy the result. * Return -1 if conversion fails. */ static int best_effort_strncat_to_utf16(struct archive_string *as16, const void *_p, size_t length, struct archive_string_conv *sc, int bigendian) { const char *s = (const char *)_p; char *utf16; size_t remaining; int ret; (void)sc; /* UNUSED */ /* * Other case, we should do the best effort. * If all character are ASCII(<0x7f), we can convert it. * if not , we set a alternative character and return -1. */ ret = 0; remaining = length; if (archive_string_ensure(as16, as16->length + (length + 1) * 2) == NULL) return (-1); utf16 = as16->s + as16->length; while (remaining--) { unsigned c = *s++; if (c > 127) { /* We cannot handle it. */ c = UNICODE_R_CHAR; ret = -1; } if (bigendian) archive_be16enc(utf16, c); else archive_le16enc(utf16, c); utf16 += 2; } as16->length = utf16 - as16->s; as16->s[as16->length] = 0; as16->s[as16->length+1] = 0; return (ret); } static int best_effort_strncat_to_utf16be(struct archive_string *as16, const void *_p, size_t length, struct archive_string_conv *sc) { return (best_effort_strncat_to_utf16(as16, _p, length, sc, 1)); } static int best_effort_strncat_to_utf16le(struct archive_string *as16, const void *_p, size_t length, struct archive_string_conv *sc) { return (best_effort_strncat_to_utf16(as16, _p, length, sc, 0)); } /* * Multistring operations. */ void archive_mstring_clean(struct archive_mstring *aes) { archive_wstring_free(&(aes->aes_wcs)); archive_string_free(&(aes->aes_mbs)); archive_string_free(&(aes->aes_utf8)); archive_string_free(&(aes->aes_mbs_in_locale)); aes->aes_set = 0; } void archive_mstring_copy(struct archive_mstring *dest, struct archive_mstring *src) { dest->aes_set = src->aes_set; archive_string_copy(&(dest->aes_mbs), &(src->aes_mbs)); archive_string_copy(&(dest->aes_utf8), &(src->aes_utf8)); archive_wstring_copy(&(dest->aes_wcs), &(src->aes_wcs)); } int archive_mstring_get_utf8(struct archive *a, struct archive_mstring *aes, const char **p) { struct archive_string_conv *sc; int r; /* If we already have a UTF8 form, return that immediately. */ if (aes->aes_set & AES_SET_UTF8) { *p = aes->aes_utf8.s; return (0); } *p = NULL; if (aes->aes_set & AES_SET_MBS) { sc = archive_string_conversion_to_charset(a, "UTF-8", 1); if (sc == NULL) return (-1);/* Couldn't allocate memory for sc. */ r = archive_strncpy_l(&(aes->aes_utf8), aes->aes_mbs.s, aes->aes_mbs.length, sc); if (a == NULL) free_sconv_object(sc); if (r == 0) { aes->aes_set |= AES_SET_UTF8; *p = aes->aes_utf8.s; return (0);/* success. */ } else return (-1);/* failure. */ } return (0);/* success. */ } int archive_mstring_get_mbs(struct archive *a, struct archive_mstring *aes, const char **p) { int r, ret = 0; (void)a; /* UNUSED */ /* If we already have an MBS form, return that immediately. */ if (aes->aes_set & AES_SET_MBS) { *p = aes->aes_mbs.s; return (ret); } *p = NULL; /* If there's a WCS form, try converting with the native locale. */ if (aes->aes_set & AES_SET_WCS) { archive_string_empty(&(aes->aes_mbs)); r = archive_string_append_from_wcs(&(aes->aes_mbs), aes->aes_wcs.s, aes->aes_wcs.length); *p = aes->aes_mbs.s; if (r == 0) { aes->aes_set |= AES_SET_MBS; return (ret); } else ret = -1; } /* * Only a UTF-8 form cannot avail because its conversion already * failed at archive_mstring_update_utf8(). */ return (ret); } int archive_mstring_get_wcs(struct archive *a, struct archive_mstring *aes, const wchar_t **wp) { int r, ret = 0; (void)a;/* UNUSED */ /* Return WCS form if we already have it. */ if (aes->aes_set & AES_SET_WCS) { *wp = aes->aes_wcs.s; return (ret); } *wp = NULL; /* Try converting MBS to WCS using native locale. */ if (aes->aes_set & AES_SET_MBS) { archive_wstring_empty(&(aes->aes_wcs)); r = archive_wstring_append_from_mbs(&(aes->aes_wcs), aes->aes_mbs.s, aes->aes_mbs.length); if (r == 0) { aes->aes_set |= AES_SET_WCS; *wp = aes->aes_wcs.s; } else ret = -1;/* failure. */ } return (ret); } int archive_mstring_get_mbs_l(struct archive_mstring *aes, const char **p, size_t *length, struct archive_string_conv *sc) { int r, ret = 0; #if defined(_WIN32) && !defined(__CYGWIN__) /* * Internationalization programming on Windows must use Wide * characters because Windows platform cannot make locale UTF-8. */ if (sc != NULL && (aes->aes_set & AES_SET_WCS) != 0) { archive_string_empty(&(aes->aes_mbs_in_locale)); r = archive_string_append_from_wcs_in_codepage( &(aes->aes_mbs_in_locale), aes->aes_wcs.s, aes->aes_wcs.length, sc); if (r == 0) { *p = aes->aes_mbs_in_locale.s; if (length != NULL) *length = aes->aes_mbs_in_locale.length; return (0); } else if (errno == ENOMEM) return (-1); else ret = -1; } #endif /* If there is not an MBS form but is a WCS form, try converting * with the native locale to be used for translating it to specified * character-set. */ if ((aes->aes_set & AES_SET_MBS) == 0 && (aes->aes_set & AES_SET_WCS) != 0) { archive_string_empty(&(aes->aes_mbs)); r = archive_string_append_from_wcs(&(aes->aes_mbs), aes->aes_wcs.s, aes->aes_wcs.length); if (r == 0) aes->aes_set |= AES_SET_MBS; else if (errno == ENOMEM) return (-1); else ret = -1; } /* If we already have an MBS form, use it to be translated to * specified character-set. */ if (aes->aes_set & AES_SET_MBS) { if (sc == NULL) { /* Conversion is unneeded. */ *p = aes->aes_mbs.s; if (length != NULL) *length = aes->aes_mbs.length; return (0); } ret = archive_strncpy_l(&(aes->aes_mbs_in_locale), aes->aes_mbs.s, aes->aes_mbs.length, sc); *p = aes->aes_mbs_in_locale.s; if (length != NULL) *length = aes->aes_mbs_in_locale.length; } else { *p = NULL; if (length != NULL) *length = 0; } return (ret); } int archive_mstring_copy_mbs(struct archive_mstring *aes, const char *mbs) { if (mbs == NULL) { aes->aes_set = 0; return (0); } return (archive_mstring_copy_mbs_len(aes, mbs, strlen(mbs))); } int archive_mstring_copy_mbs_len(struct archive_mstring *aes, const char *mbs, size_t len) { if (mbs == NULL) { aes->aes_set = 0; return (0); } aes->aes_set = AES_SET_MBS; /* Only MBS form is set now. */ archive_strncpy(&(aes->aes_mbs), mbs, len); archive_string_empty(&(aes->aes_utf8)); archive_wstring_empty(&(aes->aes_wcs)); return (0); } int archive_mstring_copy_wcs(struct archive_mstring *aes, const wchar_t *wcs) { return archive_mstring_copy_wcs_len(aes, wcs, wcs == NULL ? 0 : wcslen(wcs)); } int archive_mstring_copy_utf8(struct archive_mstring *aes, const char *utf8) { if (utf8 == NULL) { aes->aes_set = 0; } aes->aes_set = AES_SET_UTF8; archive_string_empty(&(aes->aes_mbs)); archive_string_empty(&(aes->aes_wcs)); archive_strncpy(&(aes->aes_utf8), utf8, strlen(utf8)); return (int)strlen(utf8); } int archive_mstring_copy_wcs_len(struct archive_mstring *aes, const wchar_t *wcs, size_t len) { if (wcs == NULL) { aes->aes_set = 0; } aes->aes_set = AES_SET_WCS; /* Only WCS form set. */ archive_string_empty(&(aes->aes_mbs)); archive_string_empty(&(aes->aes_utf8)); archive_wstrncpy(&(aes->aes_wcs), wcs, len); return (0); } int archive_mstring_copy_mbs_len_l(struct archive_mstring *aes, const char *mbs, size_t len, struct archive_string_conv *sc) { int r; if (mbs == NULL) { aes->aes_set = 0; return (0); } archive_string_empty(&(aes->aes_mbs)); archive_wstring_empty(&(aes->aes_wcs)); archive_string_empty(&(aes->aes_utf8)); #if defined(_WIN32) && !defined(__CYGWIN__) /* * Internationalization programming on Windows must use Wide * characters because Windows platform cannot make locale UTF-8. */ if (sc == NULL) { if (archive_string_append(&(aes->aes_mbs), mbs, mbsnbytes(mbs, len)) == NULL) { aes->aes_set = 0; r = -1; } else { aes->aes_set = AES_SET_MBS; r = 0; } #if defined(HAVE_ICONV) } else if (sc != NULL && sc->cd_w != (iconv_t)-1) { /* * This case happens only when MultiByteToWideChar() cannot * handle sc->from_cp, and we have to iconv in order to * translate character-set to wchar_t,UTF-16. */ iconv_t cd = sc->cd; unsigned from_cp; int flag; /* * Translate multi-bytes from some character-set to UTF-8. */ sc->cd = sc->cd_w; r = archive_strncpy_l(&(aes->aes_utf8), mbs, len, sc); sc->cd = cd; if (r != 0) { aes->aes_set = 0; return (r); } aes->aes_set = AES_SET_UTF8; /* * Append the UTF-8 string into wstring. */ flag = sc->flag; sc->flag &= ~(SCONV_NORMALIZATION_C | SCONV_TO_UTF16| SCONV_FROM_UTF16); from_cp = sc->from_cp; sc->from_cp = CP_UTF8; r = archive_wstring_append_from_mbs_in_codepage(&(aes->aes_wcs), aes->aes_utf8.s, aes->aes_utf8.length, sc); sc->flag = flag; sc->from_cp = from_cp; if (r == 0) aes->aes_set |= AES_SET_WCS; #endif } else { r = archive_wstring_append_from_mbs_in_codepage( &(aes->aes_wcs), mbs, len, sc); if (r == 0) aes->aes_set = AES_SET_WCS; else aes->aes_set = 0; } #else r = archive_strncpy_l(&(aes->aes_mbs), mbs, len, sc); if (r == 0) aes->aes_set = AES_SET_MBS; /* Only MBS form is set now. */ else aes->aes_set = 0; #endif return (r); } /* * The 'update' form tries to proactively update all forms of * this string (WCS and MBS) and returns an error if any of * them fail. This is used by the 'pax' handler, for instance, * to detect and report character-conversion failures early while * still allowing clients to get potentially useful values from * the more tolerant lazy conversions. (get_mbs and get_wcs will * strive to give the user something useful, so you can get hopefully * usable values even if some of the character conversions are failing.) */ int archive_mstring_update_utf8(struct archive *a, struct archive_mstring *aes, const char *utf8) { struct archive_string_conv *sc; int r; if (utf8 == NULL) { aes->aes_set = 0; return (0); /* Succeeded in clearing everything. */ } /* Save the UTF8 string. */ archive_strcpy(&(aes->aes_utf8), utf8); /* Empty the mbs and wcs strings. */ archive_string_empty(&(aes->aes_mbs)); archive_wstring_empty(&(aes->aes_wcs)); aes->aes_set = AES_SET_UTF8; /* Only UTF8 is set now. */ /* Try converting UTF-8 to MBS, return false on failure. */ sc = archive_string_conversion_from_charset(a, "UTF-8", 1); if (sc == NULL) return (-1);/* Couldn't allocate memory for sc. */ r = archive_strcpy_l(&(aes->aes_mbs), utf8, sc); if (a == NULL) free_sconv_object(sc); if (r != 0) return (-1); aes->aes_set = AES_SET_UTF8 | AES_SET_MBS; /* Both UTF8 and MBS set. */ /* Try converting MBS to WCS, return false on failure. */ if (archive_wstring_append_from_mbs(&(aes->aes_wcs), aes->aes_mbs.s, aes->aes_mbs.length)) return (-1); aes->aes_set = AES_SET_UTF8 | AES_SET_WCS | AES_SET_MBS; /* All conversions succeeded. */ return (0); } Index: stable/10/contrib/libarchive/libarchive/archive_version_details.c =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_version_details.c (revision 324417) +++ stable/10/contrib/libarchive/libarchive/archive_version_details.c (revision 324418) @@ -1,133 +1,151 @@ /*- * Copyright (c) 2009-2012,2014 Michihiro NAKAJIMA * Copyright (c) 2003-2007 Tim Kientzle * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "archive_platform.h" __FBSDID("$FreeBSD$"); #ifdef HAVE_STDLIB_H #include #endif #ifdef HAVE_STRING_H #include #endif #ifdef HAVE_ZLIB_H #include #endif #ifdef HAVE_LZMA_H #include #endif #ifdef HAVE_BZLIB_H #include #endif #ifdef HAVE_LZ4_H #include #endif +#ifdef HAVE_ZSTD_H +#include +#endif #include "archive.h" #include "archive_private.h" #include "archive_string.h" const char * archive_version_details(void) { static struct archive_string str; static int init = 0; const char *zlib = archive_zlib_version(); const char *liblzma = archive_liblzma_version(); const char *bzlib = archive_bzlib_version(); const char *liblz4 = archive_liblz4_version(); + const char *libzstd = archive_libzstd_version(); if (!init) { archive_string_init(&str); archive_strcat(&str, ARCHIVE_VERSION_STRING); if (zlib != NULL) { archive_strcat(&str, " zlib/"); archive_strcat(&str, zlib); } if (liblzma) { archive_strcat(&str, " liblzma/"); archive_strcat(&str, liblzma); } if (bzlib) { const char *p = bzlib; const char *sep = strchr(p, ','); if (sep == NULL) sep = p + strlen(p); archive_strcat(&str, " bz2lib/"); archive_strncat(&str, p, sep - p); } if (liblz4) { archive_strcat(&str, " liblz4/"); archive_strcat(&str, liblz4); } + if (libzstd) { + archive_strcat(&str, " libzstd/"); + archive_strcat(&str, libzstd); + } } return str.s; } const char * archive_zlib_version(void) { #ifdef HAVE_ZLIB_H return ZLIB_VERSION; #else return NULL; #endif } const char * archive_liblzma_version(void) { #ifdef HAVE_LZMA_H return LZMA_VERSION_STRING; #else return NULL; #endif } const char * archive_bzlib_version(void) { #ifdef HAVE_BZLIB_H return BZ2_bzlibVersion(); #else return NULL; #endif } const char * archive_liblz4_version(void) { #if defined(HAVE_LZ4_H) && defined(HAVE_LIBLZ4) #define str(s) #s #define NUMBER(x) str(x) return NUMBER(LZ4_VERSION_MAJOR) "." NUMBER(LZ4_VERSION_MINOR) "." NUMBER(LZ4_VERSION_RELEASE); #undef NUMBER #undef str +#else + return NULL; +#endif +} + +const char * +archive_libzstd_version(void) +{ +#if HAVE_ZSTD_H && HAVE_LIBZSTD + return ZSTD_VERSION_STRING; #else return NULL; #endif } Index: stable/10/contrib/libarchive/libarchive/archive_write.3 =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_write.3 (revision 324417) +++ stable/10/contrib/libarchive/libarchive/archive_write.3 (revision 324418) @@ -1,268 +1,268 @@ .\" Copyright (c) 2003-2011 Tim Kientzle .\" All rights reserved. .\" .\" Redistribution and use in source and binary forms, with or without .\" modification, are permitted provided that the following conditions .\" are met: .\" 1. Redistributions of source code must retain the above copyright .\" notice, this list of conditions and the following disclaimer. .\" 2. Redistributions in binary form must reproduce the above copyright .\" notice, this list of conditions and the following disclaimer in the .\" documentation and/or other materials provided with the distribution. .\" .\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE .\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF .\" SUCH DAMAGE. .\" .\" $FreeBSD$ .\" .Dd February 2, 2012 .Dt ARCHIVE_WRITE 3 .Os .Sh NAME .Nm archive_write .Nd functions for creating archives .Sh LIBRARY Streaming Archive Library (libarchive, -larchive) .Sh SYNOPSIS .In archive.h .Sh DESCRIPTION These functions provide a complete API for creating streaming archive files. The general process is to first create the .Tn struct archive object, set any desired options, initialize the archive, append entries, then close the archive and release all resources. .\" .Ss Create archive object See .Xr archive_write_new 3 . .Pp To write an archive, you must first obtain an initialized .Tn struct archive object from .Fn archive_write_new . .\" .Ss Enable filters and formats, configure block size and padding See .Xr archive_write_filter 3 , .Xr archive_write_format 3 and .Xr archive_write_blocksize 3 . .Pp You can then modify this object for the desired operations with the various .Fn archive_write_set_XXX functions. In particular, you will need to invoke appropriate .Fn archive_write_add_XXX and .Fn archive_write_set_XXX functions to enable the corresponding compression and format support. .\" .Ss Set options See -.Xr archive_read_set_options 3 . +.Xr archive_write_set_options 3 . .\" .Ss Open archive See .Xr archive_write_open 3 . .Pp Once you have prepared the .Tn struct archive object, you call .Fn archive_write_open to actually open the archive and prepare it for writing. There are several variants of this function; the most basic expects you to provide pointers to several functions that can provide blocks of bytes from the archive. There are convenience forms that allow you to specify a filename, file descriptor, .Ft "FILE *" object, or a block of memory from which to write the archive data. .\" .Ss Produce archive See .Xr archive_write_header 3 and .Xr archive_write_data 3 . .Pp Individual archive entries are written in a three-step process: You first initialize a .Tn struct archive_entry structure with information about the new entry. At a minimum, you should set the pathname of the entry and provide a .Va struct stat with a valid .Va st_mode field, which specifies the type of object and .Va st_size field, which specifies the size of the data portion of the object. .\" .Ss Release resources See .Xr archive_write_free 3 . .Pp After all entries have been written, use the .Fn archive_write_free function to release all resources. .\" .Sh EXAMPLE The following sketch illustrates basic usage of the library. In this example, the callback functions are simply wrappers around the standard .Xr open 2 , .Xr write 2 , and .Xr close 2 system calls. .Bd -literal -offset indent #ifdef __linux__ #define _FILE_OFFSET_BITS 64 #endif #include #include #include #include #include #include struct mydata { const char *name; int fd; }; int myopen(struct archive *a, void *client_data) { struct mydata *mydata = client_data; mydata->fd = open(mydata->name, O_WRONLY | O_CREAT, 0644); if (mydata->fd >= 0) return (ARCHIVE_OK); else return (ARCHIVE_FATAL); } la_ssize_t mywrite(struct archive *a, void *client_data, const void *buff, size_t n) { struct mydata *mydata = client_data; return (write(mydata->fd, buff, n)); } int myclose(struct archive *a, void *client_data) { struct mydata *mydata = client_data; if (mydata->fd > 0) close(mydata->fd); return (0); } void write_archive(const char *outname, const char **filename) { struct mydata *mydata = malloc(sizeof(struct mydata)); struct archive *a; struct archive_entry *entry; struct stat st; char buff[8192]; int len; int fd; a = archive_write_new(); mydata->name = outname; /* Set archive format and filter according to output file extension. * If it fails, set default format. Platform depended function. * See supported formats in archive_write_set_format_filter_by_ext.c */ if (archive_write_set_format_filter_by_ext(a, outname) != ARCHIVE_OK) { archive_write_add_filter_gzip(a); archive_write_set_format_ustar(a); } archive_write_open(a, mydata, myopen, mywrite, myclose); while (*filename) { stat(*filename, &st); entry = archive_entry_new(); archive_entry_copy_stat(entry, &st); archive_entry_set_pathname(entry, *filename); archive_write_header(a, entry); if ((fd = open(*filename, O_RDONLY)) != -1) { len = read(fd, buff, sizeof(buff)); while (len > 0) { archive_write_data(a, buff, len); len = read(fd, buff, sizeof(buff)); } close(fd); } archive_entry_free(entry); filename++; } archive_write_free(a); } int main(int argc, const char **argv) { const char *outname; argv++; outname = *argv++; write_archive(outname, argv); return 0; } .Ed .Sh SEE ALSO .Xr tar 1 , .Xr libarchive 3 , .Xr archive_write_set_options 3 , .Xr cpio 5 , .Xr mtree 5 , .Xr tar 5 .Sh HISTORY The .Nm libarchive library first appeared in .Fx 5.3 . .Sh AUTHORS .An -nosplit The .Nm libarchive library was written by .An Tim Kientzle Aq kientzle@acm.org . .Sh BUGS There are many peculiar bugs in historic tar implementations that may cause certain programs to reject archives written by this library. For example, several historic implementations calculated header checksums incorrectly and will thus reject valid archives; GNU tar does not fully support pax interchange format; some old tar implementations required specific field terminations. .Pp The default pax interchange format eliminates most of the historic tar limitations and provides a generic key/value attribute facility for vendor-defined extensions. One oversight in POSIX is the failure to provide a standard attribute for large device numbers. This library uses .Dq SCHILY.devminor and .Dq SCHILY.devmajor for device numbers that exceed the range supported by the backwards-compatible ustar header. These keys are compatible with Joerg Schilling's .Nm star archiver. Other implementations may not recognize these keys and will thus be unable to correctly restore device nodes with large device numbers from archives created by this library. Index: stable/10/contrib/libarchive/libarchive/archive_write_add_filter.c =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_write_add_filter.c (revision 324417) +++ stable/10/contrib/libarchive/libarchive/archive_write_add_filter.c (revision 324418) @@ -1,71 +1,72 @@ /*- * Copyright (c) 2012 Ondrej Holy * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "archive_platform.h" __FBSDID("$FreeBSD$"); #ifdef HAVE_SYS_TYPES_H #include #endif #ifdef HAVE_ERRNO_H #include #endif #include "archive.h" #include "archive_private.h" /* A table that maps filter codes to functions. */ static const struct { int code; int (*setter)(struct archive *); } codes[] = { { ARCHIVE_FILTER_NONE, archive_write_add_filter_none }, { ARCHIVE_FILTER_GZIP, archive_write_add_filter_gzip }, { ARCHIVE_FILTER_BZIP2, archive_write_add_filter_bzip2 }, { ARCHIVE_FILTER_COMPRESS, archive_write_add_filter_compress }, { ARCHIVE_FILTER_GRZIP, archive_write_add_filter_grzip }, { ARCHIVE_FILTER_LRZIP, archive_write_add_filter_lrzip }, { ARCHIVE_FILTER_LZ4, archive_write_add_filter_lz4 }, { ARCHIVE_FILTER_LZIP, archive_write_add_filter_lzip }, { ARCHIVE_FILTER_LZMA, archive_write_add_filter_lzma }, { ARCHIVE_FILTER_LZOP, archive_write_add_filter_lzip }, { ARCHIVE_FILTER_UU, archive_write_add_filter_uuencode }, { ARCHIVE_FILTER_XZ, archive_write_add_filter_xz }, + { ARCHIVE_FILTER_ZSTD, archive_write_add_filter_zstd }, { -1, NULL } }; int archive_write_add_filter(struct archive *a, int code) { int i; for (i = 0; codes[i].code != -1; i++) { if (code == codes[i].code) return ((codes[i].setter)(a)); } archive_set_error(a, EINVAL, "No such filter"); return (ARCHIVE_FATAL); } Index: stable/10/contrib/libarchive/libarchive/archive_write_add_filter_by_name.c =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_write_add_filter_by_name.c (revision 324417) +++ stable/10/contrib/libarchive/libarchive/archive_write_add_filter_by_name.c (revision 324418) @@ -1,76 +1,77 @@ /*- * Copyright (c) 2003-2007 Tim Kientzle * Copyright (c) 2012 Michihiro NAKAJIMA * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "archive_platform.h" __FBSDID("$FreeBSD$"); #ifdef HAVE_SYS_TYPES_H #include #endif #ifdef HAVE_ERRNO_H #include #endif #ifdef HAVE_STRING_H #include #endif #include "archive.h" #include "archive_private.h" /* A table that maps names to functions. */ static const struct { const char *name; int (*setter)(struct archive *); } names[] = { { "b64encode", archive_write_add_filter_b64encode }, { "bzip2", archive_write_add_filter_bzip2 }, { "compress", archive_write_add_filter_compress }, { "grzip", archive_write_add_filter_grzip }, { "gzip", archive_write_add_filter_gzip }, { "lrzip", archive_write_add_filter_lrzip }, { "lz4", archive_write_add_filter_lz4 }, { "lzip", archive_write_add_filter_lzip }, { "lzma", archive_write_add_filter_lzma }, { "lzop", archive_write_add_filter_lzop }, { "uuencode", archive_write_add_filter_uuencode }, { "xz", archive_write_add_filter_xz }, + { "zstd", archive_write_add_filter_zstd }, { NULL, NULL } }; int archive_write_add_filter_by_name(struct archive *a, const char *name) { int i; for (i = 0; names[i].name != NULL; i++) { if (strcmp(name, names[i].name) == 0) return ((names[i].setter)(a)); } archive_set_error(a, EINVAL, "No such filter '%s'", name); a->state = ARCHIVE_STATE_FATAL; return (ARCHIVE_FATAL); } Index: stable/10/contrib/libarchive/libarchive/archive_write_add_filter_zstd.c =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_write_add_filter_zstd.c (nonexistent) +++ stable/10/contrib/libarchive/libarchive/archive_write_add_filter_zstd.c (revision 324418) @@ -0,0 +1,335 @@ +/*- + * Copyright (c) 2017 Sean Purcell + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES + * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. + * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, + * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT + * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF + * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ + +#include "archive_platform.h" + +__FBSDID("$FreeBSD$"); + + +#ifdef HAVE_ERRNO_H +#include +#endif +#ifdef HAVE_STDLIB_H +#include +#endif +#ifdef HAVE_STRING_H +#include +#endif +#ifdef HAVE_ZSTD_H +#include +#endif + +#include "archive.h" +#include "archive_private.h" +#include "archive_string.h" +#include "archive_write_private.h" + +/* Don't compile this if we don't have zstd.h */ + +struct private_data { + int compression_level; +#if HAVE_ZSTD_H && HAVE_LIBZSTD + ZSTD_CStream *cstream; + int64_t total_in; + ZSTD_outBuffer out; +#else + struct archive_write_program_data *pdata; +#endif +}; + +static int archive_compressor_zstd_options(struct archive_write_filter *, + const char *, const char *); +static int archive_compressor_zstd_open(struct archive_write_filter *); +static int archive_compressor_zstd_write(struct archive_write_filter *, + const void *, size_t); +static int archive_compressor_zstd_close(struct archive_write_filter *); +static int archive_compressor_zstd_free(struct archive_write_filter *); +#if HAVE_ZSTD_H && HAVE_LIBZSTD +static int drive_compressor(struct archive_write_filter *, + struct private_data *, int, const void *, size_t); +#endif + + +/* + * Add a zstd compression filter to this write handle. + */ +int +archive_write_add_filter_zstd(struct archive *_a) +{ + struct archive_write *a = (struct archive_write *)_a; + struct archive_write_filter *f = __archive_write_allocate_filter(_a); + struct private_data *data; + archive_check_magic(&a->archive, ARCHIVE_WRITE_MAGIC, + ARCHIVE_STATE_NEW, "archive_write_add_filter_zstd"); + + data = calloc(1, sizeof(*data)); + if (data == NULL) { + archive_set_error(&a->archive, ENOMEM, "Out of memory"); + return (ARCHIVE_FATAL); + } + f->data = data; + f->open = &archive_compressor_zstd_open; + f->options = &archive_compressor_zstd_options; + f->close = &archive_compressor_zstd_close; + f->free = &archive_compressor_zstd_free; + f->code = ARCHIVE_FILTER_ZSTD; + f->name = "zstd"; + data->compression_level = 3; /* Default level used by the zstd CLI */ +#if HAVE_ZSTD_H && HAVE_LIBZSTD + data->cstream = ZSTD_createCStream(); + if (data->cstream == NULL) { + free(data); + archive_set_error(&a->archive, ENOMEM, + "Failed to allocate zstd compressor object"); + return (ARCHIVE_FATAL); + } + + return (ARCHIVE_OK); +#else + data->pdata = __archive_write_program_allocate("zstd"); + if (data->pdata == NULL) { + free(data); + archive_set_error(&a->archive, ENOMEM, "Out of memory"); + return (ARCHIVE_FATAL); + } + archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, + "Using external zstd program"); + return (ARCHIVE_WARN); +#endif +} + +static int +archive_compressor_zstd_free(struct archive_write_filter *f) +{ + struct private_data *data = (struct private_data *)f->data; +#if HAVE_ZSTD_H && HAVE_LIBZSTD + ZSTD_freeCStream(data->cstream); + free(data->out.dst); +#else + __archive_write_program_free(data->pdata); +#endif + free(data); + f->data = NULL; + return (ARCHIVE_OK); +} + +/* + * Set write options. + */ +static int +archive_compressor_zstd_options(struct archive_write_filter *f, const char *key, + const char *value) +{ + struct private_data *data = (struct private_data *)f->data; + + if (strcmp(key, "compression-level") == 0) { + int level = atoi(value); +#if HAVE_ZSTD_H && HAVE_LIBZSTD + if (level < 1 || level > ZSTD_maxCLevel()) { +#else + /* If we don't have the library, hard-code the max level */ + if (level < 1 || level > 22) { +#endif + return (ARCHIVE_WARN); + } + data->compression_level = level; + return (ARCHIVE_OK); + } + + /* Note: The "warn" return is just to inform the options + * supervisor that we didn't handle it. It will generate + * a suitable error if no one used this option. */ + return (ARCHIVE_WARN); +} + +#if HAVE_ZSTD_H && HAVE_LIBZSTD +/* + * Setup callback. + */ +static int +archive_compressor_zstd_open(struct archive_write_filter *f) +{ + struct private_data *data = (struct private_data *)f->data; + int ret; + + ret = __archive_write_open_filter(f->next_filter); + if (ret != ARCHIVE_OK) + return (ret); + + if (data->out.dst == NULL) { + size_t bs = ZSTD_CStreamOutSize(), bpb; + if (f->archive->magic == ARCHIVE_WRITE_MAGIC) { + /* Buffer size should be a multiple number of + * the of bytes per block for performance. */ + bpb = archive_write_get_bytes_per_block(f->archive); + if (bpb > bs) + bs = bpb; + else if (bpb != 0) + bs -= bs % bpb; + } + data->out.size = bs; + data->out.pos = 0; + data->out.dst + = (unsigned char *)malloc(data->out.size); + if (data->out.dst == NULL) { + archive_set_error(f->archive, ENOMEM, + "Can't allocate data for compression buffer"); + return (ARCHIVE_FATAL); + } + } + + f->write = archive_compressor_zstd_write; + + if (ZSTD_isError(ZSTD_initCStream(data->cstream, + data->compression_level))) { + archive_set_error(f->archive, ARCHIVE_ERRNO_MISC, + "Internal error initializing zstd compressor object"); + return (ARCHIVE_FATAL); + } + + return (ARCHIVE_OK); +} + +/* + * Write data to the compressed stream. + */ +static int +archive_compressor_zstd_write(struct archive_write_filter *f, const void *buff, + size_t length) +{ + struct private_data *data = (struct private_data *)f->data; + int ret; + + /* Update statistics */ + data->total_in += length; + + if ((ret = drive_compressor(f, data, 0, buff, length)) != ARCHIVE_OK) + return (ret); + + return (ARCHIVE_OK); +} + +/* + * Finish the compression... + */ +static int +archive_compressor_zstd_close(struct archive_write_filter *f) +{ + struct private_data *data = (struct private_data *)f->data; + int r1, r2; + + /* Finish zstd frame */ + r1 = drive_compressor(f, data, 1, NULL, 0); + + r2 = __archive_write_close_filter(f->next_filter); + + return r1 < r2 ? r1 : r2; +} + +/* + * Utility function to push input data through compressor, + * writing full output blocks as necessary. + * + * Note that this handles both the regular write case (finishing == + * false) and the end-of-archive case (finishing == true). + */ +static int +drive_compressor(struct archive_write_filter *f, + struct private_data *data, int finishing, const void *src, size_t length) +{ + ZSTD_inBuffer in = (ZSTD_inBuffer) { src, length, 0 }; + + for (;;) { + if (data->out.pos == data->out.size) { + const int ret = __archive_write_filter(f->next_filter, + data->out.dst, data->out.size); + if (ret != ARCHIVE_OK) + return (ARCHIVE_FATAL); + data->out.pos = 0; + } + + /* If there's nothing to do, we're done. */ + if (!finishing && in.pos == in.size) + return (ARCHIVE_OK); + + { + const size_t zstdret = !finishing ? + ZSTD_compressStream(data->cstream, &data->out, &in) + : ZSTD_endStream(data->cstream, &data->out); + + if (ZSTD_isError(zstdret)) { + archive_set_error(f->archive, + ARCHIVE_ERRNO_MISC, + "Zstd compression failed: %s", + ZSTD_getErrorName(zstdret)); + return (ARCHIVE_FATAL); + } + + /* If we're finishing, 0 means nothing left to flush */ + if (finishing && zstdret == 0) { + const int ret = __archive_write_filter(f->next_filter, + data->out.dst, data->out.pos); + return (ret); + } + } + } +} + +#else /* HAVE_ZSTD_H && HAVE_LIBZSTD */ + +static int +archive_compressor_zstd_open(struct archive_write_filter *f) +{ + struct private_data *data = (struct private_data *)f->data; + struct archive_string as; + int r; + + archive_string_init(&as); + archive_string_sprintf(&as, "zstd -%d", data->compression_level); + + f->write = archive_compressor_zstd_write; + r = __archive_write_program_open(f, data->pdata, as.s); + archive_string_free(&as); + return (r); +} + +static int +archive_compressor_zstd_write(struct archive_write_filter *f, const void *buff, + size_t length) +{ + struct private_data *data = (struct private_data *)f->data; + + return __archive_write_program_write(f, data->pdata, buff, length); +} + +static int +archive_compressor_zstd_close(struct archive_write_filter *f) +{ + struct private_data *data = (struct private_data *)f->data; + + return __archive_write_program_close(f, data->pdata); +} + +#endif /* HAVE_ZSTD_H && HAVE_LIBZSTD */ Property changes on: stable/10/contrib/libarchive/libarchive/archive_write_add_filter_zstd.c ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: stable/10/contrib/libarchive/libarchive/archive_write_filter.3 =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_write_filter.3 (revision 324417) +++ stable/10/contrib/libarchive/libarchive/archive_write_filter.3 (revision 324418) @@ -1,130 +1,134 @@ .\" Copyright (c) 2003-2011 Tim Kientzle .\" All rights reserved. .\" .\" Redistribution and use in source and binary forms, with or without .\" modification, are permitted provided that the following conditions .\" are met: .\" 1. Redistributions of source code must retain the above copyright .\" notice, this list of conditions and the following disclaimer. .\" 2. Redistributions in binary form must reproduce the above copyright .\" notice, this list of conditions and the following disclaimer in the .\" documentation and/or other materials provided with the distribution. .\" .\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE .\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF .\" SUCH DAMAGE. .\" .\" $FreeBSD$ .\" .Dd August 14, 2014 .Dt ARCHIVE_WRITE_FILTER 3 .Os .Sh NAME .Nm archive_write_add_filter_b64encode , .Nm archive_write_add_filter_by_name , .Nm archive_write_add_filter_bzip2 , .Nm archive_write_add_filter_compress , .Nm archive_write_add_filter_grzip , .Nm archive_write_add_filter_gzip , .Nm archive_write_add_filter_lrzip , .Nm archive_write_add_filter_lz4 , .Nm archive_write_add_filter_lzip , .Nm archive_write_add_filter_lzma , .Nm archive_write_add_filter_lzop , .Nm archive_write_add_filter_none , .Nm archive_write_add_filter_program , .Nm archive_write_add_filter_uuencode , -.Nm archive_write_add_filter_xz +.Nm archive_write_add_filter_xz , +.Nm archive_write_add_filter_zstd , .Nd functions enabling output filters .Sh LIBRARY Streaming Archive Library (libarchive, -larchive) .Sh SYNOPSIS .In archive.h .Ft int .Fn archive_write_add_filter_b64encode "struct archive *" .Ft int .Fn archive_write_add_filter_bzip2 "struct archive *" .Ft int .Fn archive_write_add_filter_compress "struct archive *" .Ft int .Fn archive_write_add_filter_grzip "struct archive *" .Ft int .Fn archive_write_add_filter_gzip "struct archive *" .Ft int .Fn archive_write_add_filter_lrzip "struct archive *" .Ft int .Fn archive_write_add_filter_lz4 "struct archive *" .Ft int .Fn archive_write_add_filter_lzip "struct archive *" .Ft int .Fn archive_write_add_filter_lzma "struct archive *" .Ft int .Fn archive_write_add_filter_lzop "struct archive *" .Ft int .Fn archive_write_add_filter_none "struct archive *" .Ft int .Fn archive_write_add_filter_program "struct archive *" "const char * cmd" .Ft int .Fn archive_write_add_filter_uuencode "struct archive *" .Ft int .Fn archive_write_add_filter_xz "struct archive *" +.Ft int +.Fn archive_write_add_filter_zstd "struct archive *" .Sh DESCRIPTION .Bl -tag -width indent .It Xo .Fn archive_write_add_filter_bzip2 , .Fn archive_write_add_filter_compress , .Fn archive_write_add_filter_grzip , .Fn archive_write_add_filter_gzip , .Fn archive_write_add_filter_lrzip , .Fn archive_write_add_filter_lz4 , .Fn archive_write_add_filter_lzip , .Fn archive_write_add_filter_lzma , .Fn archive_write_add_filter_lzop , .Fn archive_write_add_filter_xz , +.Fn archive_write_add_filter_zstd , .Xc The resulting archive will be compressed as specified. Note that the compressed output is always properly blocked. .It Xo .Fn archive_write_add_filter_b64encode , .Fn archive_write_add_filter_uuencode , .Xc The output will be encoded as specified. The encoded output is always properly blocked. .It Fn archive_write_add_filter_none This is never necessary. It is provided only for backwards compatibility. .It Fn archive_write_add_filter_program The archive will be fed into the specified compression program. The output of that program is blocked and written to the client write callbacks. .El .Sh RETURN VALUES These functions return .Cm ARCHIVE_OK on success, or .Cm ARCHIVE_FATAL . .\" .Sh ERRORS Detailed error codes and textual descriptions are available from the .Fn archive_errno and .Fn archive_error_string functions. .\" .Sh SEE ALSO .Xr tar 1 , .Xr libarchive 3 , .Xr archive_write 3 , .Xr archive_write_format 3 , .Xr archive_write_set_options 3 , .Xr cpio 5 , .Xr mtree 5 , .Xr tar 5 Index: stable/10/contrib/libarchive/libarchive/archive_write_set_format_pax.c =================================================================== --- stable/10/contrib/libarchive/libarchive/archive_write_set_format_pax.c (revision 324417) +++ stable/10/contrib/libarchive/libarchive/archive_write_set_format_pax.c (revision 324418) @@ -1,1969 +1,1969 @@ /*- * Copyright (c) 2003-2007 Tim Kientzle * Copyright (c) 2010-2012 Michihiro NAKAJIMA * Copyright (c) 2016 Martin Matuska * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "archive_platform.h" __FBSDID("$FreeBSD$"); #ifdef HAVE_ERRNO_H #include #endif #ifdef HAVE_STDLIB_H #include #endif #ifdef HAVE_STRING_H #include #endif #include "archive.h" #include "archive_entry.h" #include "archive_entry_locale.h" #include "archive_private.h" #include "archive_write_private.h" struct sparse_block { struct sparse_block *next; int is_hole; uint64_t offset; uint64_t remaining; }; struct pax { uint64_t entry_bytes_remaining; uint64_t entry_padding; struct archive_string l_url_encoded_name; struct archive_string pax_header; struct archive_string sparse_map; size_t sparse_map_padding; struct sparse_block *sparse_list; struct sparse_block *sparse_tail; struct archive_string_conv *sconv_utf8; int opt_binary; unsigned flags; #define WRITE_SCHILY_XATTR (1 << 0) #define WRITE_LIBARCHIVE_XATTR (1 << 1) }; static void add_pax_attr(struct archive_string *, const char *key, const char *value); static void add_pax_attr_binary(struct archive_string *, const char *key, const char *value, size_t value_len); static void add_pax_attr_int(struct archive_string *, const char *key, int64_t value); static void add_pax_attr_time(struct archive_string *, const char *key, int64_t sec, unsigned long nanos); static int add_pax_acl(struct archive_write *, struct archive_entry *, struct pax *, int); static ssize_t archive_write_pax_data(struct archive_write *, const void *, size_t); static int archive_write_pax_close(struct archive_write *); static int archive_write_pax_free(struct archive_write *); static int archive_write_pax_finish_entry(struct archive_write *); static int archive_write_pax_header(struct archive_write *, struct archive_entry *); static int archive_write_pax_options(struct archive_write *, const char *, const char *); static char *base64_encode(const char *src, size_t len); static char *build_gnu_sparse_name(char *dest, const char *src); static char *build_pax_attribute_name(char *dest, const char *src); static char *build_ustar_entry_name(char *dest, const char *src, size_t src_length, const char *insert); static char *format_int(char *dest, int64_t); static int has_non_ASCII(const char *); static void sparse_list_clear(struct pax *); static int sparse_list_add(struct pax *, int64_t, int64_t); static char *url_encode(const char *in); /* * Set output format to 'restricted pax' format. * * This is the same as normal 'pax', but tries to suppress * the pax header whenever possible. This is the default for * bsdtar, for instance. */ int archive_write_set_format_pax_restricted(struct archive *_a) { struct archive_write *a = (struct archive_write *)_a; int r; archive_check_magic(_a, ARCHIVE_WRITE_MAGIC, ARCHIVE_STATE_NEW, "archive_write_set_format_pax_restricted"); r = archive_write_set_format_pax(&a->archive); a->archive.archive_format = ARCHIVE_FORMAT_TAR_PAX_RESTRICTED; a->archive.archive_format_name = "restricted POSIX pax interchange"; return (r); } /* * Set output format to 'pax' format. */ int archive_write_set_format_pax(struct archive *_a) { struct archive_write *a = (struct archive_write *)_a; struct pax *pax; archive_check_magic(_a, ARCHIVE_WRITE_MAGIC, ARCHIVE_STATE_NEW, "archive_write_set_format_pax"); if (a->format_free != NULL) (a->format_free)(a); pax = (struct pax *)calloc(1, sizeof(*pax)); if (pax == NULL) { archive_set_error(&a->archive, ENOMEM, "Can't allocate pax data"); return (ARCHIVE_FATAL); } pax->flags = WRITE_LIBARCHIVE_XATTR | WRITE_SCHILY_XATTR; a->format_data = pax; a->format_name = "pax"; a->format_options = archive_write_pax_options; a->format_write_header = archive_write_pax_header; a->format_write_data = archive_write_pax_data; a->format_close = archive_write_pax_close; a->format_free = archive_write_pax_free; a->format_finish_entry = archive_write_pax_finish_entry; a->archive.archive_format = ARCHIVE_FORMAT_TAR_PAX_INTERCHANGE; a->archive.archive_format_name = "POSIX pax interchange"; return (ARCHIVE_OK); } static int archive_write_pax_options(struct archive_write *a, const char *key, const char *val) { struct pax *pax = (struct pax *)a->format_data; int ret = ARCHIVE_FAILED; if (strcmp(key, "hdrcharset") == 0) { /* * The character-set we can use are defined in * IEEE Std 1003.1-2001 */ if (val == NULL || val[0] == 0) archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "pax: hdrcharset option needs a character-set name"); else if (strcmp(val, "BINARY") == 0 || strcmp(val, "binary") == 0) { /* * Specify binary mode. We will not convert * filenames, uname and gname to any charsets. */ pax->opt_binary = 1; ret = ARCHIVE_OK; } else if (strcmp(val, "UTF-8") == 0) { /* * Specify UTF-8 character-set to be used for * filenames. This is almost the test that * running platform supports the string conversion. * Especially libarchive_test needs this trick for * its test. */ pax->sconv_utf8 = archive_string_conversion_to_charset( &(a->archive), "UTF-8", 0); if (pax->sconv_utf8 == NULL) ret = ARCHIVE_FATAL; else ret = ARCHIVE_OK; } else archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "pax: invalid charset name"); return (ret); } /* Note: The "warn" return is just to inform the options * supervisor that we didn't handle it. It will generate * a suitable error if no one used this option. */ return (ARCHIVE_WARN); } /* * Note: This code assumes that 'nanos' has the same sign as 'sec', * which implies that sec=-1, nanos=200000000 represents -1.2 seconds * and not -0.8 seconds. This is a pretty pedantic point, as we're * unlikely to encounter many real files created before Jan 1, 1970, * much less ones with timestamps recorded to sub-second resolution. */ static void add_pax_attr_time(struct archive_string *as, const char *key, int64_t sec, unsigned long nanos) { int digit, i; char *t; /* * Note that each byte contributes fewer than 3 base-10 * digits, so this will always be big enough. */ char tmp[1 + 3*sizeof(sec) + 1 + 3*sizeof(nanos)]; tmp[sizeof(tmp) - 1] = 0; t = tmp + sizeof(tmp) - 1; /* Skip trailing zeros in the fractional part. */ for (digit = 0, i = 10; i > 0 && digit == 0; i--) { digit = nanos % 10; nanos /= 10; } /* Only format the fraction if it's non-zero. */ if (i > 0) { while (i > 0) { *--t = "0123456789"[digit]; digit = nanos % 10; nanos /= 10; i--; } *--t = '.'; } t = format_int(t, sec); add_pax_attr(as, key, t); } static char * format_int(char *t, int64_t i) { uint64_t ui; if (i < 0) ui = (i == INT64_MIN) ? (uint64_t)(INT64_MAX) + 1 : (uint64_t)(-i); else ui = i; do { *--t = "0123456789"[ui % 10]; } while (ui /= 10); if (i < 0) *--t = '-'; return (t); } static void add_pax_attr_int(struct archive_string *as, const char *key, int64_t value) { char tmp[1 + 3 * sizeof(value)]; tmp[sizeof(tmp) - 1] = 0; add_pax_attr(as, key, format_int(tmp + sizeof(tmp) - 1, value)); } /* * Add a key/value attribute to the pax header. This function handles * the length field and various other syntactic requirements. */ static void add_pax_attr(struct archive_string *as, const char *key, const char *value) { add_pax_attr_binary(as, key, value, strlen(value)); } /* * Add a key/value attribute to the pax header. This function handles * binary values. */ static void add_pax_attr_binary(struct archive_string *as, const char *key, const char *value, size_t value_len) { int digits, i, len, next_ten; char tmp[1 + 3 * sizeof(int)]; /* < 3 base-10 digits per byte */ /*- * PAX attributes have the following layout: * <=> */ len = 1 + (int)strlen(key) + 1 + (int)value_len + 1; /* * The field includes the length of the field, so * computing the correct length is tricky. I start by * counting the number of base-10 digits in 'len' and * computing the next higher power of 10. */ next_ten = 1; digits = 0; i = len; while (i > 0) { i = i / 10; digits++; next_ten = next_ten * 10; } /* * For example, if string without the length field is 99 * chars, then adding the 2 digit length "99" will force the * total length past 100, requiring an extra digit. The next * statement adjusts for this effect. */ if (len + digits >= next_ten) digits++; /* Now, we have the right length so we can build the line. */ tmp[sizeof(tmp) - 1] = 0; /* Null-terminate the work area. */ archive_strcat(as, format_int(tmp + sizeof(tmp) - 1, len + digits)); archive_strappend_char(as, ' '); archive_strcat(as, key); archive_strappend_char(as, '='); archive_array_append(as, value, value_len); archive_strappend_char(as, '\n'); } static void archive_write_pax_header_xattr(struct pax *pax, const char *encoded_name, const void *value, size_t value_len) { struct archive_string s; char *encoded_value; if (pax->flags & WRITE_LIBARCHIVE_XATTR) { encoded_value = base64_encode((const char *)value, value_len); if (encoded_name != NULL && encoded_value != NULL) { archive_string_init(&s); archive_strcpy(&s, "LIBARCHIVE.xattr."); archive_strcat(&s, encoded_name); add_pax_attr(&(pax->pax_header), s.s, encoded_value); archive_string_free(&s); } free(encoded_value); } if (pax->flags & WRITE_SCHILY_XATTR) { archive_string_init(&s); archive_strcpy(&s, "SCHILY.xattr."); archive_strcat(&s, encoded_name); add_pax_attr_binary(&(pax->pax_header), s.s, value, value_len); archive_string_free(&s); } } static int archive_write_pax_header_xattrs(struct archive_write *a, struct pax *pax, struct archive_entry *entry) { int i = archive_entry_xattr_reset(entry); while (i--) { const char *name; const void *value; char *url_encoded_name = NULL, *encoded_name = NULL; size_t size; int r; archive_entry_xattr_next(entry, &name, &value, &size); url_encoded_name = url_encode(name); if (url_encoded_name != NULL) { /* Convert narrow-character to UTF-8. */ r = archive_strcpy_l(&(pax->l_url_encoded_name), url_encoded_name, pax->sconv_utf8); free(url_encoded_name); /* Done with this. */ if (r == 0) encoded_name = pax->l_url_encoded_name.s; else if (errno == ENOMEM) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory for Linkname"); return (ARCHIVE_FATAL); } } archive_write_pax_header_xattr(pax, encoded_name, value, size); } return (ARCHIVE_OK); } static int get_entry_hardlink(struct archive_write *a, struct archive_entry *entry, const char **name, size_t *length, struct archive_string_conv *sc) { int r; r = archive_entry_hardlink_l(entry, name, length, sc); if (r != 0) { if (errno == ENOMEM) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory for Linkname"); return (ARCHIVE_FATAL); } return (ARCHIVE_WARN); } return (ARCHIVE_OK); } static int get_entry_pathname(struct archive_write *a, struct archive_entry *entry, const char **name, size_t *length, struct archive_string_conv *sc) { int r; r = archive_entry_pathname_l(entry, name, length, sc); if (r != 0) { if (errno == ENOMEM) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory for Pathname"); return (ARCHIVE_FATAL); } return (ARCHIVE_WARN); } return (ARCHIVE_OK); } static int get_entry_uname(struct archive_write *a, struct archive_entry *entry, const char **name, size_t *length, struct archive_string_conv *sc) { int r; r = archive_entry_uname_l(entry, name, length, sc); if (r != 0) { if (errno == ENOMEM) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory for Uname"); return (ARCHIVE_FATAL); } return (ARCHIVE_WARN); } return (ARCHIVE_OK); } static int get_entry_gname(struct archive_write *a, struct archive_entry *entry, const char **name, size_t *length, struct archive_string_conv *sc) { int r; r = archive_entry_gname_l(entry, name, length, sc); if (r != 0) { if (errno == ENOMEM) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory for Gname"); return (ARCHIVE_FATAL); } return (ARCHIVE_WARN); } return (ARCHIVE_OK); } static int get_entry_symlink(struct archive_write *a, struct archive_entry *entry, const char **name, size_t *length, struct archive_string_conv *sc) { int r; r = archive_entry_symlink_l(entry, name, length, sc); if (r != 0) { if (errno == ENOMEM) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory for Linkname"); return (ARCHIVE_FATAL); } return (ARCHIVE_WARN); } return (ARCHIVE_OK); } /* Add ACL to pax header */ static int add_pax_acl(struct archive_write *a, struct archive_entry *entry, struct pax *pax, int flags) { char *p; const char *attr; int acl_types; acl_types = archive_entry_acl_types(entry); if ((acl_types & ARCHIVE_ENTRY_ACL_TYPE_NFS4) != 0) attr = "SCHILY.acl.ace"; else if ((flags & ARCHIVE_ENTRY_ACL_TYPE_ACCESS) != 0) attr = "SCHILY.acl.access"; else if ((flags & ARCHIVE_ENTRY_ACL_TYPE_DEFAULT) != 0) attr = "SCHILY.acl.default"; else return (ARCHIVE_FATAL); p = archive_entry_acl_to_text_l(entry, NULL, flags, pax->sconv_utf8); if (p == NULL) { if (errno == ENOMEM) { archive_set_error(&a->archive, ENOMEM, "%s %s", "Can't allocate memory for ", attr); return (ARCHIVE_FATAL); } archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "%s %s %s", "Can't translate ", attr, " to UTF-8"); return(ARCHIVE_WARN); } else if (*p != '\0') { add_pax_attr(&(pax->pax_header), attr, p); free(p); } return(ARCHIVE_OK); } /* * TODO: Consider adding 'comment' and 'charset' fields to * archive_entry so that clients can specify them. Also, consider * adding generic key/value tags so clients can add arbitrary * key/value data. * * TODO: Break up this 700-line function!!!! Yowza! */ static int archive_write_pax_header(struct archive_write *a, struct archive_entry *entry_original) { struct archive_entry *entry_main; const char *p; const char *suffix; int need_extension, r, ret; int acl_types; int sparse_count; uint64_t sparse_total, real_size; struct pax *pax; const char *hardlink; const char *path = NULL, *linkpath = NULL; const char *uname = NULL, *gname = NULL; const void *mac_metadata; size_t mac_metadata_size; struct archive_string_conv *sconv; size_t hardlink_length, path_length, linkpath_length; size_t uname_length, gname_length; char paxbuff[512]; char ustarbuff[512]; char ustar_entry_name[256]; char pax_entry_name[256]; char gnu_sparse_name[256]; struct archive_string entry_name; ret = ARCHIVE_OK; need_extension = 0; pax = (struct pax *)a->format_data; /* Sanity check. */ if (archive_entry_pathname(entry_original) == NULL) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "Can't record entry in tar file without pathname"); return (ARCHIVE_FAILED); } /* * Choose a header encoding. */ if (pax->opt_binary) sconv = NULL;/* Binary mode. */ else { /* Header encoding is UTF-8. */ if (pax->sconv_utf8 == NULL) { /* Initialize the string conversion object * we must need */ pax->sconv_utf8 = archive_string_conversion_to_charset( &(a->archive), "UTF-8", 1); if (pax->sconv_utf8 == NULL) /* Couldn't allocate memory */ return (ARCHIVE_FAILED); } sconv = pax->sconv_utf8; } r = get_entry_hardlink(a, entry_original, &hardlink, &hardlink_length, sconv); if (r == ARCHIVE_FATAL) return (r); else if (r != ARCHIVE_OK) { r = get_entry_hardlink(a, entry_original, &hardlink, &hardlink_length, NULL); if (r == ARCHIVE_FATAL) return (r); archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Can't translate linkname '%s' to %s", hardlink, archive_string_conversion_charset_name(sconv)); ret = ARCHIVE_WARN; sconv = NULL;/* The header charset switches to binary mode. */ } /* Make sure this is a type of entry that we can handle here */ if (hardlink == NULL) { switch (archive_entry_filetype(entry_original)) { case AE_IFBLK: case AE_IFCHR: case AE_IFIFO: case AE_IFLNK: case AE_IFREG: break; case AE_IFDIR: { /* * Ensure a trailing '/'. Modify the original * entry so the client sees the change. */ #if defined(_WIN32) && !defined(__CYGWIN__) const wchar_t *wp; wp = archive_entry_pathname_w(entry_original); if (wp != NULL && wp[wcslen(wp) -1] != L'/') { struct archive_wstring ws; archive_string_init(&ws); path_length = wcslen(wp); if (archive_wstring_ensure(&ws, path_length + 2) == NULL) { archive_set_error(&a->archive, ENOMEM, "Can't allocate pax data"); archive_wstring_free(&ws); return(ARCHIVE_FATAL); } /* Should we keep '\' ? */ if (wp[path_length -1] == L'\\') path_length--; archive_wstrncpy(&ws, wp, path_length); archive_wstrappend_wchar(&ws, L'/'); archive_entry_copy_pathname_w( entry_original, ws.s); archive_wstring_free(&ws); p = NULL; } else #endif p = archive_entry_pathname(entry_original); /* * On Windows, this is a backup operation just in * case getting WCS failed. On POSIX, this is a * normal operation. */ if (p != NULL && p[strlen(p) - 1] != '/') { struct archive_string as; archive_string_init(&as); path_length = strlen(p); if (archive_string_ensure(&as, path_length + 2) == NULL) { archive_set_error(&a->archive, ENOMEM, "Can't allocate pax data"); archive_string_free(&as); return(ARCHIVE_FATAL); } #if defined(_WIN32) && !defined(__CYGWIN__) /* NOTE: This might break the pathname * if the current code page is CP932 and * the pathname includes a character '\' * as a part of its multibyte pathname. */ if (p[strlen(p) -1] == '\\') path_length--; else #endif archive_strncpy(&as, p, path_length); archive_strappend_char(&as, '/'); archive_entry_copy_pathname( entry_original, as.s); archive_string_free(&as); } break; } case AE_IFSOCK: archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "tar format cannot archive socket"); return (ARCHIVE_FAILED); default: archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "tar format cannot archive this (type=0%lo)", (unsigned long) archive_entry_filetype(entry_original)); return (ARCHIVE_FAILED); } } /* * If Mac OS metadata blob is here, recurse to write that * as a separate entry. This is really a pretty poor design: * In particular, it doubles the overhead for long filenames. * TODO: Help Apple folks design something better and figure * out how to transition from this legacy format. * * Note that this code is present on every platform; clients * on non-Mac are unlikely to ever provide this data, but * applications that copy entries from one archive to another * should not lose data just because the local filesystem * can't store it. */ mac_metadata = archive_entry_mac_metadata(entry_original, &mac_metadata_size); if (mac_metadata != NULL) { const char *oname; char *name, *bname; size_t name_length; struct archive_entry *extra = archive_entry_new2(&a->archive); oname = archive_entry_pathname(entry_original); name_length = strlen(oname); name = malloc(name_length + 3); if (name == NULL || extra == NULL) { /* XXX error message */ archive_entry_free(extra); free(name); return (ARCHIVE_FAILED); } strcpy(name, oname); /* Find last '/'; strip trailing '/' characters */ bname = strrchr(name, '/'); while (bname != NULL && bname[1] == '\0') { *bname = '\0'; bname = strrchr(name, '/'); } if (bname == NULL) { memmove(name + 2, name, name_length + 1); memmove(name, "._", 2); } else { bname += 1; memmove(bname + 2, bname, strlen(bname) + 1); memmove(bname, "._", 2); } archive_entry_copy_pathname(extra, name); free(name); archive_entry_set_size(extra, mac_metadata_size); archive_entry_set_filetype(extra, AE_IFREG); archive_entry_set_perm(extra, archive_entry_perm(entry_original)); archive_entry_set_mtime(extra, archive_entry_mtime(entry_original), archive_entry_mtime_nsec(entry_original)); archive_entry_set_gid(extra, archive_entry_gid(entry_original)); archive_entry_set_gname(extra, archive_entry_gname(entry_original)); archive_entry_set_uid(extra, archive_entry_uid(entry_original)); archive_entry_set_uname(extra, archive_entry_uname(entry_original)); /* Recurse to write the special copyfile entry. */ r = archive_write_pax_header(a, extra); archive_entry_free(extra); if (r < ARCHIVE_WARN) return (r); if (r < ret) ret = r; r = (int)archive_write_pax_data(a, mac_metadata, mac_metadata_size); if (r < ARCHIVE_WARN) return (r); if (r < ret) ret = r; r = archive_write_pax_finish_entry(a); if (r < ARCHIVE_WARN) return (r); if (r < ret) ret = r; } /* Copy entry so we can modify it as needed. */ #if defined(_WIN32) && !defined(__CYGWIN__) /* Make sure the path separators in pathname, hardlink and symlink * are all slash '/', not the Windows path separator '\'. */ entry_main = __la_win_entry_in_posix_pathseparator(entry_original); if (entry_main == entry_original) entry_main = archive_entry_clone(entry_original); #else entry_main = archive_entry_clone(entry_original); #endif if (entry_main == NULL) { archive_set_error(&a->archive, ENOMEM, "Can't allocate pax data"); return(ARCHIVE_FATAL); } archive_string_empty(&(pax->pax_header)); /* Blank our work area. */ archive_string_empty(&(pax->sparse_map)); sparse_total = 0; sparse_list_clear(pax); if (hardlink == NULL && archive_entry_filetype(entry_main) == AE_IFREG) sparse_count = archive_entry_sparse_reset(entry_main); else sparse_count = 0; if (sparse_count) { int64_t offset, length, last_offset = 0; /* Get the last entry of sparse block. */ while (archive_entry_sparse_next( entry_main, &offset, &length) == ARCHIVE_OK) last_offset = offset + length; /* If the last sparse block does not reach the end of file, * We have to add a empty sparse block as the last entry to * manage storing file data. */ if (last_offset < archive_entry_size(entry_main)) archive_entry_sparse_add_entry(entry_main, archive_entry_size(entry_main), 0); sparse_count = archive_entry_sparse_reset(entry_main); } /* * First, check the name fields and see if any of them * require binary coding. If any of them does, then all of * them do. */ r = get_entry_pathname(a, entry_main, &path, &path_length, sconv); if (r == ARCHIVE_FATAL) return (r); else if (r != ARCHIVE_OK) { r = get_entry_pathname(a, entry_main, &path, &path_length, NULL); if (r == ARCHIVE_FATAL) return (r); archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Can't translate pathname '%s' to %s", path, archive_string_conversion_charset_name(sconv)); ret = ARCHIVE_WARN; sconv = NULL;/* The header charset switches to binary mode. */ } r = get_entry_uname(a, entry_main, &uname, &uname_length, sconv); if (r == ARCHIVE_FATAL) return (r); else if (r != ARCHIVE_OK) { r = get_entry_uname(a, entry_main, &uname, &uname_length, NULL); if (r == ARCHIVE_FATAL) return (r); archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Can't translate uname '%s' to %s", uname, archive_string_conversion_charset_name(sconv)); ret = ARCHIVE_WARN; sconv = NULL;/* The header charset switches to binary mode. */ } r = get_entry_gname(a, entry_main, &gname, &gname_length, sconv); if (r == ARCHIVE_FATAL) return (r); else if (r != ARCHIVE_OK) { r = get_entry_gname(a, entry_main, &gname, &gname_length, NULL); if (r == ARCHIVE_FATAL) return (r); archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Can't translate gname '%s' to %s", gname, archive_string_conversion_charset_name(sconv)); ret = ARCHIVE_WARN; sconv = NULL;/* The header charset switches to binary mode. */ } linkpath = hardlink; linkpath_length = hardlink_length; if (linkpath == NULL) { r = get_entry_symlink(a, entry_main, &linkpath, &linkpath_length, sconv); if (r == ARCHIVE_FATAL) return (r); else if (r != ARCHIVE_OK) { r = get_entry_symlink(a, entry_main, &linkpath, &linkpath_length, NULL); if (r == ARCHIVE_FATAL) return (r); archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, "Can't translate linkname '%s' to %s", linkpath, archive_string_conversion_charset_name(sconv)); ret = ARCHIVE_WARN; sconv = NULL; } } /* If any string conversions failed, get all attributes * in binary-mode. */ if (sconv == NULL && !pax->opt_binary) { if (hardlink != NULL) { r = get_entry_hardlink(a, entry_main, &hardlink, &hardlink_length, NULL); if (r == ARCHIVE_FATAL) return (r); linkpath = hardlink; linkpath_length = hardlink_length; } r = get_entry_pathname(a, entry_main, &path, &path_length, NULL); if (r == ARCHIVE_FATAL) return (r); r = get_entry_uname(a, entry_main, &uname, &uname_length, NULL); if (r == ARCHIVE_FATAL) return (r); r = get_entry_gname(a, entry_main, &gname, &gname_length, NULL); if (r == ARCHIVE_FATAL) return (r); } /* Store the header encoding first, to be nice to readers. */ if (sconv == NULL) add_pax_attr(&(pax->pax_header), "hdrcharset", "BINARY"); /* * If name is too long, or has non-ASCII characters, add * 'path' to pax extended attrs. (Note that an unconvertible * name must have non-ASCII characters.) */ if (has_non_ASCII(path)) { /* We have non-ASCII characters. */ add_pax_attr(&(pax->pax_header), "path", path); archive_entry_set_pathname(entry_main, build_ustar_entry_name(ustar_entry_name, path, path_length, NULL)); need_extension = 1; } else { /* We have an all-ASCII path; we'd like to just store * it in the ustar header if it will fit. Yes, this * duplicates some of the logic in * archive_write_set_format_ustar.c */ if (path_length <= 100) { /* Fits in the old 100-char tar name field. */ } else { /* Find largest suffix that will fit. */ /* Note: strlen() > 100, so strlen() - 100 - 1 >= 0 */ suffix = strchr(path + path_length - 100 - 1, '/'); /* Don't attempt an empty prefix. */ if (suffix == path) suffix = strchr(suffix + 1, '/'); /* We can put it in the ustar header if it's * all ASCII and it's either <= 100 characters * or can be split at a '/' into a prefix <= * 155 chars and a suffix <= 100 chars. (Note * the strchr() above will return NULL exactly * when the path can't be split.) */ if (suffix == NULL /* Suffix > 100 chars. */ || suffix[1] == '\0' /* empty suffix */ || suffix - path > 155) /* Prefix > 155 chars */ { add_pax_attr(&(pax->pax_header), "path", path); archive_entry_set_pathname(entry_main, build_ustar_entry_name(ustar_entry_name, path, path_length, NULL)); need_extension = 1; } } } if (linkpath != NULL) { /* If link name is too long or has non-ASCII characters, add * 'linkpath' to pax extended attrs. */ if (linkpath_length > 100 || has_non_ASCII(linkpath)) { add_pax_attr(&(pax->pax_header), "linkpath", linkpath); if (linkpath_length > 100) { if (hardlink != NULL) archive_entry_set_hardlink(entry_main, "././@LongHardLink"); else archive_entry_set_symlink(entry_main, "././@LongSymLink"); } need_extension = 1; } } /* Save a pathname since it will be renamed if `entry_main` has * sparse blocks. */ archive_string_init(&entry_name); archive_strcpy(&entry_name, archive_entry_pathname(entry_main)); /* If file size is too large, add 'size' to pax extended attrs. */ if (archive_entry_size(entry_main) >= (((int64_t)1) << 33)) { add_pax_attr_int(&(pax->pax_header), "size", archive_entry_size(entry_main)); need_extension = 1; } /* If numeric GID is too large, add 'gid' to pax extended attrs. */ if ((unsigned int)archive_entry_gid(entry_main) >= (1 << 18)) { add_pax_attr_int(&(pax->pax_header), "gid", archive_entry_gid(entry_main)); need_extension = 1; } /* If group name is too large or has non-ASCII characters, add * 'gname' to pax extended attrs. */ if (gname != NULL) { if (gname_length > 31 || has_non_ASCII(gname)) { add_pax_attr(&(pax->pax_header), "gname", gname); need_extension = 1; } } /* If numeric UID is too large, add 'uid' to pax extended attrs. */ if ((unsigned int)archive_entry_uid(entry_main) >= (1 << 18)) { add_pax_attr_int(&(pax->pax_header), "uid", archive_entry_uid(entry_main)); need_extension = 1; } /* Add 'uname' to pax extended attrs if necessary. */ if (uname != NULL) { if (uname_length > 31 || has_non_ASCII(uname)) { add_pax_attr(&(pax->pax_header), "uname", uname); need_extension = 1; } } /* * POSIX/SUSv3 doesn't provide a standard key for large device * numbers. I use the same keys here that Joerg Schilling * used for 'star.' (Which, somewhat confusingly, are called * "devXXX" even though they code "rdev" values.) No doubt, * other implementations use other keys. Note that there's no * reason we can't write the same information into a number of * different keys. * * Of course, this is only needed for block or char device entries. */ if (archive_entry_filetype(entry_main) == AE_IFBLK || archive_entry_filetype(entry_main) == AE_IFCHR) { /* * If rdevmajor is too large, add 'SCHILY.devmajor' to * extended attributes. */ int rdevmajor, rdevminor; rdevmajor = archive_entry_rdevmajor(entry_main); rdevminor = archive_entry_rdevminor(entry_main); if (rdevmajor >= (1 << 18)) { add_pax_attr_int(&(pax->pax_header), "SCHILY.devmajor", rdevmajor); /* * Non-strict formatting below means we don't * have to truncate here. Not truncating improves * the chance that some more modern tar archivers * (such as GNU tar 1.13) can restore the full * value even if they don't understand the pax * extended attributes. See my rant below about * file size fields for additional details. */ /* archive_entry_set_rdevmajor(entry_main, rdevmajor & ((1 << 18) - 1)); */ need_extension = 1; } /* * If devminor is too large, add 'SCHILY.devminor' to * extended attributes. */ if (rdevminor >= (1 << 18)) { add_pax_attr_int(&(pax->pax_header), "SCHILY.devminor", rdevminor); /* Truncation is not necessary here, either. */ /* archive_entry_set_rdevminor(entry_main, rdevminor & ((1 << 18) - 1)); */ need_extension = 1; } } /* * Technically, the mtime field in the ustar header can * support 33 bits, but many platforms use signed 32-bit time * values. The cutoff of 0x7fffffff here is a compromise. * Yes, this check is duplicated just below; this helps to * avoid writing an mtime attribute just to handle a * high-resolution timestamp in "restricted pax" mode. */ if (!need_extension && ((archive_entry_mtime(entry_main) < 0) || (archive_entry_mtime(entry_main) >= 0x7fffffff))) need_extension = 1; /* I use a star-compatible file flag attribute. */ p = archive_entry_fflags_text(entry_main); if (!need_extension && p != NULL && *p != '\0') need_extension = 1; /* If there are extended attributes, we need an extension */ if (!need_extension && archive_entry_xattr_count(entry_original) > 0) need_extension = 1; /* If there are sparse info, we need an extension */ if (!need_extension && sparse_count > 0) need_extension = 1; acl_types = archive_entry_acl_types(entry_original); /* If there are any ACL entries, we need an extension */ if (!need_extension && acl_types != 0) need_extension = 1; /* * Libarchive used to include these in extended headers for * restricted pax format, but that confused people who * expected ustar-like time semantics. So now we only include * them in full pax format. */ if (a->archive.archive_format != ARCHIVE_FORMAT_TAR_PAX_RESTRICTED) { if (archive_entry_ctime(entry_main) != 0 || archive_entry_ctime_nsec(entry_main) != 0) add_pax_attr_time(&(pax->pax_header), "ctime", archive_entry_ctime(entry_main), archive_entry_ctime_nsec(entry_main)); if (archive_entry_atime(entry_main) != 0 || archive_entry_atime_nsec(entry_main) != 0) add_pax_attr_time(&(pax->pax_header), "atime", archive_entry_atime(entry_main), archive_entry_atime_nsec(entry_main)); /* Store birth/creationtime only if it's earlier than mtime */ if (archive_entry_birthtime_is_set(entry_main) && archive_entry_birthtime(entry_main) < archive_entry_mtime(entry_main)) add_pax_attr_time(&(pax->pax_header), "LIBARCHIVE.creationtime", archive_entry_birthtime(entry_main), archive_entry_birthtime_nsec(entry_main)); } /* * The following items are handled differently in "pax * restricted" format. In particular, in "pax restricted" * format they won't be added unless need_extension is * already set (we're already generating an extended header, so * may as well include these). */ if (a->archive.archive_format != ARCHIVE_FORMAT_TAR_PAX_RESTRICTED || need_extension) { if (archive_entry_mtime(entry_main) < 0 || archive_entry_mtime(entry_main) >= 0x7fffffff || archive_entry_mtime_nsec(entry_main) != 0) add_pax_attr_time(&(pax->pax_header), "mtime", archive_entry_mtime(entry_main), archive_entry_mtime_nsec(entry_main)); /* I use a star-compatible file flag attribute. */ p = archive_entry_fflags_text(entry_main); if (p != NULL && *p != '\0') add_pax_attr(&(pax->pax_header), "SCHILY.fflags", p); /* I use star-compatible ACL attributes. */ if ((acl_types & ARCHIVE_ENTRY_ACL_TYPE_NFS4) != 0) { ret = add_pax_acl(a, entry_original, pax, ARCHIVE_ENTRY_ACL_STYLE_EXTRA_ID | ARCHIVE_ENTRY_ACL_STYLE_SEPARATOR_COMMA | ARCHIVE_ENTRY_ACL_STYLE_COMPACT); if (ret == ARCHIVE_FATAL) return (ARCHIVE_FATAL); } if (acl_types & ARCHIVE_ENTRY_ACL_TYPE_ACCESS) { ret = add_pax_acl(a, entry_original, pax, ARCHIVE_ENTRY_ACL_TYPE_ACCESS | ARCHIVE_ENTRY_ACL_STYLE_EXTRA_ID | ARCHIVE_ENTRY_ACL_STYLE_SEPARATOR_COMMA); if (ret == ARCHIVE_FATAL) return (ARCHIVE_FATAL); } if (acl_types & ARCHIVE_ENTRY_ACL_TYPE_DEFAULT) { ret = add_pax_acl(a, entry_original, pax, ARCHIVE_ENTRY_ACL_TYPE_DEFAULT | ARCHIVE_ENTRY_ACL_STYLE_EXTRA_ID | ARCHIVE_ENTRY_ACL_STYLE_SEPARATOR_COMMA); if (ret == ARCHIVE_FATAL) return (ARCHIVE_FATAL); } /* We use GNU-tar-compatible sparse attributes. */ if (sparse_count > 0) { int64_t soffset, slength; add_pax_attr_int(&(pax->pax_header), "GNU.sparse.major", 1); add_pax_attr_int(&(pax->pax_header), "GNU.sparse.minor", 0); /* * Make sure to store the original path, since * truncation to ustar limit happened already. */ add_pax_attr(&(pax->pax_header), "GNU.sparse.name", path); add_pax_attr_int(&(pax->pax_header), "GNU.sparse.realsize", archive_entry_size(entry_main)); /* Rename the file name which will be used for * ustar header to a special name, which GNU * PAX Format 1.0 requires */ archive_entry_set_pathname(entry_main, build_gnu_sparse_name(gnu_sparse_name, entry_name.s)); /* * - Make a sparse map, which will precede a file data. * - Get the total size of available data of sparse. */ archive_string_sprintf(&(pax->sparse_map), "%d\n", sparse_count); while (archive_entry_sparse_next(entry_main, &soffset, &slength) == ARCHIVE_OK) { archive_string_sprintf(&(pax->sparse_map), "%jd\n%jd\n", (intmax_t)soffset, (intmax_t)slength); sparse_total += slength; if (sparse_list_add(pax, soffset, slength) != ARCHIVE_OK) { archive_set_error(&a->archive, ENOMEM, "Can't allocate memory"); archive_entry_free(entry_main); archive_string_free(&entry_name); return (ARCHIVE_FATAL); } } } /* Store extended attributes */ if (archive_write_pax_header_xattrs(a, pax, entry_original) == ARCHIVE_FATAL) { archive_entry_free(entry_main); archive_string_free(&entry_name); return (ARCHIVE_FATAL); } } /* Only regular files have data. */ if (archive_entry_filetype(entry_main) != AE_IFREG) archive_entry_set_size(entry_main, 0); /* * Pax-restricted does not store data for hardlinks, in order * to improve compatibility with ustar. */ if (a->archive.archive_format != ARCHIVE_FORMAT_TAR_PAX_INTERCHANGE && hardlink != NULL) archive_entry_set_size(entry_main, 0); /* * XXX Full pax interchange format does permit a hardlink * entry to have data associated with it. I'm not supporting * that here because the client expects me to tell them whether * or not this format expects data for hardlinks. If I * don't check here, then every pax archive will end up with * duplicated data for hardlinks. Someday, there may be * need to select this behavior, in which case the following * will need to be revisited. XXX */ if (hardlink != NULL) archive_entry_set_size(entry_main, 0); /* Save a real file size. */ real_size = archive_entry_size(entry_main); /* * Overwrite a file size by the total size of sparse blocks and * the size of sparse map info. That file size is the length of * the data, which we will exactly store into an archive file. */ if (archive_strlen(&(pax->sparse_map))) { size_t mapsize = archive_strlen(&(pax->sparse_map)); pax->sparse_map_padding = 0x1ff & (-(ssize_t)mapsize); archive_entry_set_size(entry_main, mapsize + pax->sparse_map_padding + sparse_total); } /* Format 'ustar' header for main entry. * * The trouble with file size: If the reader can't understand * the file size, they may not be able to locate the next * entry and the rest of the archive is toast. Pax-compliant * readers are supposed to ignore the file size in the main * header, so the question becomes how to maximize portability * for readers that don't support pax attribute extensions. * For maximum compatibility, I permit numeric extensions in * the main header so that the file size stored will always be * correct, even if it's in a format that only some * implementations understand. The technique used here is: * * a) If possible, follow the standard exactly. This handles * files up to 8 gigabytes minus 1. * * b) If that fails, try octal but omit the field terminator. * That handles files up to 64 gigabytes minus 1. * * c) Otherwise, use base-256 extensions. That handles files * up to 2^63 in this implementation, with the potential to * go up to 2^94. That should hold us for a while. ;-) * * The non-strict formatter uses similar logic for other * numeric fields, though they're less critical. */ if (__archive_write_format_header_ustar(a, ustarbuff, entry_main, -1, 0, NULL) == ARCHIVE_FATAL) return (ARCHIVE_FATAL); /* If we built any extended attributes, write that entry first. */ if (archive_strlen(&(pax->pax_header)) > 0) { struct archive_entry *pax_attr_entry; time_t s; int64_t uid, gid; int mode; pax_attr_entry = archive_entry_new2(&a->archive); p = entry_name.s; archive_entry_set_pathname(pax_attr_entry, build_pax_attribute_name(pax_entry_name, p)); archive_entry_set_size(pax_attr_entry, archive_strlen(&(pax->pax_header))); /* Copy uid/gid (but clip to ustar limits). */ uid = archive_entry_uid(entry_main); if (uid >= 1 << 18) uid = (1 << 18) - 1; archive_entry_set_uid(pax_attr_entry, uid); gid = archive_entry_gid(entry_main); if (gid >= 1 << 18) gid = (1 << 18) - 1; archive_entry_set_gid(pax_attr_entry, gid); /* Copy mode over (but not setuid/setgid bits) */ mode = archive_entry_mode(entry_main); #ifdef S_ISUID mode &= ~S_ISUID; #endif #ifdef S_ISGID mode &= ~S_ISGID; #endif #ifdef S_ISVTX mode &= ~S_ISVTX; #endif archive_entry_set_mode(pax_attr_entry, mode); /* Copy uname/gname. */ archive_entry_set_uname(pax_attr_entry, archive_entry_uname(entry_main)); archive_entry_set_gname(pax_attr_entry, archive_entry_gname(entry_main)); /* Copy mtime, but clip to ustar limits. */ s = archive_entry_mtime(entry_main); if (s < 0) { s = 0; } if (s >= 0x7fffffff) { s = 0x7fffffff; } archive_entry_set_mtime(pax_attr_entry, s, 0); /* Standard ustar doesn't support atime. */ archive_entry_set_atime(pax_attr_entry, 0, 0); /* Standard ustar doesn't support ctime. */ archive_entry_set_ctime(pax_attr_entry, 0, 0); r = __archive_write_format_header_ustar(a, paxbuff, pax_attr_entry, 'x', 1, NULL); archive_entry_free(pax_attr_entry); /* Note that the 'x' header shouldn't ever fail to format */ if (r < ARCHIVE_WARN) { archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, "archive_write_pax_header: " "'x' header failed?! This can't happen.\n"); return (ARCHIVE_FATAL); } else if (r < ret) ret = r; r = __archive_write_output(a, paxbuff, 512); if (r != ARCHIVE_OK) { sparse_list_clear(pax); pax->entry_bytes_remaining = 0; pax->entry_padding = 0; return (ARCHIVE_FATAL); } pax->entry_bytes_remaining = archive_strlen(&(pax->pax_header)); pax->entry_padding = 0x1ff & (-(int64_t)pax->entry_bytes_remaining); r = __archive_write_output(a, pax->pax_header.s, archive_strlen(&(pax->pax_header))); if (r != ARCHIVE_OK) { /* If a write fails, we're pretty much toast. */ return (ARCHIVE_FATAL); } /* Pad out the end of the entry. */ r = __archive_write_nulls(a, (size_t)pax->entry_padding); if (r != ARCHIVE_OK) { /* If a write fails, we're pretty much toast. */ return (ARCHIVE_FATAL); } pax->entry_bytes_remaining = pax->entry_padding = 0; } /* Write the header for main entry. */ r = __archive_write_output(a, ustarbuff, 512); if (r != ARCHIVE_OK) return (r); /* * Inform the client of the on-disk size we're using, so * they can avoid unnecessarily writing a body for something * that we're just going to ignore. */ archive_entry_set_size(entry_original, real_size); if (pax->sparse_list == NULL && real_size > 0) { /* This is not a sparse file but we handle its data as * a sparse block. */ sparse_list_add(pax, 0, real_size); sparse_total = real_size; } pax->entry_padding = 0x1ff & (-(int64_t)sparse_total); archive_entry_free(entry_main); archive_string_free(&entry_name); return (ret); } /* * We need a valid name for the regular 'ustar' entry. This routine * tries to hack something more-or-less reasonable. * * The approach here tries to preserve leading dir names. We do so by * working with four sections: * 1) "prefix" directory names, * 2) "suffix" directory names, * 3) inserted dir name (optional), * 4) filename. * * These sections must satisfy the following requirements: * * Parts 1 & 2 together form an initial portion of the dir name. * * Part 3 is specified by the caller. (It should not contain a leading * or trailing '/'.) * * Part 4 forms an initial portion of the base filename. * * The filename must be <= 99 chars to fit the ustar 'name' field. * * Parts 2, 3, 4 together must be <= 99 chars to fit the ustar 'name' fld. * * Part 1 must be <= 155 chars to fit the ustar 'prefix' field. * * If the original name ends in a '/', the new name must also end in a '/' * * Trailing '/.' sequences may be stripped. * * Note: Recall that the ustar format does not store the '/' separating * parts 1 & 2, but does store the '/' separating parts 2 & 3. */ static char * build_ustar_entry_name(char *dest, const char *src, size_t src_length, const char *insert) { const char *prefix, *prefix_end; const char *suffix, *suffix_end; const char *filename, *filename_end; char *p; int need_slash = 0; /* Was there a trailing slash? */ size_t suffix_length = 99; size_t insert_length; /* Length of additional dir element to be added. */ if (insert == NULL) insert_length = 0; else /* +2 here allows for '/' before and after the insert. */ insert_length = strlen(insert) + 2; /* Step 0: Quick bailout in a common case. */ if (src_length < 100 && insert == NULL) { strncpy(dest, src, src_length); dest[src_length] = '\0'; return (dest); } /* Step 1: Locate filename and enforce the length restriction. */ filename_end = src + src_length; /* Remove trailing '/' chars and '/.' pairs. */ for (;;) { if (filename_end > src && filename_end[-1] == '/') { filename_end --; need_slash = 1; /* Remember to restore trailing '/'. */ continue; } if (filename_end > src + 1 && filename_end[-1] == '.' && filename_end[-2] == '/') { filename_end -= 2; need_slash = 1; /* "foo/." will become "foo/" */ continue; } break; } if (need_slash) suffix_length--; /* Find start of filename. */ filename = filename_end - 1; while ((filename > src) && (*filename != '/')) filename --; if ((*filename == '/') && (filename < filename_end - 1)) filename ++; /* Adjust filename_end so that filename + insert fits in 99 chars. */ suffix_length -= insert_length; if (filename_end > filename + suffix_length) filename_end = filename + suffix_length; /* Calculate max size for "suffix" section (#3 above). */ suffix_length -= filename_end - filename; /* Step 2: Locate the "prefix" section of the dirname, including * trailing '/'. */ prefix = src; prefix_end = prefix + 155; if (prefix_end > filename) prefix_end = filename; while (prefix_end > prefix && *prefix_end != '/') prefix_end--; if ((prefix_end < filename) && (*prefix_end == '/')) prefix_end++; /* Step 3: Locate the "suffix" section of the dirname, * including trailing '/'. */ suffix = prefix_end; suffix_end = suffix + suffix_length; /* Enforce limit. */ if (suffix_end > filename) suffix_end = filename; if (suffix_end < suffix) suffix_end = suffix; while (suffix_end > suffix && *suffix_end != '/') suffix_end--; if ((suffix_end < filename) && (*suffix_end == '/')) suffix_end++; /* Step 4: Build the new name. */ /* The OpenBSD strlcpy function is safer, but less portable. */ /* Rather than maintain two versions, just use the strncpy version. */ p = dest; if (prefix_end > prefix) { strncpy(p, prefix, prefix_end - prefix); p += prefix_end - prefix; } if (suffix_end > suffix) { strncpy(p, suffix, suffix_end - suffix); p += suffix_end - suffix; } if (insert != NULL) { /* Note: assume insert does not have leading or trailing '/' */ strcpy(p, insert); p += strlen(insert); *p++ = '/'; } strncpy(p, filename, filename_end - filename); p += filename_end - filename; if (need_slash) *p++ = '/'; *p = '\0'; return (dest); } /* * The ustar header for the pax extended attributes must have a * reasonable name: SUSv3 requires 'dirname'/PaxHeader.'pid'/'filename' * where 'pid' is the PID of the archiving process. Unfortunately, * that makes testing a pain since the output varies for each run, * so I'm sticking with the simpler 'dirname'/PaxHeader/'filename' * for now. (Someday, I'll make this settable. Then I can use the * SUS recommendation as default and test harnesses can override it * to get predictable results.) * * Joerg Schilling has argued that this is unnecessary because, in * practice, if the pax extended attributes get extracted as regular * files, no one is going to bother reading those attributes to * manually restore them. Based on this, 'star' uses * /tmp/PaxHeader/'basename' as the ustar header name. This is a * tempting argument, in part because it's simpler than the SUSv3 * recommendation, but I'm not entirely convinced. I'm also * uncomfortable with the fact that "/tmp" is a Unix-ism. * * The following routine leverages build_ustar_entry_name() above and * so is simpler than you might think. It just needs to provide the * additional path element and handle a few pathological cases). */ static char * build_pax_attribute_name(char *dest, const char *src) { char buff[64]; const char *p; /* Handle the null filename case. */ if (src == NULL || *src == '\0') { strcpy(dest, "PaxHeader/blank"); return (dest); } /* Prune final '/' and other unwanted final elements. */ p = src + strlen(src); for (;;) { /* Ends in "/", remove the '/' */ if (p > src && p[-1] == '/') { --p; continue; } /* Ends in "/.", remove the '.' */ if (p > src + 1 && p[-1] == '.' && p[-2] == '/') { --p; continue; } break; } /* Pathological case: After above, there was nothing left. * This includes "/." "/./." "/.//./." etc. */ if (p == src) { strcpy(dest, "/PaxHeader/rootdir"); return (dest); } /* Convert unadorned "." into a suitable filename. */ if (*src == '.' && p == src + 1) { strcpy(dest, "PaxHeader/currentdir"); return (dest); } /* * TODO: Push this string into the 'pax' structure to avoid * recomputing it every time. That will also open the door * to having clients override it. */ #if HAVE_GETPID && 0 /* Disable this for now; see above comment. */ sprintf(buff, "PaxHeader.%d", getpid()); #else /* If the platform can't fetch the pid, don't include it. */ strcpy(buff, "PaxHeader"); #endif /* General case: build a ustar-compatible name adding * "/PaxHeader/". */ build_ustar_entry_name(dest, src, p - src, buff); return (dest); } /* * GNU PAX Format 1.0 requires the special name, which pattern is: * /GNUSparseFile./ * - * Since reproducable archives are more important, use 0 as pid. + * Since reproducible archives are more important, use 0 as pid. * * This function is used for only Sparse file, a file type of which * is regular file. */ static char * build_gnu_sparse_name(char *dest, const char *src) { const char *p; /* Handle the null filename case. */ if (src == NULL || *src == '\0') { strcpy(dest, "GNUSparseFile/blank"); return (dest); } /* Prune final '/' and other unwanted final elements. */ p = src + strlen(src); for (;;) { /* Ends in "/", remove the '/' */ if (p > src && p[-1] == '/') { --p; continue; } /* Ends in "/.", remove the '.' */ if (p > src + 1 && p[-1] == '.' && p[-2] == '/') { --p; continue; } break; } /* General case: build a ustar-compatible name adding * "/GNUSparseFile/". */ build_ustar_entry_name(dest, src, p - src, "GNUSparseFile.0"); return (dest); } /* Write two null blocks for the end of archive */ static int archive_write_pax_close(struct archive_write *a) { return (__archive_write_nulls(a, 512 * 2)); } static int archive_write_pax_free(struct archive_write *a) { struct pax *pax; pax = (struct pax *)a->format_data; if (pax == NULL) return (ARCHIVE_OK); archive_string_free(&pax->pax_header); archive_string_free(&pax->sparse_map); archive_string_free(&pax->l_url_encoded_name); sparse_list_clear(pax); free(pax); a->format_data = NULL; return (ARCHIVE_OK); } static int archive_write_pax_finish_entry(struct archive_write *a) { struct pax *pax; uint64_t remaining; int ret; pax = (struct pax *)a->format_data; remaining = pax->entry_bytes_remaining; if (remaining == 0) { while (pax->sparse_list) { struct sparse_block *sb; if (!pax->sparse_list->is_hole) remaining += pax->sparse_list->remaining; sb = pax->sparse_list->next; free(pax->sparse_list); pax->sparse_list = sb; } } ret = __archive_write_nulls(a, (size_t)(remaining + pax->entry_padding)); pax->entry_bytes_remaining = pax->entry_padding = 0; return (ret); } static ssize_t archive_write_pax_data(struct archive_write *a, const void *buff, size_t s) { struct pax *pax; size_t ws; size_t total; int ret; pax = (struct pax *)a->format_data; /* * According to GNU PAX format 1.0, write a sparse map * before the body. */ if (archive_strlen(&(pax->sparse_map))) { ret = __archive_write_output(a, pax->sparse_map.s, archive_strlen(&(pax->sparse_map))); if (ret != ARCHIVE_OK) return (ret); ret = __archive_write_nulls(a, pax->sparse_map_padding); if (ret != ARCHIVE_OK) return (ret); archive_string_empty(&(pax->sparse_map)); } total = 0; while (total < s) { const unsigned char *p; while (pax->sparse_list != NULL && pax->sparse_list->remaining == 0) { struct sparse_block *sb = pax->sparse_list->next; free(pax->sparse_list); pax->sparse_list = sb; } if (pax->sparse_list == NULL) return (total); p = ((const unsigned char *)buff) + total; ws = s - total; if (ws > pax->sparse_list->remaining) ws = (size_t)pax->sparse_list->remaining; if (pax->sparse_list->is_hole) { /* Current block is hole thus we do not write * the body. */ pax->sparse_list->remaining -= ws; total += ws; continue; } ret = __archive_write_output(a, p, ws); pax->sparse_list->remaining -= ws; total += ws; if (ret != ARCHIVE_OK) return (ret); } return (total); } static int has_non_ASCII(const char *_p) { const unsigned char *p = (const unsigned char *)_p; if (p == NULL) return (1); while (*p != '\0' && *p < 128) p++; return (*p != '\0'); } /* * Used by extended attribute support; encodes the name * so that there will be no '=' characters in the result. */ static char * url_encode(const char *in) { const char *s; char *d; int out_len = 0; char *out; for (s = in; *s != '\0'; s++) { if (*s < 33 || *s > 126 || *s == '%' || *s == '=') out_len += 3; else out_len++; } out = (char *)malloc(out_len + 1); if (out == NULL) return (NULL); for (s = in, d = out; *s != '\0'; s++) { /* encode any non-printable ASCII character or '%' or '=' */ if (*s < 33 || *s > 126 || *s == '%' || *s == '=') { /* URL encoding is '%' followed by two hex digits */ *d++ = '%'; *d++ = "0123456789ABCDEF"[0x0f & (*s >> 4)]; *d++ = "0123456789ABCDEF"[0x0f & *s]; } else { *d++ = *s; } } *d = '\0'; return (out); } /* * Encode a sequence of bytes into a C string using base-64 encoding. * * Returns a null-terminated C string allocated with malloc(); caller * is responsible for freeing the result. */ static char * base64_encode(const char *s, size_t len) { static const char digits[64] = { 'A','B','C','D','E','F','G','H','I','J','K','L','M','N','O', 'P','Q','R','S','T','U','V','W','X','Y','Z','a','b','c','d', 'e','f','g','h','i','j','k','l','m','n','o','p','q','r','s', 't','u','v','w','x','y','z','0','1','2','3','4','5','6','7', '8','9','+','/' }; int v; char *d, *out; /* 3 bytes becomes 4 chars, but round up and allow for trailing NUL */ out = (char *)malloc((len * 4 + 2) / 3 + 1); if (out == NULL) return (NULL); d = out; /* Convert each group of 3 bytes into 4 characters. */ while (len >= 3) { v = (((int)s[0] << 16) & 0xff0000) | (((int)s[1] << 8) & 0xff00) | (((int)s[2]) & 0x00ff); s += 3; len -= 3; *d++ = digits[(v >> 18) & 0x3f]; *d++ = digits[(v >> 12) & 0x3f]; *d++ = digits[(v >> 6) & 0x3f]; *d++ = digits[(v) & 0x3f]; } /* Handle final group of 1 byte (2 chars) or 2 bytes (3 chars). */ switch (len) { case 0: break; case 1: v = (((int)s[0] << 16) & 0xff0000); *d++ = digits[(v >> 18) & 0x3f]; *d++ = digits[(v >> 12) & 0x3f]; break; case 2: v = (((int)s[0] << 16) & 0xff0000) | (((int)s[1] << 8) & 0xff00); *d++ = digits[(v >> 18) & 0x3f]; *d++ = digits[(v >> 12) & 0x3f]; *d++ = digits[(v >> 6) & 0x3f]; break; } /* Add trailing NUL character so output is a valid C string. */ *d = '\0'; return (out); } static void sparse_list_clear(struct pax *pax) { while (pax->sparse_list != NULL) { struct sparse_block *sb = pax->sparse_list; pax->sparse_list = sb->next; free(sb); } pax->sparse_tail = NULL; } static int _sparse_list_add_block(struct pax *pax, int64_t offset, int64_t length, int is_hole) { struct sparse_block *sb; sb = (struct sparse_block *)malloc(sizeof(*sb)); if (sb == NULL) return (ARCHIVE_FATAL); sb->next = NULL; sb->is_hole = is_hole; sb->offset = offset; sb->remaining = length; if (pax->sparse_list == NULL || pax->sparse_tail == NULL) pax->sparse_list = pax->sparse_tail = sb; else { pax->sparse_tail->next = sb; pax->sparse_tail = sb; } return (ARCHIVE_OK); } static int sparse_list_add(struct pax *pax, int64_t offset, int64_t length) { int64_t last_offset; int r; if (pax->sparse_tail == NULL) last_offset = 0; else { last_offset = pax->sparse_tail->offset + pax->sparse_tail->remaining; } if (last_offset < offset) { /* Add a hole block. */ r = _sparse_list_add_block(pax, last_offset, offset - last_offset, 1); if (r != ARCHIVE_OK) return (r); } /* Add data block. */ return (_sparse_list_add_block(pax, offset, length, 0)); } Index: stable/10/contrib/libarchive/libarchive/test/test_archive_write_add_filter_by_name.c =================================================================== --- stable/10/contrib/libarchive/libarchive/test/test_archive_write_add_filter_by_name.c (revision 324417) +++ stable/10/contrib/libarchive/libarchive/test/test_archive_write_add_filter_by_name.c (revision 324418) @@ -1,209 +1,214 @@ /*- * Copyright (c) 2012 Michihiro NAKAJIMA * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "test.h" __FBSDID("$FreeBSD$"); static void test_filter_by_name(const char *filter_name, int filter_code, int (*can_filter_prog)(void)) { struct archive_entry *ae; struct archive *a; size_t used; size_t buffsize = 1024 * 128; char *buff; int r; assert((buff = calloc(1, buffsize)) != NULL); if (buff == NULL) return; /* Create a new archive in memory. */ assert((a = archive_write_new()) != NULL); assertEqualIntA(a, ARCHIVE_OK, archive_write_set_format_ustar(a)); r = archive_write_add_filter_by_name(a, filter_name); if (r == ARCHIVE_WARN) { if (!can_filter_prog()) { skipping("%s filter not suported on this platform", filter_name); assertEqualInt(ARCHIVE_OK, archive_write_free(a)); free(buff); return; } } else if (r == ARCHIVE_FATAL && (strcmp(archive_error_string(a), "lzma compression not supported on this platform") == 0 || strcmp(archive_error_string(a), "xz compression not supported on this platform") == 0)) { skipping("%s filter not suported on this platform", filter_name); assertEqualInt(ARCHIVE_OK, archive_write_free(a)); free(buff); return; } else { if (!assertEqualIntA(a, ARCHIVE_OK, r)) { assertEqualInt(ARCHIVE_OK, archive_write_free(a)); free(buff); return; } } if (filter_code == ARCHIVE_FILTER_LRZIP) { /* * There's a bug in lrzip (as of release 0.612) where 2nd stage * compression can't be performed on smaller files. Set lrzip to * use no 2nd stage compression. */ assertEqualIntA(a, ARCHIVE_OK, archive_write_set_options(a, "lrzip:compression=none")); } assertEqualIntA(a, ARCHIVE_OK, archive_write_set_bytes_per_block(a, 10)); assertEqualIntA(a, ARCHIVE_OK, archive_write_open_memory(a, buff, buffsize, &used)); /* * Write a file to it. */ assert((ae = archive_entry_new()) != NULL); archive_entry_set_mtime(ae, 1, 0); assertEqualInt(1, archive_entry_mtime(ae)); archive_entry_set_ctime(ae, 1, 0); assertEqualInt(1, archive_entry_ctime(ae)); archive_entry_set_atime(ae, 1, 0); assertEqualInt(1, archive_entry_atime(ae)); archive_entry_copy_pathname(ae, "file"); assertEqualString("file", archive_entry_pathname(ae)); archive_entry_set_mode(ae, AE_IFREG | 0755); assertEqualInt((AE_IFREG | 0755), archive_entry_mode(ae)); archive_entry_set_size(ae, 8); assertEqualInt(0, archive_write_header(a, ae)); archive_entry_free(ae); assertEqualInt(8, archive_write_data(a, "12345678", 8)); /* Close out the archive. */ assertEqualIntA(a, ARCHIVE_OK, archive_write_close(a)); assertEqualInt(ARCHIVE_OK, archive_write_free(a)); /* * Now, read the data back. */ assert((a = archive_read_new()) != NULL); assertEqualIntA(a, ARCHIVE_OK, archive_read_support_format_all(a)); assertEqualIntA(a, ARCHIVE_OK, archive_read_support_filter_all(a)); assertEqualIntA(a, ARCHIVE_OK, archive_read_open_memory(a, buff, used)); /* * Read and verify the file. */ assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); assertEqualInt(1, archive_entry_mtime(ae)); assertEqualString("file", archive_entry_pathname(ae)); assertEqualInt(AE_IFREG, archive_entry_filetype(ae)); assertEqualInt(8, archive_entry_size(ae)); /* Verify the end of the archive. */ assertEqualIntA(a, ARCHIVE_EOF, archive_read_next_header(a, &ae)); /* Verify archive format. */ assertEqualIntA(a, filter_code, archive_filter_code(a, 0)); assertEqualIntA(a, ARCHIVE_FORMAT_TAR_USTAR, archive_format(a)); assertEqualInt(ARCHIVE_OK, archive_read_close(a)); assertEqualInt(ARCHIVE_OK, archive_read_free(a)); free(buff); } static int canAlways(void) { return 1; } static int cannot(void) { return 0; } DEFINE_TEST(test_archive_write_add_filter_by_name_b64encode) { test_filter_by_name("b64encode", ARCHIVE_FILTER_UU, canAlways); } DEFINE_TEST(test_archive_write_add_filter_by_name_bzip2) { test_filter_by_name("bzip2", ARCHIVE_FILTER_BZIP2, canBzip2); } DEFINE_TEST(test_archive_write_add_filter_by_name_compress) { test_filter_by_name("compress", ARCHIVE_FILTER_COMPRESS, canAlways); } DEFINE_TEST(test_archive_write_add_filter_by_name_grzip) { test_filter_by_name("grzip", ARCHIVE_FILTER_GRZIP, canGrzip); } DEFINE_TEST(test_archive_write_add_filter_by_name_gzip) { test_filter_by_name("gzip", ARCHIVE_FILTER_GZIP, canGzip); } DEFINE_TEST(test_archive_write_add_filter_by_name_lrzip) { test_filter_by_name("lrzip", ARCHIVE_FILTER_LRZIP, canLrzip); } DEFINE_TEST(test_archive_write_add_filter_by_name_lz4) { test_filter_by_name("lz4", ARCHIVE_FILTER_LZ4, canLz4); } DEFINE_TEST(test_archive_write_add_filter_by_name_lzip) { test_filter_by_name("lzip", ARCHIVE_FILTER_LZIP, cannot); } DEFINE_TEST(test_archive_write_add_filter_by_name_lzma) { test_filter_by_name("lzma", ARCHIVE_FILTER_LZMA, cannot); } DEFINE_TEST(test_archive_write_add_filter_by_name_lzop) { test_filter_by_name("lzop", ARCHIVE_FILTER_LZOP, canLzop); } DEFINE_TEST(test_archive_write_add_filter_by_name_uuencode) { test_filter_by_name("uuencode", ARCHIVE_FILTER_UU, canAlways); } DEFINE_TEST(test_archive_write_add_filter_by_name_xz) { test_filter_by_name("xz", ARCHIVE_FILTER_XZ, cannot); } + +DEFINE_TEST(test_archive_write_add_filter_by_name_zstd) +{ + test_filter_by_name("zstd", ARCHIVE_FILTER_ZSTD, canZstd); +} Index: stable/10/contrib/libarchive/libarchive/test/test_compat_zstd.c =================================================================== --- stable/10/contrib/libarchive/libarchive/test/test_compat_zstd.c (nonexistent) +++ stable/10/contrib/libarchive/libarchive/test/test_compat_zstd.c (revision 324418) @@ -0,0 +1,82 @@ +/*- + * Copyright (c) 2017 Sean Purcell + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES + * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. + * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, + * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT + * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF + * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ +#include "test.h" +__FBSDID("$FreeBSD$"); + +/* + * Verify our ability to read sample files compatibly with 'zstd -d'. + * + * In particular: + * * zstd -d will read multiple zstd streams, concatenating the output + * * zstd -d will skip over zstd skippable frames + */ + +static void +compat_zstd(const char *name) +{ + const char *n[7] = { "f1", "f2", "f3", "d1/f1", "d1/f2", "d1/f3", NULL }; + struct archive_entry *ae; + struct archive *a; + int i, r; + + assert((a = archive_read_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_read_support_filter_all(a)); + r = archive_read_support_filter_zstd(a); + if (r == ARCHIVE_WARN) { + skipping("zstd reading not fully supported on this platform"); + assertEqualInt(ARCHIVE_OK, archive_read_free(a)); + return; + } + assertEqualIntA(a, ARCHIVE_OK, archive_read_support_format_all(a)); + extract_reference_file(name); + assertEqualIntA(a, ARCHIVE_OK, archive_read_open_filename(a, name, 2)); + + /* Read entries, match up names with list above. */ + for (i = 0; i < 6; ++i) { + failure("Could not read file %d (%s) from %s", i, n[i], name); + assertEqualIntA(a, ARCHIVE_OK, + archive_read_next_header(a, &ae)); + assertEqualString(n[i], archive_entry_pathname(ae)); + } + + /* Verify the end-of-archive. */ + assertEqualIntA(a, ARCHIVE_EOF, archive_read_next_header(a, &ae)); + + /* Verify that the format detection worked. */ + assertEqualInt(archive_filter_code(a, 0), ARCHIVE_FILTER_ZSTD); + assertEqualString(archive_filter_name(a, 0), "zstd"); + assertEqualInt(archive_format(a), ARCHIVE_FORMAT_TAR_USTAR); + + assertEqualInt(ARCHIVE_OK, archive_read_close(a)); + assertEqualInt(ARCHIVE_OK, archive_read_free(a)); +} + + +DEFINE_TEST(test_compat_zstd) +{ + /* This sample was compressed as 3 separate streams with a zstd skippable + * frame placed in the middle */ + compat_zstd("test_compat_zstd_1.tar.zst"); +} Property changes on: stable/10/contrib/libarchive/libarchive/test/test_compat_zstd.c ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: stable/10/contrib/libarchive/libarchive/test/test_compat_zstd_1.tar.zst.uu =================================================================== --- stable/10/contrib/libarchive/libarchive/test/test_compat_zstd_1.tar.zst.uu (nonexistent) +++ stable/10/contrib/libarchive/libarchive/test/test_compat_zstd_1.tar.zst.uu (revision 324418) @@ -0,0 +1,12 @@ +begin 644 test_compat_zstd_1.tar.zst +M*+4O_010)0,`HL0.%;`Q&>>\/$2[#IQF[<1+Z3T<0CX]!77&0@R.6+/F,0+I +M.$1A$QE2`J!+*_6[_YT9_W_M1KC-EG*V>10.`,M`%3*@#F#\`-FT#J:1#U1" +M`H1!&R#A5\*'"JYIJ;C&4=B2CL(L)*E-IJT/RV?.:A_]_N +MB&[7SDG;/=4&#P";0!5D0`=8T0&R&19,)1^HA`0(@S9`N`G.!0!IP&NM<,K! +M-#8!%A]U]K10*DT8!`````$"`P0HM2_]!%!]`P`B11`6H+$)"%]@,Z6OH`"L +MM$R2MAN&*MSG`W?OJ7+4P*B::VXR`NM(1&$7&58"J*U'_&V^S$/_O]U1N%T[ +M)VW7J'+4!A``_4$%^T`],J`8P.0!L@D63"4?J(0$"(,V0+@)S@4`: 0); if (filesets[n].names[0] == NULL || filesets[n].names[1] == NULL) { failure("Internal buffer is not big enough for " "uncompressed test files"); } else { failure("Internal buffer is not big enough for " "uncompressed test file: %s", filesets[n].names[0]); } if (!assert(size < buffsize)) { free(rawimage); rawimage = NULL; continue; } } else { for (i = 0; filesets[n].names[i] != NULL; ++i) { char *newraw; tmp = slurpfile(&size, filesets[n].names[i]); newraw = realloc(rawimage, oldsize + size); if (!assert(newraw != NULL)) { free(rawimage); rawimage = NULL; free(tmp); continue; } rawimage = newraw; memcpy(rawimage + oldsize, tmp, size); oldsize += size; size = oldsize; free(tmp); } } if (size == 0) { free(rawimage); rawimage = NULL; continue; } image = malloc(size); assert(image != NULL); if (image == NULL) { free(rawimage); rawimage = NULL; return; } assert(rawimage != NULL); srand((unsigned)time(NULL)); for (i = 0; i < 1000; ++i) { FILE *f; int j, numbytes, trycnt; /* Fuzz < 1% of the bytes in the archive. */ memcpy(image, rawimage, size); q = (int)size / 100; if (q < 4) q = 4; numbytes = (int)(rand() % q); for (j = 0; j < numbytes; ++j) image[rand() % size] = (char)rand(); /* Save the messed-up image to a file. * If we crash, that file will be useful. */ for (trycnt = 0; trycnt < 3; trycnt++) { f = fopen("after.test.failure.send.this.file." "to.libarchive.maintainers.with.system.details", "wb"); if (f != NULL) break; #if defined(_WIN32) && !defined(__CYGWIN__) /* * Sometimes previous close operation does not completely * end at this time. So we should take a wait while * the operation running. */ Sleep(100); #endif } assert(f != NULL); assertEqualInt((size_t)size, fwrite(image, 1, (size_t)size, f)); fclose(f); // Try to read all headers and bodies. assert((a = archive_read_new()) != NULL); assertEqualIntA(a, ARCHIVE_OK, archive_read_support_filter_all(a)); assertEqualIntA(a, ARCHIVE_OK, archive_read_support_format_all(a)); if (0 == archive_read_open_memory(a, image, size)) { while(0 == archive_read_next_header(a, &ae)) { while (0 == archive_read_data_block(a, &blk, &blk_size, &blk_offset)) continue; } archive_read_close(a); } archive_read_free(a); // Just list headers, skip bodies. assert((a = archive_read_new()) != NULL); assertEqualIntA(a, ARCHIVE_OK, archive_read_support_filter_all(a)); assertEqualIntA(a, ARCHIVE_OK, archive_read_support_format_all(a)); if (0 == archive_read_open_memory(a, image, size)) { while(0 == archive_read_next_header(a, &ae)) { } archive_read_close(a); } archive_read_free(a); } free(image); free(rawimage); } } DEFINE_TEST(test_fuzz_ar) { static const char *fileset1[] = { "test_read_format_ar.ar", NULL }; static const struct files filesets[] = { {0, fileset1}, {1, NULL} }; test_fuzz(filesets); } DEFINE_TEST(test_fuzz_cab) { static const char *fileset1[] = { "test_fuzz.cab", NULL }; static const struct files filesets[] = { {0, fileset1}, {1, NULL} }; test_fuzz(filesets); } DEFINE_TEST(test_fuzz_cpio) { static const char *fileset1[] = { "test_read_format_cpio_bin_be.cpio", NULL }; static const char *fileset2[] = { "test_read_format_cpio_bin_le.cpio", NULL }; static const char *fileset3[] = { /* Test RPM unwrapper */ "test_read_format_cpio_svr4_gzip_rpm.rpm", NULL }; static const struct files filesets[] = { {0, fileset1}, {0, fileset2}, {0, fileset3}, {1, NULL} }; test_fuzz(filesets); } DEFINE_TEST(test_fuzz_iso9660) { static const char *fileset1[] = { "test_fuzz_1.iso.Z", NULL }; static const struct files filesets[] = { {0, fileset1}, /* Exercise compress decompressor. */ {1, fileset1}, {1, NULL} }; test_fuzz(filesets); } DEFINE_TEST(test_fuzz_lzh) { static const char *fileset1[] = { "test_fuzz.lzh", NULL }; static const struct files filesets[] = { {0, fileset1}, {1, NULL} }; test_fuzz(filesets); } DEFINE_TEST(test_fuzz_mtree) { static const char *fileset1[] = { "test_read_format_mtree.mtree", NULL }; static const struct files filesets[] = { {0, fileset1}, {1, NULL} }; test_fuzz(filesets); } DEFINE_TEST(test_fuzz_rar) { static const char *fileset1[] = { /* Uncompressed RAR test */ "test_read_format_rar.rar", NULL }; static const char *fileset2[] = { /* RAR file with binary data */ "test_read_format_rar_binary_data.rar", NULL }; static const char *fileset3[] = { /* Best Compressed RAR test */ "test_read_format_rar_compress_best.rar", NULL }; static const char *fileset4[] = { /* Normal Compressed RAR test */ "test_read_format_rar_compress_normal.rar", NULL }; static const char *fileset5[] = { /* Normal Compressed Multi LZSS blocks RAR test */ "test_read_format_rar_multi_lzss_blocks.rar", NULL }; static const char *fileset6[] = { /* RAR with no EOF header */ "test_read_format_rar_noeof.rar", NULL }; static const char *fileset7[] = { /* Best Compressed RAR file with both PPMd and LZSS blocks */ "test_read_format_rar_ppmd_lzss_conversion.rar", NULL }; static const char *fileset8[] = { /* RAR with subblocks */ "test_read_format_rar_subblock.rar", NULL }; static const char *fileset9[] = { /* RAR with Unicode filenames */ "test_read_format_rar_unicode.rar", NULL }; static const char *fileset10[] = { "test_read_format_rar_multivolume.part0001.rar", "test_read_format_rar_multivolume.part0002.rar", "test_read_format_rar_multivolume.part0003.rar", "test_read_format_rar_multivolume.part0004.rar", NULL }; static const struct files filesets[] = { {0, fileset1}, {0, fileset2}, {0, fileset3}, {0, fileset4}, {0, fileset5}, {0, fileset6}, {0, fileset7}, {0, fileset8}, {0, fileset9}, {0, fileset10}, {1, NULL} }; test_fuzz(filesets); } DEFINE_TEST(test_fuzz_tar) { static const char *fileset1[] = { "test_compat_bzip2_1.tbz", NULL }; static const char *fileset2[] = { "test_compat_gtar_1.tar", NULL }; static const char *fileset3[] = { "test_compat_gzip_1.tgz", NULL }; static const char *fileset4[] = { "test_compat_gzip_2.tgz", NULL }; static const char *fileset5[] = { "test_compat_tar_hardlink_1.tar", NULL }; static const char *fileset6[] = { "test_compat_xz_1.txz", NULL }; static const char *fileset7[] = { "test_read_format_gtar_sparse_1_17_posix10_modified.tar", NULL }; static const char *fileset8[] = { "test_read_format_tar_empty_filename.tar", NULL }; #if HAVE_LIBLZO2 && HAVE_LZO_LZO1X_H && HAVE_LZO_LZOCONF_H static const char *fileset9[] = { "test_compat_lzop_1.tar.lzo", NULL }; #endif +#if HAVE_ZSTD_H && HAVE_LIBZSTD + static const char *fileset10[] = { + "test_compat_zstd_1.tar.zst", + NULL + }; +#endif static const struct files filesets[] = { {0, fileset1}, /* Exercise bzip2 decompressor. */ {1, fileset1}, {0, fileset2}, {0, fileset3}, /* Exercise gzip decompressor. */ {0, fileset4}, /* Exercise gzip decompressor. */ {0, fileset5}, {0, fileset6}, /* Exercise xz decompressor. */ {0, fileset7}, {0, fileset8}, #if HAVE_LIBLZO2 && HAVE_LZO_LZO1X_H && HAVE_LZO_LZOCONF_H {0, fileset9}, /* Exercise lzo decompressor. */ +#endif +#if HAVE_ZSTD_H && HAVE_LIBZSTD + {0, fileset10}, /* Excercise zstd decompressor. */ #endif {1, NULL} }; test_fuzz(filesets); } DEFINE_TEST(test_fuzz_zip) { static const char *fileset1[] = { "test_compat_zip_1.zip", NULL }; static const char *fileset2[] = { "test_compat_zip_2.zip", NULL }; static const char *fileset3[] = { "test_compat_zip_3.zip", NULL }; static const char *fileset4[] = { "test_compat_zip_4.zip", NULL }; static const char *fileset5[] = { "test_compat_zip_5.zip", NULL }; static const char *fileset6[] = { "test_compat_zip_6.zip", NULL }; static const char *fileset7[] = { "test_read_format_zip.zip", NULL }; static const char *fileset8[] = { "test_read_format_zip_comment_stored_1.zip", NULL }; static const char *fileset9[] = { "test_read_format_zip_comment_stored_2.zip", NULL }; static const char *fileset10[] = { "test_read_format_zip_encryption_data.zip", NULL }; static const char *fileset11[] = { "test_read_format_zip_encryption_header.zip", NULL }; static const char *fileset12[] = { "test_read_format_zip_encryption_partially.zip", NULL }; static const char *fileset13[] = { "test_read_format_zip_filename_cp866.zip", NULL }; static const char *fileset14[] = { "test_read_format_zip_filename_cp932.zip", NULL }; static const char *fileset15[] = { "test_read_format_zip_filename_koi8r.zip", NULL }; static const char *fileset16[] = { "test_read_format_zip_filename_utf8_jp.zip", NULL }; static const char *fileset17[] = { "test_read_format_zip_filename_utf8_ru.zip", NULL }; static const char *fileset18[] = { "test_read_format_zip_filename_utf8_ru2.zip", NULL }; static const char *fileset19[] = { "test_read_format_zip_length_at_end.zip", NULL }; static const char *fileset20[] = { "test_read_format_zip_mac_metadata.zip", NULL }; static const char *fileset21[] = { "test_read_format_zip_malformed1.zip", NULL }; static const char *fileset22[] = { "test_read_format_zip_msdos.zip", NULL }; static const char *fileset23[] = { "test_read_format_zip_nested.zip", NULL }; static const char *fileset24[] = { "test_read_format_zip_nofiletype.zip", NULL }; static const char *fileset25[] = { "test_read_format_zip_padded1.zip", NULL }; static const char *fileset26[] = { "test_read_format_zip_padded2.zip", NULL }; static const char *fileset27[] = { "test_read_format_zip_padded3.zip", NULL }; static const char *fileset28[] = { "test_read_format_zip_symlink.zip", NULL }; static const char *fileset29[] = { "test_read_format_zip_traditional_encryption_data.zip", NULL }; static const char *fileset30[] = { "test_read_format_zip_ux.zip", NULL }; static const char *fileset31[] = { "test_read_format_zip_winzip_aes128.zip", NULL }; static const char *fileset32[] = { "test_read_format_zip_winzip_aes256.zip", NULL }; static const char *fileset33[] = { "test_read_format_zip_winzip_aes256_large.zip", NULL }; static const char *fileset34[] = { "test_read_format_zip_winzip_aes256_stored.zip", NULL }; static const char *fileset35[] = { "test_read_format_zip_zip64a.zip", NULL }; static const char *fileset36[] = { "test_read_format_zip_zip64b.zip", NULL }; static const struct files filesets[] = { {0, fileset1}, {0, fileset2}, {0, fileset3}, {0, fileset4}, {0, fileset5}, {0, fileset6}, {0, fileset7}, {0, fileset8}, {0, fileset9}, {0, fileset10}, {0, fileset11}, {0, fileset12}, {0, fileset13}, {0, fileset14}, {0, fileset15}, {0, fileset16}, {0, fileset17}, {0, fileset18}, {0, fileset19}, {0, fileset20}, {0, fileset21}, {0, fileset22}, {0, fileset23}, {0, fileset24}, {0, fileset25}, {0, fileset26}, {0, fileset27}, {0, fileset28}, {0, fileset29}, {0, fileset30}, {0, fileset31}, {0, fileset32}, {0, fileset33}, {0, fileset34}, {0, fileset35}, {0, fileset36}, {1, NULL} }; test_fuzz(filesets); } Index: stable/10/contrib/libarchive/libarchive/test/test_write_filter_zstd.c =================================================================== --- stable/10/contrib/libarchive/libarchive/test/test_write_filter_zstd.c (nonexistent) +++ stable/10/contrib/libarchive/libarchive/test/test_write_filter_zstd.c (revision 324418) @@ -0,0 +1,201 @@ +/*- + * Copyright (c) 2017 Sean Purcell + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer + * in this position and unchanged. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES + * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. + * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, + * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT + * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF + * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ + +#include "test.h" +__FBSDID("$FreeBSD$"); + +DEFINE_TEST(test_write_filter_zstd) +{ + struct archive_entry *ae; + struct archive *a; + char *buff, *data; + size_t buffsize, datasize; + char path[16]; + size_t used1, used2; + int i, r; + + buffsize = 2000000; + assert(NULL != (buff = (char *)malloc(buffsize))); + if (buff == NULL) + return; + + datasize = 10000; + assert(NULL != (data = (char *)malloc(datasize))); + if (data == NULL) { + free(buff); + return; + } + memset(data, 0, datasize); + + /* + * Write a 100 files and read them all back. + */ + assert((a = archive_write_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_write_set_format_ustar(a)); + r = archive_write_add_filter_zstd(a); + if (r != ARCHIVE_OK) { + skipping("zstd writing not supported on this platform"); + assertEqualInt(ARCHIVE_OK, archive_write_free(a)); + free(buff); + free(data); + return; + } + assertEqualIntA(a, ARCHIVE_OK, + archive_write_set_bytes_per_block(a, 10)); + assertEqualInt(ARCHIVE_FILTER_ZSTD, archive_filter_code(a, 0)); + assertEqualString("zstd", archive_filter_name(a, 0)); + assertEqualIntA(a, ARCHIVE_OK, archive_write_open_memory(a, buff, buffsize, &used1)); + assertEqualInt(ARCHIVE_FILTER_ZSTD, archive_filter_code(a, 0)); + assertEqualString("zstd", archive_filter_name(a, 0)); + assert((ae = archive_entry_new()) != NULL); + archive_entry_set_filetype(ae, AE_IFREG); + archive_entry_set_size(ae, datasize); + for (i = 0; i < 100; i++) { + sprintf(path, "file%03d", i); + archive_entry_copy_pathname(ae, path); + assertEqualIntA(a, ARCHIVE_OK, archive_write_header(a, ae)); + assertA(datasize + == (size_t)archive_write_data(a, data, datasize)); + } + archive_entry_free(ae); + assertEqualIntA(a, ARCHIVE_OK, archive_write_close(a)); + assertEqualInt(ARCHIVE_OK, archive_write_free(a)); + + assert((a = archive_read_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_read_support_format_all(a)); + r = archive_read_support_filter_zstd(a); + if (r == ARCHIVE_WARN) { + skipping("Can't verify zstd writing by reading back;" + " zstd reading not fully supported on this platform"); + } else { + assertEqualIntA(a, ARCHIVE_OK, + archive_read_support_filter_all(a)); + assertEqualIntA(a, ARCHIVE_OK, + archive_read_open_memory(a, buff, used1)); + for (i = 0; i < 100; i++) { + sprintf(path, "file%03d", i); + if (!assertEqualInt(ARCHIVE_OK, + archive_read_next_header(a, &ae))) + break; + assertEqualString(path, archive_entry_pathname(ae)); + assertEqualInt((int)datasize, archive_entry_size(ae)); + } + assertEqualIntA(a, ARCHIVE_OK, archive_read_close(a)); + } + assertEqualInt(ARCHIVE_OK, archive_read_free(a)); + + /* + * Repeat the cycle again, this time setting some compression + * options. + */ + assert((a = archive_write_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_write_set_format_ustar(a)); + assertEqualIntA(a, ARCHIVE_OK, + archive_write_set_bytes_per_block(a, 10)); + assertEqualIntA(a, ARCHIVE_OK, archive_write_add_filter_zstd(a)); + assertEqualIntA(a, ARCHIVE_FAILED, + archive_write_set_filter_option(a, NULL, "nonexistent-option", "0")); + assertEqualIntA(a, ARCHIVE_FAILED, + archive_write_set_filter_option(a, NULL, "compression-level", "abc")); + assertEqualIntA(a, ARCHIVE_FAILED, + archive_write_set_filter_option(a, NULL, "compression-level", "25")); /* too big */ + assertEqualIntA(a, ARCHIVE_OK, + archive_write_set_filter_option(a, NULL, "compression-level", "9")); + assertEqualIntA(a, ARCHIVE_OK, + archive_write_set_filter_option(a, NULL, "compression-level", "15")); + assertEqualIntA(a, ARCHIVE_OK, archive_write_open_memory(a, buff, buffsize, &used2)); + for (i = 0; i < 100; i++) { + sprintf(path, "file%03d", i); + assert((ae = archive_entry_new()) != NULL); + archive_entry_copy_pathname(ae, path); + archive_entry_set_size(ae, datasize); + archive_entry_set_filetype(ae, AE_IFREG); + assertEqualIntA(a, ARCHIVE_OK, archive_write_header(a, ae)); + assertA(datasize == (size_t)archive_write_data(a, data, datasize)); + archive_entry_free(ae); + } + assertEqualIntA(a, ARCHIVE_OK, archive_write_close(a)); + assertEqualInt(ARCHIVE_OK, archive_write_free(a)); + + failure("compression-level=15 wrote %d bytes, default wrote %d bytes", + (int)used2, (int)used1); + assert(used2 < used1); + + assert((a = archive_read_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_read_support_format_all(a)); + r = archive_read_support_filter_zstd(a); + if (r == ARCHIVE_WARN) { + skipping("zstd reading not fully supported on this platform"); + } else { + assertEqualIntA(a, ARCHIVE_OK, + archive_read_support_filter_all(a)); + assertEqualIntA(a, ARCHIVE_OK, + archive_read_open_memory(a, buff, used2)); + for (i = 0; i < 100; i++) { + sprintf(path, "file%03d", i); + failure("Trying to read %s", path); + if (!assertEqualIntA(a, ARCHIVE_OK, + archive_read_next_header(a, &ae))) + break; + assertEqualString(path, archive_entry_pathname(ae)); + assertEqualInt((int)datasize, archive_entry_size(ae)); + } + assertEqualIntA(a, ARCHIVE_OK, archive_read_close(a)); + } + assertEqualInt(ARCHIVE_OK, archive_read_free(a)); + + /* + * Test various premature shutdown scenarios to make sure we + * don't crash or leak memory. + */ + assert((a = archive_write_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_write_add_filter_zstd(a)); + assertEqualInt(ARCHIVE_OK, archive_write_free(a)); + + assert((a = archive_write_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_write_add_filter_zstd(a)); + assertEqualInt(ARCHIVE_OK, archive_write_close(a)); + assertEqualInt(ARCHIVE_OK, archive_write_free(a)); + + assert((a = archive_write_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_write_set_format_ustar(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_write_add_filter_zstd(a)); + assertEqualInt(ARCHIVE_OK, archive_write_close(a)); + assertEqualInt(ARCHIVE_OK, archive_write_free(a)); + + assert((a = archive_write_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_write_set_format_ustar(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_write_add_filter_zstd(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_write_open_memory(a, buff, buffsize, &used2)); + assertEqualInt(ARCHIVE_OK, archive_write_close(a)); + assertEqualInt(ARCHIVE_OK, archive_write_free(a)); + + /* + * Clean up. + */ + free(data); + free(buff); +} Property changes on: stable/10/contrib/libarchive/libarchive/test/test_write_filter_zstd.c ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: stable/10/contrib/libarchive/tar/bsdtar.1 =================================================================== --- stable/10/contrib/libarchive/tar/bsdtar.1 (revision 324417) +++ stable/10/contrib/libarchive/tar/bsdtar.1 (revision 324418) @@ -1,1242 +1,1259 @@ .\" Copyright (c) 2003-2007 Tim Kientzle .\" Copyright (c) 2017 Martin Matuska .\" All rights reserved. .\" .\" Redistribution and use in source and binary forms, with or without .\" modification, are permitted provided that the following conditions .\" are met: .\" 1. Redistributions of source code must retain the above copyright .\" notice, this list of conditions and the following disclaimer. .\" 2. Redistributions in binary form must reproduce the above copyright .\" notice, this list of conditions and the following disclaimer in the .\" documentation and/or other materials provided with the distribution. .\" .\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE .\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF .\" SUCH DAMAGE. .\" .\" $FreeBSD$ .\" -.Dd February 25, 2017 +.Dd October 1, 2017 .Dt TAR 1 .Os .Sh NAME .Nm tar .Nd manipulate tape archives .Sh SYNOPSIS .Nm .Op Ar bundled-flags Ao args Ac .Op Ao Ar file Ac | Ao Ar pattern Ac ... .Nm .Brq Fl c .Op Ar options .Op Ar files | Ar directories .Nm .Brq Fl r | Fl u .Fl f Ar archive-file .Op Ar options .Op Ar files | Ar directories .Nm .Brq Fl t | Fl x .Op Ar options .Op Ar patterns .Sh DESCRIPTION .Nm creates and manipulates streaming archive files. This implementation can extract from tar, pax, cpio, zip, jar, ar, xar, rpm, 7-zip, and ISO 9660 cdrom images and can create tar, pax, cpio, ar, zip, 7-zip, and shar archives. .Pp The first synopsis form shows a .Dq bundled option word. This usage is provided for compatibility with historical implementations. See COMPATIBILITY below for details. .Pp The other synopsis forms show the preferred usage. The first option to .Nm is a mode indicator from the following list: .Bl -tag -compact -width indent .It Fl c Create a new archive containing the specified items. The long option form is .Fl Fl create . .It Fl r Like .Fl c , but new entries are appended to the archive. Note that this only works on uncompressed archives stored in regular files. The .Fl f option is required. The long option form is .Fl Fl append . .It Fl t List archive contents to stdout. The long option form is .Fl Fl list . .It Fl u Like .Fl r , but new entries are added only if they have a modification date newer than the corresponding entry in the archive. Note that this only works on uncompressed archives stored in regular files. The .Fl f option is required. The long form is .Fl Fl update . .It Fl x Extract to disk from the archive. If a file with the same name appears more than once in the archive, each copy will be extracted, with later copies overwriting (replacing) earlier copies. The long option form is .Fl Fl extract . .El .Pp In .Fl c , .Fl r , or .Fl u mode, each specified file or directory is added to the archive in the order specified on the command line. By default, the contents of each directory are also archived. .Pp In extract or list mode, the entire command line is read and parsed before the archive is opened. The pathnames or patterns on the command line indicate which items in the archive should be processed. Patterns are shell-style globbing patterns as documented in .Xr tcsh 1 . .Sh OPTIONS Unless specifically stated otherwise, options are applicable in all operating modes. .Bl -tag -width indent .It Cm @ Ns Pa archive (c and r modes only) The specified archive is opened and the entries in it will be appended to the current archive. As a simple example, .Dl Nm Fl c Fl f Pa - Pa newfile Cm @ Ns Pa original.tar writes a new archive to standard output containing a file .Pa newfile and all of the entries from .Pa original.tar . In contrast, .Dl Nm Fl c Fl f Pa - Pa newfile Pa original.tar creates a new archive with only two entries. Similarly, .Dl Nm Fl czf Pa - Fl Fl format Cm pax Cm @ Ns Pa - reads an archive from standard input (whose format will be determined automatically) and converts it into a gzip-compressed pax-format archive on stdout. In this way, .Nm can be used to convert archives from one format to another. .It Fl a , Fl Fl auto-compress (c mode only) Use the archive suffix to decide a set of the format and the compressions. As a simple example, .Dl Nm Fl a Fl cf Pa archive.tgz source.c source.h creates a new archive with restricted pax format and gzip compression, .Dl Nm Fl a Fl cf Pa archive.tar.bz2.uu source.c source.h creates a new archive with restricted pax format and bzip2 compression and uuencode compression, .Dl Nm Fl a Fl cf Pa archive.zip source.c source.h creates a new archive with zip format, .Dl Nm Fl a Fl jcf Pa archive.tgz source.c source.h ignores the .Dq -j option, and creates a new archive with restricted pax format and gzip compression, .Dl Nm Fl a Fl jcf Pa archive.xxx source.c source.h if it is unknown suffix or no suffix, creates a new archive with restricted pax format and bzip2 compression. .It Fl Fl acls (c, r, u, x modes only) Archive or extract POSIX.1e or NFSv4 ACLs. This is the reverse of .Fl Fl no-acls and the default behavior in c, r, and u modes (except on Mac OS X) or if .Nm is run in x mode as root. On Mac OS X this option translates extended ACLs to NFSv4 ACLs. To store extended ACLs the .Fl Fl mac-metadata option is preferred. .It Fl B , Fl Fl read-full-blocks Ignored for compatibility with other .Xr tar 1 implementations. .It Fl b Ar blocksize , Fl Fl block-size Ar blocksize Specify the block size, in 512-byte records, for tape drive I/O. As a rule, this argument is only needed when reading from or writing to tape drives, and usually not even then as the default block size of 20 records (10240 bytes) is very common. .It Fl C Ar directory , Fl Fl cd Ar directory , Fl Fl directory Ar directory In c and r mode, this changes the directory before adding the following files. In x mode, change directories after opening the archive but before extracting entries from the archive. .It Fl Fl chroot (x mode only) .Fn chroot to the current directory after processing any .Fl C options and before extracting any files. .It Fl Fl clear-nochange-fflags (x mode only) Before removing file system objects to replace them, clear platform-specific file flags that might prevent removal. .It Fl Fl exclude Ar pattern Do not process files or directories that match the specified pattern. Note that exclusions take precedence over patterns or filenames specified on the command line. .It Fl Fl fflags (c, r, u, x modes only) Archive or extract file flags. This is the reverse of .Fl Fl no-fflags and the default behavior in c, r, and u modes or if .Nm is run in x mode as root. .It Fl Fl format Ar format (c, r, u mode only) Use the specified format for the created archive. Supported formats include .Dq cpio , .Dq pax , .Dq shar , and .Dq ustar . Other formats may also be supported; see .Xr libarchive-formats 5 for more information about currently-supported formats. In r and u modes, when extending an existing archive, the format specified here must be compatible with the format of the existing archive on disk. .It Fl f Ar file , Fl Fl file Ar file Read the archive from or write the archive to the specified file. The filename can be .Pa - for standard input or standard output. The default varies by system; on .Fx , the default is .Pa /dev/sa0 ; on Linux, the default is .Pa /dev/st0 . .It Fl Fl gid Ar id Use the provided group id number. On extract, this overrides the group id in the archive; the group name in the archive will be ignored. On create, this overrides the group id read from disk; if .Fl Fl gname is not also specified, the group name will be set to match the group id. .It Fl Fl gname Ar name Use the provided group name. On extract, this overrides the group name in the archive; if the provided group name does not exist on the system, the group id (from the archive or from the .Fl Fl gid option) will be used instead. On create, this sets the group name that will be stored in the archive; the name will not be verified against the system group database. .It Fl H (c and r modes only) Symbolic links named on the command line will be followed; the target of the link will be archived, not the link itself. .It Fl h (c and r modes only) Synonym for .Fl L . .It Fl I Synonym for .Fl T . .It Fl Fl help Show usage. .It Fl Fl hfsCompression (x mode only) Mac OS X specific (v10.6 or later). Compress extracted regular files with HFS+ compression. .It Fl Fl ignore-zeros An alias of .Fl Fl options Cm read_concatenated_archives for compatibility with GNU tar. .It Fl Fl include Ar pattern Process only files or directories that match the specified pattern. Note that exclusions specified with .Fl Fl exclude take precedence over inclusions. If no inclusions are explicitly specified, all entries are processed by default. The .Fl Fl include option is especially useful when filtering archives. For example, the command .Dl Nm Fl c Fl f Pa new.tar Fl Fl include='*foo*' Cm @ Ns Pa old.tgz creates a new archive .Pa new.tar containing only the entries from .Pa old.tgz containing the string .Sq foo . .It Fl J , Fl Fl xz (c mode only) Compress the resulting archive with .Xr xz 1 . In extract or list modes, this option is ignored. -Note that, unlike other +Note that this .Nm tar -implementations, this implementation recognizes XZ compression -automatically when reading archives. +implementation recognizes XZ compression automatically when reading archives. .It Fl j , Fl Fl bzip , Fl Fl bzip2 , Fl Fl bunzip2 (c mode only) Compress the resulting archive with .Xr bzip2 1 . In extract or list modes, this option is ignored. -Note that, unlike other +Note that this .Nm tar -implementations, this implementation recognizes bzip2 compression -automatically when reading archives. +implementation recognizes bzip2 compression automatically when reading +archives. .It Fl k , Fl Fl keep-old-files (x mode only) Do not overwrite existing files. In particular, if a file appears more than once in an archive, later copies will not overwrite earlier copies. .It Fl Fl keep-newer-files (x mode only) Do not overwrite existing files that are newer than the versions appearing in the archive being extracted. .It Fl L , Fl Fl dereference (c and r modes only) All symbolic links will be followed. Normally, symbolic links are archived as such. With this option, the target of the link will be archived instead. .It Fl l , Fl Fl check-links (c and r modes only) Issue a warning message unless all links to each file are archived. .It Fl Fl lrzip (c mode only) Compress the resulting archive with .Xr lrzip 1 . In extract or list modes, this option is ignored. +Note that this +.Nm tar +implementation recognizes lrzip compression automatically when reading +archives. .It Fl Fl lz4 (c mode only) Compress the archive with lz4-compatible compression before writing it. -In input mode, this option is ignored; lz4 compression is recognized -automatically on input. +In extract or list modes, this option is ignored. +Note that this +.Nm tar +implementation recognizes lz4 compression automatically when reading archives. +.It Fl Fl zstd +(c mode only) +Compress the archive with zstd-compatible compression before writing it. +In extract or list modes, this option is ignored. +Note that this +.Nm tar +implementation recognizes zstd compression automatically when reading archives. .It Fl Fl lzma (c mode only) Compress the resulting archive with the original LZMA algorithm. +In extract or list modes, this option is ignored. Use of this option is discouraged and new archives should be created with .Fl Fl xz instead. -Note that, unlike other +Note that this .Nm tar -implementations, this implementation recognizes LZMA compression -automatically when reading archives. +implementation recognizes LZMA compression automatically when reading archives. .It Fl Fl lzop (c mode only) Compress the resulting archive with .Xr lzop 1 . In extract or list modes, this option is ignored. +Note that this +.Nm tar +implementation recognizes LZO compression automatically when reading archives. .It Fl m , Fl Fl modification-time (x mode only) Do not extract modification time. By default, the modification time is set to the time stored in the archive. .It Fl Fl mac-metadata (c, r, u and x mode only) Mac OS X specific. Archive or extract extended ACLs and extended attributes using .Xr copyfile 3 in AppleDouble format. This is the reverse of .Fl Fl no-mac-metadata . and the default behavior in c, r, and u modes or if .Nm is run in x mode as root. .It Fl n , Fl Fl norecurse , Fl Fl no-recursion (c, r, u modes only) Do not recursively archive the contents of directories. .It Fl Fl newer Ar date (c, r, u modes only) Only include files and directories newer than the specified date. This compares ctime entries. .It Fl Fl newer-mtime Ar date (c, r, u modes only) Like .Fl Fl newer , except it compares mtime entries instead of ctime entries. .It Fl Fl newer-than Pa file (c, r, u modes only) Only include files and directories newer than the specified file. This compares ctime entries. .It Fl Fl newer-mtime-than Pa file (c, r, u modes only) Like .Fl Fl newer-than , except it compares mtime entries instead of ctime entries. .It Fl Fl nodump (c and r modes only) Honor the nodump file flag by skipping this file. .It Fl Fl nopreserveHFSCompression (x mode only) Mac OS X specific (v10.6 or later). Do not compress extracted regular files which were compressed with HFS+ compression before archived. By default, compress the regular files again with HFS+ compression. .It Fl Fl null (use with .Fl I or .Fl T ) Filenames or patterns are separated by null characters, not by newlines. This is often used to read filenames output by the .Fl print0 option to .Xr find 1 . .It Fl Fl no-acls (c, r, u, x modes only) Do not archive or extract POSIX.1e or NFSv4 ACLs. This is the reverse of .Fl Fl acls and the default behavior if .Nm is run as non-root in x mode (on Mac OS X as any user in c, r, u and x modes). .It Fl Fl no-fflags (c, r, u, x modes only) Do not archive or extract file flags. This is the reverse of .Fl Fl fflags and the default behavior if .Nm is run as non-root in x mode. .It Fl Fl no-mac-metadata (x mode only) Mac OS X specific. Do not archive or extract ACLs and extended attributes using .Xr copyfile 3 in AppleDouble format. This is the reverse of .Fl Fl mac-metadata . and the default behavior if .Nm is run as non-root in x mode. .It Fl n , Fl Fl norecurse , Fl Fl no-recursion .It Fl Fl no-same-owner (x mode only) Do not extract owner and group IDs. This is the reverse of .Fl Fl same-owner and the default behavior if .Nm is run as non-root. .It Fl Fl no-same-permissions (x mode only) Do not extract full permissions (SGID, SUID, sticky bit, ACLs, extended attributes or extended file flags). This is the reverse of .Fl p and the default behavior if .Nm is run as non-root. .It Fl Fl no-xattrs (c, r, u, x modes only) Do not archive or extract extended attributes. This is the reverse of .Fl Fl xattrs and the default behavior if .Nm is run as non-root in x mode. .It Fl Fl numeric-owner This is equivalent to .Fl Fl uname .Qq .Fl Fl gname .Qq . On extract, it causes user and group names in the archive to be ignored in favor of the numeric user and group ids. On create, it causes user and group names to not be stored in the archive. .It Fl O , Fl Fl to-stdout (x, t modes only) In extract (-x) mode, files will be written to standard out rather than being extracted to disk. In list (-t) mode, the file listing will be written to stderr rather than the usual stdout. .It Fl o (x mode) Use the user and group of the user running the program rather than those specified in the archive. Note that this has no significance unless .Fl p is specified, and the program is being run by the root user. In this case, the file modes and flags from the archive will be restored, but ACLs or owner information in the archive will be discarded. .It Fl o (c, r, u mode) A synonym for .Fl Fl format Ar ustar .It Fl Fl older Ar date (c, r, u modes only) Only include files and directories older than the specified date. This compares ctime entries. .It Fl Fl older-mtime Ar date (c, r, u modes only) Like .Fl Fl older , except it compares mtime entries instead of ctime entries. .It Fl Fl older-than Pa file (c, r, u modes only) Only include files and directories older than the specified file. This compares ctime entries. .It Fl Fl older-mtime-than Pa file (c, r, u modes only) Like .Fl Fl older-than , except it compares mtime entries instead of ctime entries. .It Fl Fl one-file-system (c, r, and u modes) Do not cross mount points. .It Fl Fl options Ar options Select optional behaviors for particular modules. The argument is a text string containing comma-separated keywords and values. These are passed to the modules that handle particular formats to control how those formats will behave. Each option has one of the following forms: .Bl -tag -compact -width indent .It Ar key=value The key will be set to the specified value in every module that supports it. Modules that do not support this key will ignore it. .It Ar key The key will be enabled in every module that supports it. This is equivalent to .Ar key Ns Cm =1 . .It Ar !key The key will be disabled in every module that supports it. .It Ar module:key=value , Ar module:key , Ar module:!key As above, but the corresponding key and value will be provided only to modules whose name matches .Ar module . .El The currently supported modules and keys are: .Bl -tag -compact -width indent .It Cm iso9660:joliet Support Joliet extensions. This is enabled by default, use .Cm !joliet or .Cm iso9660:!joliet to disable. .It Cm iso9660:rockridge Support Rock Ridge extensions. This is enabled by default, use .Cm !rockridge or .Cm iso9660:!rockridge to disable. .It Cm gzip:compression-level A decimal integer from 1 to 9 specifying the gzip compression level. .It Cm gzip:timestamp Store timestamp. This is enabled by default, use .Cm !timestamp or .Cm gzip:!timestamp to disable. .It Cm lrzip:compression Ns = Ns Ar type Use .Ar type as compression method. Supported values are bzip2, gzip, lzo (ultra fast), and zpaq (best, extremely slow). .It Cm lrzip:compression-level A decimal integer from 1 to 9 specifying the lrzip compression level. .It Cm lz4:compression-level A decimal integer from 1 to 9 specifying the lzop compression level. .It Cm lz4:stream-checksum Enable stream checksum. This is by default, use .Cm lz4:!stream-checksum to disable. .It Cm lz4:block-checksum Enable block checksum (Disabled by default). .It Cm lz4:block-size A decimal integer from 4 to 7 specifying the lz4 compression block size (7 is set by default). .It Cm lz4:block-dependence Use the previous block of the block being compressed for a compression dictionary to improve compression ratio. +.It Cm zstd:compression-level +A decimal integer from 1 to 22 specifying the zstd compression level. .It Cm lzop:compression-level A decimal integer from 1 to 9 specifying the lzop compression level. .It Cm xz:compression-level A decimal integer from 0 to 9 specifying the xz compression level. .It Cm mtree: Ns Ar keyword The mtree writer module allows you to specify which mtree keywords will be included in the output. Supported keywords include: .Cm cksum , Cm device , Cm flags , Cm gid , Cm gname , Cm indent , .Cm link , Cm md5 , Cm mode , Cm nlink , Cm rmd160 , Cm sha1 , Cm sha256 , .Cm sha384 , Cm sha512 , Cm size , Cm time , Cm uid , Cm uname . The default is equivalent to: .Dq device, flags, gid, gname, link, mode, nlink, size, time, type, uid, uname . .It Cm mtree:all Enables all of the above keywords. You can also use .Cm mtree:!all to disable all keywords. .It Cm mtree:use-set Enable generation of .Cm /set lines in the output. .It Cm mtree:indent Produce human-readable output by indenting options and splitting lines to fit into 80 columns. .It Cm zip:compression Ns = Ns Ar type Use .Ar type as compression method. Supported values are store (uncompressed) and deflate (gzip algorithm). .It Cm zip:encryption Enable encryption using traditional zip encryption. .It Cm zip:encryption Ns = Ns Ar type Use .Ar type as encryption type. Supported values are zipcrypt (traditional zip encryption), aes128 (WinZip AES-128 encryption) and aes256 (WinZip AES-256 encryption). .It Cm read_concatenated_archives Ignore zeroed blocks in the archive, which occurs when multiple tar archives have been concatenated together. Without this option, only the contents of the first concatenated archive would be read. This option is comparable to the .Fl i , Fl Fl ignore-zeros option of GNU tar. .El If a provided option is not supported by any module, that is a fatal error. .It Fl P , Fl Fl absolute-paths Preserve pathnames. By default, absolute pathnames (those that begin with a / character) have the leading slash removed both when creating archives and extracting from them. Also, .Nm will refuse to extract archive entries whose pathnames contain .Pa .. or whose target directory would be altered by a symlink. This option suppresses these behaviors. .It Fl p , Fl Fl insecure , Fl Fl preserve-permissions (x mode only) Preserve file permissions. Attempt to restore the full permissions, including owner, file modes, ACLs, extended attributes and extended file flags, if available, for each item extracted from the archive. This is te reverse of .Fl Fl no-same-permissions and the default if .Nm is being run by root and can be partially overridden by also specifying .Fl Fl no-acls , .Fl Fl no-fflags , .Fl Fl no-mac-metadata or .Fl Fl no-xattrs . .It Fl Fl passphrase Ar passphrase The .Pa passphrase is used to extract or create an encrypted archive. Currently, zip is the only supported format that supports encryption. You shouldn't use this option unless you realize how insecure use of this option is. .It Fl Fl posix (c, r, u mode only) Synonym for .Fl Fl format Ar pax .It Fl q , Fl Fl fast-read (x and t mode only) Extract or list only the first archive entry that matches each pattern or filename operand. Exit as soon as each specified pattern or filename has been matched. By default, the archive is always read to the very end, since there can be multiple entries with the same name and, by convention, later entries overwrite earlier entries. This option is provided as a performance optimization. .It Fl S (x mode only) Extract files as sparse files. For every block on disk, check first if it contains only NULL bytes and seek over it otherwise. This works similar to the conv=sparse option of dd. .It Fl s Ar pattern Modify file or archive member names according to .Pa pattern . The pattern has the format .Ar /old/new/ Ns Op ghHprRsS where .Ar old is a basic regular expression, .Ar new is the replacement string of the matched part, and the optional trailing letters modify how the replacement is handled. If .Ar old is not matched, the pattern is skipped. Within .Ar new , ~ is substituted with the match, \e1 to \e9 with the content of the corresponding captured group. The optional trailing g specifies that matching should continue after the matched part and stop on the first unmatched pattern. The optional trailing s specifies that the pattern applies to the value of symbolic links. The optional trailing p specifies that after a successful substitution the original path name and the new path name should be printed to standard error. Optional trailing H, R, or S characters suppress substitutions for hardlink targets, regular filenames, or symlink targets, respectively. Optional trailing h, r, or s characters enable substitutions for hardlink targets, regular filenames, or symlink targets, respectively. The default is .Ar hrs which applies substitutions to all names. In particular, it is never necessary to specify h, r, or s. .It Fl Fl same-owner (x mode only) Extract owner and group IDs. This is the reverse of .Fl Fl no-same-owner and the default behavior if .Nm is run as root. .It Fl Fl strip-components Ar count Remove the specified number of leading path elements. Pathnames with fewer elements will be silently skipped. Note that the pathname is edited after checking inclusion/exclusion patterns but before security checks. .It Fl T Ar filename , Fl Fl files-from Ar filename In x or t mode, .Nm will read the list of names to be extracted from .Pa filename . In c mode, .Nm will read names to be archived from .Pa filename . The special name .Dq -C on a line by itself will cause the current directory to be changed to the directory specified on the following line. Names are terminated by newlines unless .Fl Fl null is specified. Note that .Fl Fl null also disables the special handling of lines containing .Dq -C . Note: If you are generating lists of files using .Xr find 1 , you probably want to use .Fl n as well. .It Fl Fl totals (c, r, u modes only) After archiving all files, print a summary to stderr. .It Fl U , Fl Fl unlink , Fl Fl unlink-first (x mode only) Unlink files before creating them. This can be a minor performance optimization if most files already exist, but can make things slower if most files do not already exist. This flag also causes .Nm to remove intervening directory symlinks instead of reporting an error. See the SECURITY section below for more details. .It Fl Fl uid Ar id Use the provided user id number and ignore the user name from the archive. On create, if .Fl Fl uname is not also specified, the user name will be set to match the user id. .It Fl Fl uname Ar name Use the provided user name. On extract, this overrides the user name in the archive; if the provided user name does not exist on the system, it will be ignored and the user id (from the archive or from the .Fl Fl uid option) will be used instead. On create, this sets the user name that will be stored in the archive; the name is not verified against the system user database. .It Fl Fl use-compress-program Ar program Pipe the input (in x or t mode) or the output (in c mode) through .Pa program instead of using the builtin compression support. .It Fl v , Fl Fl verbose Produce verbose output. In create and extract modes, .Nm will list each file name as it is read from or written to the archive. In list mode, .Nm will produce output similar to that of .Xr ls 1 . An additional .Fl v option will also provide ls-like details in create and extract mode. .It Fl Fl version Print version of .Nm and .Nm libarchive , and exit. .It Fl w , Fl Fl confirmation , Fl Fl interactive Ask for confirmation for every action. .It Fl X Ar filename , Fl Fl exclude-from Ar filename Read a list of exclusion patterns from the specified file. See .Fl Fl exclude for more information about the handling of exclusions. .It Fl Fl xattrs (c, r, u, x modes only) Archive or extract extended attributes. This is the reverse of .Fl Fl no-xattrs and the default behavior in c, r, and u modes or if .Nm is run in x mode as root. .It Fl y (c mode only) Compress the resulting archive with .Xr bzip2 1 . In extract or list modes, this option is ignored. -Note that, unlike other +Note that this .Nm tar -implementations, this implementation recognizes bzip2 compression -automatically when reading archives. +implementation recognizes bzip2 compression automatically when reading +archives. .It Fl Z , Fl Fl compress , Fl Fl uncompress (c mode only) Compress the resulting archive with .Xr compress 1 . In extract or list modes, this option is ignored. -Note that, unlike other +Note that this .Nm tar -implementations, this implementation recognizes compress compression -automatically when reading archives. +implementation recognizes compress compression automatically when reading +archives. .It Fl z , Fl Fl gunzip , Fl Fl gzip (c mode only) Compress the resulting archive with .Xr gzip 1 . In extract or list modes, this option is ignored. -Note that, unlike other +Note that this .Nm tar -implementations, this implementation recognizes gzip compression -automatically when reading archives. +implementation recognizes gzip compression automatically when reading +archives. .El .Sh ENVIRONMENT The following environment variables affect the execution of .Nm : .Bl -tag -width ".Ev BLOCKSIZE" .It Ev TAR_READER_OPTIONS The default options for format readers and compression readers. The .Fl Fl options option overrides this. .It Ev TAR_WRITER_OPTIONS The default options for format writers and compression writers. The .Fl Fl options option overrides this. .It Ev LANG The locale to use. See .Xr environ 7 for more information. .It Ev TAPE The default device. The .Fl f option overrides this. Please see the description of the .Fl f option above for more details. .It Ev TZ The timezone to use when displaying dates. See .Xr environ 7 for more information. .El .Sh EXIT STATUS .Ex -std .Sh EXAMPLES The following creates a new archive called .Ar file.tar.gz that contains two files .Ar source.c and .Ar source.h : .Dl Nm Fl czf Pa file.tar.gz Pa source.c Pa source.h .Pp To view a detailed table of contents for this archive: .Dl Nm Fl tvf Pa file.tar.gz .Pp To extract all entries from the archive on the default tape drive: .Dl Nm Fl x .Pp To examine the contents of an ISO 9660 cdrom image: .Dl Nm Fl tf Pa image.iso .Pp To move file hierarchies, invoke .Nm as .Dl Nm Fl cf Pa - Fl C Pa srcdir\ . | Nm Fl xpf Pa - Fl C Pa destdir or more traditionally .Dl cd srcdir \&; Nm Fl cf Pa -\ . | ( cd destdir \&; Nm Fl xpf Pa - ) .Pp In create mode, the list of files and directories to be archived can also include directory change instructions of the form .Cm -C Ns Pa foo/baz and archive inclusions of the form .Cm @ Ns Pa archive-file . For example, the command line .Dl Nm Fl c Fl f Pa new.tar Pa foo1 Cm @ Ns Pa old.tgz Cm -C Ns Pa /tmp Pa foo2 will create a new archive .Pa new.tar . .Nm will read the file .Pa foo1 from the current directory and add it to the output archive. It will then read each entry from .Pa old.tgz and add those entries to the output archive. Finally, it will switch to the .Pa /tmp directory and add .Pa foo2 to the output archive. .Pp An input file in .Xr mtree 5 format can be used to create an output archive with arbitrary ownership, permissions, or names that differ from existing data on disk: .Pp .Bd -literal -offset indent $ cat input.mtree #mtree usr/bin uid=0 gid=0 mode=0755 type=dir usr/bin/ls uid=0 gid=0 mode=0755 type=file content=myls $ tar -cvf output.tar @input.mtree .Ed .Pp The .Fl Fl newer and .Fl Fl newer-mtime switches accept a variety of common date and time specifications, including .Dq 12 Mar 2005 7:14:29pm , .Dq 2005-03-12 19:14 , .Dq 5 minutes ago , and .Dq 19:14 PST May 1 . .Pp The .Fl Fl options argument can be used to control various details of archive generation or reading. For example, you can generate mtree output which only contains .Cm type , Cm time , and .Cm uid keywords: .Dl Nm Fl cf Pa file.tar Fl Fl format=mtree Fl Fl options='!all,type,time,uid' Pa dir or you can set the compression level used by gzip or xz compression: .Dl Nm Fl czf Pa file.tar Fl Fl options='compression-level=9' . For more details, see the explanation of the .Fn archive_read_set_options and .Fn archive_write_set_options API calls that are described in .Xr archive_read 3 and .Xr archive_write 3 . .Sh COMPATIBILITY The bundled-arguments format is supported for compatibility with historic implementations. It consists of an initial word (with no leading - character) in which each character indicates an option. Arguments follow as separate words. The order of the arguments must match the order of the corresponding characters in the bundled command word. For example, .Dl Nm Cm tbf 32 Pa file.tar specifies three flags .Cm t , .Cm b , and .Cm f . The .Cm b and .Cm f flags both require arguments, so there must be two additional items on the command line. The .Ar 32 is the argument to the .Cm b flag, and .Ar file.tar is the argument to the .Cm f flag. .Pp The mode options c, r, t, u, and x and the options b, f, l, m, o, v, and w comply with SUSv2. .Pp For maximum portability, scripts that invoke .Nm tar should use the bundled-argument format above, should limit themselves to the .Cm c , .Cm t , and .Cm x modes, and the .Cm b , .Cm f , .Cm m , .Cm v , and .Cm w options. .Pp Additional long options are provided to improve compatibility with other tar implementations. .Sh SECURITY Certain security issues are common to many archiving programs, including .Nm . In particular, carefully-crafted archives can request that .Nm extract files to locations outside of the target directory. This can potentially be used to cause unwitting users to overwrite files they did not intend to overwrite. If the archive is being extracted by the superuser, any file on the system can potentially be overwritten. There are three ways this can happen. Although .Nm has mechanisms to protect against each one, savvy users should be aware of the implications: .Bl -bullet -width indent .It Archive entries can have absolute pathnames. By default, .Nm removes the leading .Pa / character from filenames before restoring them to guard against this problem. .It Archive entries can have pathnames that include .Pa .. components. By default, .Nm will not extract files containing .Pa .. components in their pathname. .It Archive entries can exploit symbolic links to restore files to other directories. An archive can restore a symbolic link to another directory, then use that link to restore a file into that directory. To guard against this, .Nm checks each extracted path for symlinks. If the final path element is a symlink, it will be removed and replaced with the archive entry. If .Fl U is specified, any intermediate symlink will also be unconditionally removed. If neither .Fl U nor .Fl P is specified, .Nm will refuse to extract the entry. .El To protect yourself, you should be wary of any archives that come from untrusted sources. You should examine the contents of an archive with .Dl Nm Fl tf Pa filename before extraction. You should use the .Fl k option to ensure that .Nm will not overwrite any existing files or the .Fl U option to remove any pre-existing files. You should generally not extract archives while running with super-user privileges. Note that the .Fl P option to .Nm disables the security checks above and allows you to extract an archive while preserving any absolute pathnames, .Pa .. components, or symlinks to other directories. .Sh SEE ALSO .Xr bzip2 1 , .Xr compress 1 , .Xr cpio 1 , .Xr gzip 1 , .Xr mt 1 , .Xr pax 1 , .Xr shar 1 , .Xr xz 1 , .Xr libarchive 3 , .Xr libarchive-formats 5 , .Xr tar 5 .Sh STANDARDS There is no current POSIX standard for the tar command; it appeared in .St -p1003.1-96 but was dropped from .St -p1003.1-2001 . The options supported by this implementation were developed by surveying a number of existing tar implementations as well as the old POSIX specification for tar and the current POSIX specification for pax. .Pp The ustar and pax interchange file formats are defined by .St -p1003.1-2001 for the pax command. .Sh HISTORY A .Nm tar command appeared in Seventh Edition Unix, which was released in January, 1979. There have been numerous other implementations, many of which extended the file format. John Gilmore's .Nm pdtar public-domain implementation (circa November, 1987) was quite influential, and formed the basis of GNU tar. GNU tar was included as the standard system tar in .Fx beginning with .Fx 1.0 . .Pp This is a complete re-implementation based on the .Xr libarchive 3 library. It was first released with .Fx 5.4 in May, 2005. .Sh BUGS This program follows .St -p1003.1-96 for the definition of the .Fl l option. Note that GNU tar prior to version 1.15 treated .Fl l as a synonym for the .Fl Fl one-file-system option. .Pp The .Fl C Pa dir option may differ from historic implementations. .Pp All archive output is written in correctly-sized blocks, even if the output is being compressed. Whether or not the last output block is padded to a full block size varies depending on the format and the output device. For tar and cpio formats, the last block of output is padded to a full block size if the output is being written to standard output or to a character or block device such as a tape drive. If the output is being written to a regular file, the last block will not be padded. Many compressors, including .Xr gzip 1 and .Xr bzip2 1 , complain about the null padding when decompressing an archive created by .Nm , although they still extract it correctly. .Pp The compression and decompression is implemented internally, so there may be insignificant differences between the compressed output generated by .Dl Nm Fl czf Pa - file and that generated by .Dl Nm Fl cf Pa - file | Nm gzip .Pp The default should be to read and write archives to the standard I/O paths, but tradition (and POSIX) dictates otherwise. .Pp The .Cm r and .Cm u modes require that the archive be uncompressed and located in a regular file on disk. Other archives can be modified using .Cm c mode with the .Pa @archive-file extension. .Pp To archive a file called .Pa @foo or .Pa -foo you must specify it as .Pa ./@foo or .Pa ./-foo , respectively. .Pp In create mode, a leading .Pa ./ is always removed. A leading .Pa / is stripped unless the .Fl P option is specified. .Pp There needs to be better support for file selection on both create and extract. .Pp There is not yet any support for multi-volume archives. .Pp Converting between dissimilar archive formats (such as tar and cpio) using the .Cm @ Ns Pa - convention can cause hard link information to be lost. (This is a consequence of the incompatible ways that different archive formats store hardlink information.) Index: stable/10/contrib/libarchive/tar/bsdtar.c =================================================================== --- stable/10/contrib/libarchive/tar/bsdtar.c (revision 324417) +++ stable/10/contrib/libarchive/tar/bsdtar.c (revision 324418) @@ -1,986 +1,988 @@ /*- * Copyright (c) 2003-2008 Tim Kientzle * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "bsdtar_platform.h" __FBSDID("$FreeBSD$"); #ifdef HAVE_SYS_PARAM_H #include #endif #ifdef HAVE_SYS_STAT_H #include #endif #ifdef HAVE_COPYFILE_H #include #endif #ifdef HAVE_ERRNO_H #include #endif #ifdef HAVE_FCNTL_H #include #endif #ifdef HAVE_LANGINFO_H #include #endif #ifdef HAVE_LOCALE_H #include #endif #ifdef HAVE_PATHS_H #include #endif #ifdef HAVE_SIGNAL_H #include #endif #include #ifdef HAVE_STDLIB_H #include #endif #ifdef HAVE_STRING_H #include #endif #ifdef HAVE_TIME_H #include #endif #ifdef HAVE_UNISTD_H #include #endif #include "bsdtar.h" #include "err.h" /* * Per POSIX.1-1988, tar defaults to reading/writing archives to/from * the default tape device for the system. Pick something reasonable here. */ #ifdef __linux #define _PATH_DEFTAPE "/dev/st0" #endif #if defined(_WIN32) && !defined(__CYGWIN__) #define _PATH_DEFTAPE "\\\\.\\tape0" #endif #if defined(__APPLE__) #undef _PATH_DEFTAPE #define _PATH_DEFTAPE "-" /* Mac OS has no tape support, default to stdio. */ #endif #ifndef _PATH_DEFTAPE #define _PATH_DEFTAPE "/dev/tape" #endif #ifdef __MINGW32__ int _CRT_glob = 0; /* Disable broken CRT globbing. */ #endif #if defined(HAVE_SIGACTION) && (defined(SIGINFO) || defined(SIGUSR1)) static volatile int siginfo_occurred; static void siginfo_handler(int sig) { (void)sig; /* UNUSED */ siginfo_occurred = 1; } int need_report(void) { int r = siginfo_occurred; siginfo_occurred = 0; return (r); } #else int need_report(void) { return (0); } #endif static void long_help(void) __LA_DEAD; static void only_mode(struct bsdtar *, const char *opt, const char *valid); static void set_mode(struct bsdtar *, char opt); static void version(void) __LA_DEAD; /* A basic set of security flags to request from libarchive. */ #define SECURITY \ (ARCHIVE_EXTRACT_SECURE_SYMLINKS \ | ARCHIVE_EXTRACT_SECURE_NODOTDOT) int main(int argc, char **argv) { struct bsdtar *bsdtar, bsdtar_storage; int opt, t; char compression, compression2; const char *compression_name, *compression2_name; const char *compress_program; char possible_help_request; char buff[16]; /* * Use a pointer for consistency, but stack-allocated storage * for ease of cleanup. */ bsdtar = &bsdtar_storage; memset(bsdtar, 0, sizeof(*bsdtar)); bsdtar->fd = -1; /* Mark as "unused" */ bsdtar->gid = -1; bsdtar->uid = -1; bsdtar->flags = 0; compression = compression2 = '\0'; compression_name = compression2_name = NULL; compress_program = NULL; #if defined(HAVE_SIGACTION) { /* Set up signal handling. */ struct sigaction sa; sa.sa_handler = siginfo_handler; sigemptyset(&sa.sa_mask); sa.sa_flags = 0; #ifdef SIGINFO if (sigaction(SIGINFO, &sa, NULL)) lafe_errc(1, errno, "sigaction(SIGINFO) failed"); #endif #ifdef SIGUSR1 /* ... and treat SIGUSR1 the same way as SIGINFO. */ if (sigaction(SIGUSR1, &sa, NULL)) lafe_errc(1, errno, "sigaction(SIGUSR1) failed"); #endif #ifdef SIGPIPE /* Ignore SIGPIPE signals. */ sa.sa_handler = SIG_IGN; sigaction(SIGPIPE, &sa, NULL); #endif } #endif /* Set lafe_progname before calling lafe_warnc. */ lafe_setprogname(*argv, "bsdtar"); #if HAVE_SETLOCALE if (setlocale(LC_ALL, "") == NULL) lafe_warnc(0, "Failed to set default locale"); #endif #if defined(HAVE_NL_LANGINFO) && defined(HAVE_D_MD_ORDER) bsdtar->day_first = (*nl_langinfo(D_MD_ORDER) == 'd'); #endif possible_help_request = 0; /* Look up uid of current user for future reference */ bsdtar->user_uid = geteuid(); /* Default: open tape drive. */ bsdtar->filename = getenv("TAPE"); if (bsdtar->filename == NULL) bsdtar->filename = _PATH_DEFTAPE; /* Default block size settings. */ bsdtar->bytes_per_block = DEFAULT_BYTES_PER_BLOCK; /* Allow library to default this unless user specifies -b. */ bsdtar->bytes_in_last_block = -1; /* Default: preserve mod time on extract */ bsdtar->extract_flags = ARCHIVE_EXTRACT_TIME; /* Default: Perform basic security checks. */ bsdtar->extract_flags |= SECURITY; #ifndef _WIN32 /* On POSIX systems, assume --same-owner and -p when run by * the root user. This doesn't make any sense on Windows. */ if (bsdtar->user_uid == 0) { /* --same-owner */ bsdtar->extract_flags |= ARCHIVE_EXTRACT_OWNER; /* -p */ bsdtar->extract_flags |= ARCHIVE_EXTRACT_PERM; bsdtar->extract_flags |= ARCHIVE_EXTRACT_ACL; bsdtar->extract_flags |= ARCHIVE_EXTRACT_XATTR; bsdtar->extract_flags |= ARCHIVE_EXTRACT_FFLAGS; bsdtar->extract_flags |= ARCHIVE_EXTRACT_MAC_METADATA; } #endif /* * Enable Mac OS "copyfile()" extension by default. * This has no effect on other platforms. */ bsdtar->readdisk_flags |= ARCHIVE_READDISK_MAC_COPYFILE; #ifdef COPYFILE_DISABLE_VAR if (getenv(COPYFILE_DISABLE_VAR)) bsdtar->readdisk_flags &= ~ARCHIVE_READDISK_MAC_COPYFILE; #endif #if defined(__APPLE__) /* * On Mac OS ACLs are archived with copyfile() (--mac-metadata) * Translation to NFSv4 ACLs has to be requested explicitly with --acls */ bsdtar->readdisk_flags |= ARCHIVE_READDISK_NO_ACL; #endif bsdtar->matching = archive_match_new(); if (bsdtar->matching == NULL) lafe_errc(1, errno, "Out of memory"); bsdtar->cset = cset_new(); if (bsdtar->cset == NULL) lafe_errc(1, errno, "Out of memory"); bsdtar->argv = argv; bsdtar->argc = argc; /* * Comments following each option indicate where that option * originated: SUSv2, POSIX, GNU tar, star, etc. If there's * no such comment, then I don't know of anyone else who * implements that option. */ while ((opt = bsdtar_getopt(bsdtar)) != -1) { switch (opt) { case 'a': /* GNU tar */ bsdtar->flags |= OPTFLAG_AUTO_COMPRESS; break; case OPTION_ACLS: /* GNU tar */ bsdtar->extract_flags |= ARCHIVE_EXTRACT_ACL; bsdtar->readdisk_flags &= ~ARCHIVE_READDISK_NO_ACL; bsdtar->flags |= OPTFLAG_ACLS; break; case 'B': /* GNU tar */ /* libarchive doesn't need this; just ignore it. */ break; case 'b': /* SUSv2 */ t = atoi(bsdtar->argument); if (t <= 0 || t > 8192) lafe_errc(1, 0, "Argument to -b is out of range (1..8192)"); bsdtar->bytes_per_block = 512 * t; /* Explicit -b forces last block size. */ bsdtar->bytes_in_last_block = bsdtar->bytes_per_block; break; case OPTION_B64ENCODE: if (compression2 != '\0') lafe_errc(1, 0, "Can't specify both --uuencode and " "--b64encode"); compression2 = opt; compression2_name = "b64encode"; break; case 'C': /* GNU tar */ if (strlen(bsdtar->argument) == 0) lafe_errc(1, 0, "Meaningless option: -C ''"); set_chdir(bsdtar, bsdtar->argument); break; case 'c': /* SUSv2 */ set_mode(bsdtar, opt); break; case OPTION_CHECK_LINKS: /* GNU tar */ bsdtar->flags |= OPTFLAG_WARN_LINKS; break; case OPTION_CHROOT: /* NetBSD */ bsdtar->flags |= OPTFLAG_CHROOT; break; case OPTION_CLEAR_NOCHANGE_FFLAGS: bsdtar->extract_flags |= ARCHIVE_EXTRACT_CLEAR_NOCHANGE_FFLAGS; break; case OPTION_EXCLUDE: /* GNU tar */ if (archive_match_exclude_pattern( bsdtar->matching, bsdtar->argument) != ARCHIVE_OK) lafe_errc(1, 0, "Couldn't exclude %s\n", bsdtar->argument); break; case OPTION_FFLAGS: bsdtar->extract_flags |= ARCHIVE_EXTRACT_FFLAGS; bsdtar->readdisk_flags &= ~ARCHIVE_READDISK_NO_FFLAGS; bsdtar->flags |= OPTFLAG_FFLAGS; break; case OPTION_FORMAT: /* GNU tar, others */ cset_set_format(bsdtar->cset, bsdtar->argument); break; case 'f': /* SUSv2 */ bsdtar->filename = bsdtar->argument; break; case OPTION_GID: /* cpio */ t = atoi(bsdtar->argument); if (t < 0) lafe_errc(1, 0, "Argument to --gid must be positive"); bsdtar->gid = t; break; case OPTION_GNAME: /* cpio */ bsdtar->gname = bsdtar->argument; break; case OPTION_GRZIP: if (compression != '\0') lafe_errc(1, 0, "Can't specify both -%c and -%c", opt, compression); compression = opt; compression_name = "grzip"; break; case 'H': /* BSD convention */ bsdtar->symlink_mode = 'H'; break; case 'h': /* Linux Standards Base, gtar; synonym for -L */ bsdtar->symlink_mode = 'L'; /* Hack: -h by itself is the "help" command. */ possible_help_request = 1; break; case OPTION_HELP: /* GNU tar, others */ long_help(); exit(0); break; case OPTION_HFS_COMPRESSION: /* Mac OS X v10.6 or later */ bsdtar->extract_flags |= ARCHIVE_EXTRACT_HFS_COMPRESSION_FORCED; break; case OPTION_IGNORE_ZEROS: bsdtar->flags |= OPTFLAG_IGNORE_ZEROS; break; case 'I': /* GNU tar */ /* * TODO: Allow 'names' to come from an archive, * not just a text file. Design a good UI for * allowing names and mode/owner to be read * from an archive, with contents coming from * disk. This can be used to "refresh" an * archive or to design archives with special * permissions without having to create those * permissions on disk. */ bsdtar->names_from_file = bsdtar->argument; break; case OPTION_INCLUDE: /* * No one else has the @archive extension, so * no one else needs this to filter entries * when transforming archives. */ if (archive_match_include_pattern(bsdtar->matching, bsdtar->argument) != ARCHIVE_OK) lafe_errc(1, 0, "Failed to add %s to inclusion list", bsdtar->argument); break; case 'j': /* GNU tar */ if (compression != '\0') lafe_errc(1, 0, "Can't specify both -%c and -%c", opt, compression); compression = opt; compression_name = "bzip2"; break; case 'J': /* GNU tar 1.21 and later */ if (compression != '\0') lafe_errc(1, 0, "Can't specify both -%c and -%c", opt, compression); compression = opt; compression_name = "xz"; break; case 'k': /* GNU tar */ bsdtar->extract_flags |= ARCHIVE_EXTRACT_NO_OVERWRITE; break; case OPTION_KEEP_NEWER_FILES: /* GNU tar */ bsdtar->extract_flags |= ARCHIVE_EXTRACT_NO_OVERWRITE_NEWER; break; case 'L': /* BSD convention */ bsdtar->symlink_mode = 'L'; break; case 'l': /* SUSv2 and GNU tar beginning with 1.16 */ /* GNU tar 1.13 used -l for --one-file-system */ bsdtar->flags |= OPTFLAG_WARN_LINKS; break; case OPTION_LRZIP: case OPTION_LZ4: case OPTION_LZIP: /* GNU tar beginning with 1.23 */ case OPTION_LZMA: /* GNU tar beginning with 1.20 */ case OPTION_LZOP: /* GNU tar beginning with 1.21 */ + case OPTION_ZSTD: if (compression != '\0') lafe_errc(1, 0, "Can't specify both -%c and -%c", opt, compression); compression = opt; switch (opt) { case OPTION_LRZIP: compression_name = "lrzip"; break; case OPTION_LZ4: compression_name = "lz4"; break; - case OPTION_LZIP: compression_name = "lzip"; break; - case OPTION_LZMA: compression_name = "lzma"; break; - case OPTION_LZOP: compression_name = "lzop"; break; + case OPTION_LZIP: compression_name = "lzip"; break; + case OPTION_LZMA: compression_name = "lzma"; break; + case OPTION_LZOP: compression_name = "lzop"; break; + case OPTION_ZSTD: compression_name = "zstd"; break; } break; case 'm': /* SUSv2 */ bsdtar->extract_flags &= ~ARCHIVE_EXTRACT_TIME; break; case OPTION_MAC_METADATA: /* Mac OS X */ bsdtar->readdisk_flags |= ARCHIVE_READDISK_MAC_COPYFILE; bsdtar->extract_flags |= ARCHIVE_EXTRACT_MAC_METADATA; bsdtar->flags |= OPTFLAG_MAC_METADATA; break; case 'n': /* GNU tar */ bsdtar->flags |= OPTFLAG_NO_SUBDIRS; break; /* * Selecting files by time: * --newer-?time='date' Only files newer than 'date' * --newer-?time-than='file' Only files newer than time * on specified file (useful for incremental backups) */ case OPTION_NEWER_CTIME: /* GNU tar */ if (archive_match_include_date(bsdtar->matching, ARCHIVE_MATCH_CTIME | ARCHIVE_MATCH_NEWER, bsdtar->argument) != ARCHIVE_OK) lafe_errc(1, 0, "Error : %s", archive_error_string(bsdtar->matching)); break; case OPTION_NEWER_CTIME_THAN: if (archive_match_include_file_time(bsdtar->matching, ARCHIVE_MATCH_CTIME | ARCHIVE_MATCH_NEWER, bsdtar->argument) != ARCHIVE_OK) lafe_errc(1, 0, "Error : %s", archive_error_string(bsdtar->matching)); break; case OPTION_NEWER_MTIME: /* GNU tar */ if (archive_match_include_date(bsdtar->matching, ARCHIVE_MATCH_MTIME | ARCHIVE_MATCH_NEWER, bsdtar->argument) != ARCHIVE_OK) lafe_errc(1, 0, "Error : %s", archive_error_string(bsdtar->matching)); break; case OPTION_NEWER_MTIME_THAN: if (archive_match_include_file_time(bsdtar->matching, ARCHIVE_MATCH_MTIME | ARCHIVE_MATCH_NEWER, bsdtar->argument) != ARCHIVE_OK) lafe_errc(1, 0, "Error : %s", archive_error_string(bsdtar->matching)); break; case OPTION_NODUMP: /* star */ bsdtar->readdisk_flags |= ARCHIVE_READDISK_HONOR_NODUMP; break; case OPTION_NOPRESERVE_HFS_COMPRESSION: /* Mac OS X v10.6 or later */ bsdtar->extract_flags |= ARCHIVE_EXTRACT_NO_HFS_COMPRESSION; break; case OPTION_NO_ACLS: /* GNU tar */ bsdtar->extract_flags &= ~ARCHIVE_EXTRACT_ACL; bsdtar->readdisk_flags |= ARCHIVE_READDISK_NO_ACL; bsdtar->flags |= OPTFLAG_NO_ACLS; break; case OPTION_NO_FFLAGS: bsdtar->extract_flags &= ~ARCHIVE_EXTRACT_FFLAGS; bsdtar->readdisk_flags |= ARCHIVE_READDISK_NO_FFLAGS; bsdtar->flags |= OPTFLAG_NO_FFLAGS; break; case OPTION_NO_MAC_METADATA: /* Mac OS X */ bsdtar->readdisk_flags &= ~ARCHIVE_READDISK_MAC_COPYFILE; bsdtar->extract_flags &= ~ARCHIVE_EXTRACT_MAC_METADATA; bsdtar->flags |= OPTFLAG_NO_MAC_METADATA; break; case OPTION_NO_SAME_OWNER: /* GNU tar */ bsdtar->extract_flags &= ~ARCHIVE_EXTRACT_OWNER; break; case OPTION_NO_SAME_PERMISSIONS: /* GNU tar */ bsdtar->extract_flags &= ~ARCHIVE_EXTRACT_PERM; bsdtar->extract_flags &= ~ARCHIVE_EXTRACT_ACL; bsdtar->extract_flags &= ~ARCHIVE_EXTRACT_XATTR; bsdtar->extract_flags &= ~ARCHIVE_EXTRACT_FFLAGS; bsdtar->extract_flags &= ~ARCHIVE_EXTRACT_MAC_METADATA; break; case OPTION_NO_XATTRS: /* GNU tar */ bsdtar->extract_flags &= ~ARCHIVE_EXTRACT_XATTR; bsdtar->readdisk_flags |= ARCHIVE_READDISK_NO_XATTR; bsdtar->flags |= OPTFLAG_NO_XATTRS; break; case OPTION_NULL: /* GNU tar */ bsdtar->flags |= OPTFLAG_NULL; break; case OPTION_NUMERIC_OWNER: /* GNU tar */ bsdtar->uname = ""; bsdtar->gname = ""; bsdtar->flags |= OPTFLAG_NUMERIC_OWNER; break; case 'O': /* GNU tar */ bsdtar->flags |= OPTFLAG_STDOUT; break; case 'o': /* SUSv2 and GNU conflict here, but not fatally */ bsdtar->flags |= OPTFLAG_O; break; /* * Selecting files by time: * --older-?time='date' Only files older than 'date' * --older-?time-than='file' Only files older than time * on specified file */ case OPTION_OLDER_CTIME: if (archive_match_include_date(bsdtar->matching, ARCHIVE_MATCH_CTIME | ARCHIVE_MATCH_OLDER, bsdtar->argument) != ARCHIVE_OK) lafe_errc(1, 0, "Error : %s", archive_error_string(bsdtar->matching)); break; case OPTION_OLDER_CTIME_THAN: if (archive_match_include_file_time(bsdtar->matching, ARCHIVE_MATCH_CTIME | ARCHIVE_MATCH_OLDER, bsdtar->argument) != ARCHIVE_OK) lafe_errc(1, 0, "Error : %s", archive_error_string(bsdtar->matching)); break; case OPTION_OLDER_MTIME: if (archive_match_include_date(bsdtar->matching, ARCHIVE_MATCH_MTIME | ARCHIVE_MATCH_OLDER, bsdtar->argument) != ARCHIVE_OK) lafe_errc(1, 0, "Error : %s", archive_error_string(bsdtar->matching)); break; case OPTION_OLDER_MTIME_THAN: if (archive_match_include_file_time(bsdtar->matching, ARCHIVE_MATCH_MTIME | ARCHIVE_MATCH_OLDER, bsdtar->argument) != ARCHIVE_OK) lafe_errc(1, 0, "Error : %s", archive_error_string(bsdtar->matching)); break; case OPTION_ONE_FILE_SYSTEM: /* GNU tar */ bsdtar->readdisk_flags |= ARCHIVE_READDISK_NO_TRAVERSE_MOUNTS; break; case OPTION_OPTIONS: bsdtar->option_options = bsdtar->argument; break; #if 0 /* * The common BSD -P option is not necessary, since * our default is to archive symlinks, not follow * them. This is convenient, as -P conflicts with GNU * tar anyway. */ case 'P': /* BSD convention */ /* Default behavior, no option necessary. */ break; #endif case 'P': /* GNU tar */ bsdtar->extract_flags &= ~SECURITY; bsdtar->flags |= OPTFLAG_ABSOLUTE_PATHS; break; case 'p': /* GNU tar, star */ bsdtar->extract_flags |= ARCHIVE_EXTRACT_PERM; bsdtar->extract_flags |= ARCHIVE_EXTRACT_ACL; bsdtar->extract_flags |= ARCHIVE_EXTRACT_XATTR; bsdtar->extract_flags |= ARCHIVE_EXTRACT_FFLAGS; bsdtar->extract_flags |= ARCHIVE_EXTRACT_MAC_METADATA; break; case OPTION_PASSPHRASE: bsdtar->passphrase = bsdtar->argument; break; case OPTION_POSIX: /* GNU tar */ cset_set_format(bsdtar->cset, "pax"); break; case 'q': /* FreeBSD GNU tar --fast-read, NetBSD -q */ bsdtar->flags |= OPTFLAG_FAST_READ; break; case 'r': /* SUSv2 */ set_mode(bsdtar, opt); break; case 'S': /* NetBSD pax-as-tar */ bsdtar->extract_flags |= ARCHIVE_EXTRACT_SPARSE; break; case 's': /* NetBSD pax-as-tar */ #if defined(HAVE_REGEX_H) || defined(HAVE_PCREPOSIX_H) add_substitution(bsdtar, bsdtar->argument); #else lafe_warnc(0, "-s is not supported by this version of bsdtar"); usage(); #endif break; case OPTION_SAME_OWNER: /* GNU tar */ bsdtar->extract_flags |= ARCHIVE_EXTRACT_OWNER; break; case OPTION_STRIP_COMPONENTS: /* GNU tar 1.15 */ errno = 0; bsdtar->strip_components = strtol(bsdtar->argument, NULL, 0); if (errno) lafe_errc(1, 0, "Invalid --strip-components argument: %s", bsdtar->argument); break; case 'T': /* GNU tar */ bsdtar->names_from_file = bsdtar->argument; break; case 't': /* SUSv2 */ set_mode(bsdtar, opt); bsdtar->verbose++; break; case OPTION_TOTALS: /* GNU tar */ bsdtar->flags |= OPTFLAG_TOTALS; break; case 'U': /* GNU tar */ bsdtar->extract_flags |= ARCHIVE_EXTRACT_UNLINK; bsdtar->flags |= OPTFLAG_UNLINK_FIRST; break; case 'u': /* SUSv2 */ set_mode(bsdtar, opt); break; case OPTION_UID: /* cpio */ t = atoi(bsdtar->argument); if (t < 0) lafe_errc(1, 0, "Argument to --uid must be positive"); bsdtar->uid = t; break; case OPTION_UNAME: /* cpio */ bsdtar->uname = bsdtar->argument; break; case OPTION_UUENCODE: if (compression2 != '\0') lafe_errc(1, 0, "Can't specify both --uuencode and " "--b64encode"); compression2 = opt; compression2_name = "uuencode"; break; case 'v': /* SUSv2 */ bsdtar->verbose++; break; case OPTION_VERSION: /* GNU convention */ version(); break; #if 0 /* * The -W longopt feature is handled inside of * bsdtar_getopt(), so -W is not available here. */ case 'W': /* Obscure GNU convention. */ break; #endif case 'w': /* SUSv2 */ bsdtar->flags |= OPTFLAG_INTERACTIVE; break; case 'X': /* GNU tar */ if (archive_match_exclude_pattern_from_file( bsdtar->matching, bsdtar->argument, 0) != ARCHIVE_OK) lafe_errc(1, 0, "Error : %s", archive_error_string(bsdtar->matching)); break; case 'x': /* SUSv2 */ set_mode(bsdtar, opt); break; case OPTION_XATTRS: /* GNU tar */ bsdtar->extract_flags |= ARCHIVE_EXTRACT_XATTR; bsdtar->readdisk_flags &= ~ARCHIVE_READDISK_NO_XATTR; bsdtar->flags |= OPTFLAG_XATTRS; break; case 'y': /* FreeBSD version of GNU tar */ if (compression != '\0') lafe_errc(1, 0, "Can't specify both -%c and -%c", opt, compression); compression = opt; compression_name = "bzip2"; break; case 'Z': /* GNU tar */ if (compression != '\0') lafe_errc(1, 0, "Can't specify both -%c and -%c", opt, compression); compression = opt; compression_name = "compress"; break; case 'z': /* GNU tar, star, many others */ if (compression != '\0') lafe_errc(1, 0, "Can't specify both -%c and -%c", opt, compression); compression = opt; compression_name = "gzip"; break; case OPTION_USE_COMPRESS_PROGRAM: compress_program = bsdtar->argument; break; default: usage(); } } /* * Sanity-check options. */ /* If no "real" mode was specified, treat -h as --help. */ if ((bsdtar->mode == '\0') && possible_help_request) { long_help(); exit(0); } /* Otherwise, a mode is required. */ if (bsdtar->mode == '\0') lafe_errc(1, 0, "Must specify one of -c, -r, -t, -u, -x"); /* Check boolean options only permitted in certain modes. */ if (bsdtar->flags & OPTFLAG_AUTO_COMPRESS) only_mode(bsdtar, "-a", "c"); if (bsdtar->readdisk_flags & ARCHIVE_READDISK_NO_TRAVERSE_MOUNTS) only_mode(bsdtar, "--one-file-system", "cru"); if (bsdtar->flags & OPTFLAG_FAST_READ) only_mode(bsdtar, "--fast-read", "xt"); if (bsdtar->extract_flags & ARCHIVE_EXTRACT_HFS_COMPRESSION_FORCED) only_mode(bsdtar, "--hfsCompression", "x"); if (bsdtar->extract_flags & ARCHIVE_EXTRACT_NO_HFS_COMPRESSION) only_mode(bsdtar, "--nopreserveHFSCompression", "x"); if (bsdtar->readdisk_flags & ARCHIVE_READDISK_HONOR_NODUMP) only_mode(bsdtar, "--nodump", "cru"); if (bsdtar->flags & OPTFLAG_ACLS) only_mode(bsdtar, "--acls", "crux"); if (bsdtar->flags & OPTFLAG_NO_ACLS) only_mode(bsdtar, "--no-acls", "crux"); if (bsdtar->flags & OPTFLAG_XATTRS) only_mode(bsdtar, "--xattrs", "crux"); if (bsdtar->flags & OPTFLAG_NO_XATTRS) only_mode(bsdtar, "--no-xattrs", "crux"); if (bsdtar->flags & OPTFLAG_FFLAGS) only_mode(bsdtar, "--fflags", "crux"); if (bsdtar->flags & OPTFLAG_NO_FFLAGS) only_mode(bsdtar, "--no-fflags", "crux"); if (bsdtar->flags & OPTFLAG_MAC_METADATA) only_mode(bsdtar, "--mac-metadata", "crux"); if (bsdtar->flags & OPTFLAG_NO_MAC_METADATA) only_mode(bsdtar, "--no-mac-metadata", "crux"); if (bsdtar->flags & OPTFLAG_O) { switch (bsdtar->mode) { case 'c': /* * In GNU tar, -o means "old format." The * "ustar" format is the closest thing * supported by libarchive. */ cset_set_format(bsdtar->cset, "ustar"); /* TODO: bsdtar->create_format = "v7"; */ break; case 'x': /* POSIX-compatible behavior. */ bsdtar->flags |= OPTFLAG_NO_OWNER; bsdtar->extract_flags &= ~ARCHIVE_EXTRACT_OWNER; break; default: only_mode(bsdtar, "-o", "xc"); break; } } if (bsdtar->flags & OPTFLAG_NO_SUBDIRS) only_mode(bsdtar, "-n", "cru"); if (bsdtar->flags & OPTFLAG_STDOUT) only_mode(bsdtar, "-O", "xt"); if (bsdtar->flags & OPTFLAG_UNLINK_FIRST) only_mode(bsdtar, "-U", "x"); if (bsdtar->flags & OPTFLAG_WARN_LINKS) only_mode(bsdtar, "--check-links", "cr"); if ((bsdtar->flags & OPTFLAG_AUTO_COMPRESS) && cset_auto_compress(bsdtar->cset, bsdtar->filename)) { /* Ignore specified compressions if auto-compress works. */ compression = '\0'; compression2 = '\0'; } /* Check other parameters only permitted in certain modes. */ if (compress_program != NULL) { only_mode(bsdtar, "--use-compress-program", "cxt"); cset_add_filter_program(bsdtar->cset, compress_program); /* Ignore specified compressions. */ compression = '\0'; compression2 = '\0'; } if (compression != '\0') { switch (compression) { case 'J': case 'j': case 'y': case 'Z': case 'z': strcpy(buff, "-?"); buff[1] = compression; break; default: strcpy(buff, "--"); strcat(buff, compression_name); break; } only_mode(bsdtar, buff, "cxt"); cset_add_filter(bsdtar->cset, compression_name); } if (compression2 != '\0') { strcpy(buff, "--"); strcat(buff, compression2_name); only_mode(bsdtar, buff, "cxt"); cset_add_filter(bsdtar->cset, compression2_name); } if (cset_get_format(bsdtar->cset) != NULL) only_mode(bsdtar, "--format", "cru"); if (bsdtar->symlink_mode != '\0') { strcpy(buff, "-?"); buff[1] = bsdtar->symlink_mode; only_mode(bsdtar, buff, "cru"); } /* Filename "-" implies stdio. */ if (strcmp(bsdtar->filename, "-") == 0) bsdtar->filename = NULL; switch(bsdtar->mode) { case 'c': tar_mode_c(bsdtar); break; case 'r': tar_mode_r(bsdtar); break; case 't': tar_mode_t(bsdtar); break; case 'u': tar_mode_u(bsdtar); break; case 'x': tar_mode_x(bsdtar); break; } archive_match_free(bsdtar->matching); #if defined(HAVE_REGEX_H) || defined(HAVE_PCREPOSIX_H) cleanup_substitution(bsdtar); #endif cset_free(bsdtar->cset); passphrase_free(bsdtar->ppbuff); if (bsdtar->return_value != 0) lafe_warnc(0, "Error exit delayed from previous errors."); return (bsdtar->return_value); } static void set_mode(struct bsdtar *bsdtar, char opt) { if (bsdtar->mode != '\0' && bsdtar->mode != opt) lafe_errc(1, 0, "Can't specify both -%c and -%c", opt, bsdtar->mode); bsdtar->mode = opt; } /* * Verify that the mode is correct. */ static void only_mode(struct bsdtar *bsdtar, const char *opt, const char *valid_modes) { if (strchr(valid_modes, bsdtar->mode) == NULL) lafe_errc(1, 0, "Option %s is not permitted in mode -%c", opt, bsdtar->mode); } void usage(void) { const char *p; p = lafe_getprogname(); fprintf(stderr, "Usage:\n"); fprintf(stderr, " List: %s -tf \n", p); fprintf(stderr, " Extract: %s -xf \n", p); fprintf(stderr, " Create: %s -cf [filenames...]\n", p); fprintf(stderr, " Help: %s --help\n", p); exit(1); } static void version(void) { printf("bsdtar %s - %s\n", BSDTAR_VERSION_STRING, archive_version_details()); exit(0); } static const char *long_help_msg = "First option must be a mode specifier:\n" " -c Create -r Add/Replace -t List -u Update -x Extract\n" "Common Options:\n" " -b # Use # 512-byte records per I/O block\n" " -f Location of archive (default " _PATH_DEFTAPE ")\n" " -v Verbose\n" " -w Interactive\n" "Create: %p -c [options] [ | | @ | -C ]\n" " , add these items to archive\n" " -z, -j, -J, --lzma Compress archive with gzip/bzip2/xz/lzma\n" " --format {ustar|pax|cpio|shar} Select archive format\n" " --exclude Skip files that match pattern\n" " -C Change to before processing remaining files\n" " @ Add entries from to output\n" "List: %p -t [options] []\n" " If specified, list only entries that match\n" "Extract: %p -x [options] []\n" " If specified, extract only entries that match\n" " -k Keep (don't overwrite) existing files\n" " -m Don't restore modification times\n" " -O Write entries to stdout, don't restore to disk\n" " -p Restore permissions (including ACLs, owner, file flags)\n"; /* * Note that the word 'bsdtar' will always appear in the first line * of output. * * In particular, /bin/sh scripts that need to test for the presence * of bsdtar can use the following template: * * if (tar --help 2>&1 | grep bsdtar >/dev/null 2>&1 ) then \ * echo bsdtar; else echo not bsdtar; fi */ static void long_help(void) { const char *prog; const char *p; prog = lafe_getprogname(); fflush(stderr); p = (strcmp(prog,"bsdtar") != 0) ? "(bsdtar)" : ""; printf("%s%s: manipulate archive files\n", prog, p); for (p = long_help_msg; *p != '\0'; p++) { if (*p == '%') { if (p[1] == 'p') { fputs(prog, stdout); p++; } else putchar('%'); } else putchar(*p); } version(); } Index: stable/10/contrib/libarchive/tar/bsdtar.h =================================================================== --- stable/10/contrib/libarchive/tar/bsdtar.h (revision 324417) +++ stable/10/contrib/libarchive/tar/bsdtar.h (revision 324418) @@ -1,224 +1,225 @@ /*- * Copyright (c) 2003-2007 Tim Kientzle * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. * * $FreeBSD$ */ #include "bsdtar_platform.h" #include #define DEFAULT_BYTES_PER_BLOCK (20*512) #define ENV_READER_OPTIONS "TAR_READER_OPTIONS" #define ENV_WRITER_OPTIONS "TAR_WRITER_OPTIONS" #define IGNORE_WRONG_MODULE_NAME "__ignore_wrong_module_name__," struct creation_set; /* * The internal state for the "bsdtar" program. * * Keeping all of the state in a structure like this simplifies memory * leak testing (at exit, anything left on the heap is suspect). A * pointer to this structure is passed to most bsdtar internal * functions. */ struct bsdtar { /* Options */ const char *filename; /* -f filename */ char *pending_chdir; /* -C dir */ const char *names_from_file; /* -T file */ int bytes_per_block; /* -b block_size */ int bytes_in_last_block; /* See -b handling. */ int verbose; /* -v */ unsigned int flags; /* Bitfield of boolean options */ int extract_flags; /* Flags for extract operation */ int readdisk_flags; /* Flags for read disk operation */ int strip_components; /* Remove this many leading dirs */ int gid; /* --gid */ const char *gname; /* --gname */ int uid; /* --uid */ const char *uname; /* --uname */ const char *passphrase; /* --passphrase */ char mode; /* Program mode: 'c', 't', 'r', 'u', 'x' */ char symlink_mode; /* H or L, per BSD conventions */ const char *option_options; /* --options */ char day_first; /* show day before month in -tv output */ struct creation_set *cset; /* Option parser state */ int getopt_state; char *getopt_word; /* If >= 0, then close this when done. */ int fd; /* Miscellaneous state information */ int argc; char **argv; const char *argument; size_t gs_width; /* For 'list_item' in read.c */ size_t u_width; /* for 'list_item' in read.c */ uid_t user_uid; /* UID running this program */ int return_value; /* Value returned by main() */ char warned_lead_slash; /* Already displayed warning */ char next_line_is_dir; /* Used for -C parsing in -cT */ /* * Data for various subsystems. Full definitions are located in * the file where they are used. */ struct archive *diskreader; /* for write.c */ struct archive_entry_linkresolver *resolver; /* for write.c */ struct archive_dir *archive_dir; /* for write.c */ struct name_cache *gname_cache; /* for write.c */ char *buff; /* for write.c */ size_t buff_size; /* for write.c */ int first_fs; /* for write.c */ struct archive *matching; /* for matching.c */ struct security *security; /* for read.c */ struct name_cache *uname_cache; /* for write.c */ struct siginfo_data *siginfo; /* for siginfo.c */ struct substitution *substitution; /* for subst.c */ char *ppbuff; /* for util.c */ }; /* Options for flags bitfield */ #define OPTFLAG_AUTO_COMPRESS (0x00000001) /* -a */ #define OPTFLAG_ABSOLUTE_PATHS (0x00000002) /* -P */ #define OPTFLAG_CHROOT (0x00000004) /* --chroot */ #define OPTFLAG_FAST_READ (0x00000008) /* --fast-read */ #define OPTFLAG_IGNORE_ZEROS (0x00000010) /* --ignore-zeros */ #define OPTFLAG_INTERACTIVE (0x00000020) /* -w */ #define OPTFLAG_NO_OWNER (0x00000040) /* -o */ #define OPTFLAG_NO_SUBDIRS (0x00000080) /* -n */ #define OPTFLAG_NULL (0x00000100) /* --null */ #define OPTFLAG_NUMERIC_OWNER (0x00000200) /* --numeric-owner */ #define OPTFLAG_O (0x00000400) /* -o */ #define OPTFLAG_STDOUT (0x00000800) /* -O */ #define OPTFLAG_TOTALS (0x00001000) /* --totals */ #define OPTFLAG_UNLINK_FIRST (0x00002000) /* -U */ #define OPTFLAG_WARN_LINKS (0x00004000) /* --check-links */ #define OPTFLAG_NO_XATTRS (0x00008000) /* --no-xattrs */ #define OPTFLAG_XATTRS (0x00010000) /* --xattrs */ #define OPTFLAG_NO_ACLS (0x00020000) /* --no-acls */ #define OPTFLAG_ACLS (0x00040000) /* --acls */ #define OPTFLAG_NO_FFLAGS (0x00080000) /* --no-fflags */ #define OPTFLAG_FFLAGS (0x00100000) /* --fflags */ #define OPTFLAG_NO_MAC_METADATA (0x00200000) /* --no-mac-metadata */ #define OPTFLAG_MAC_METADATA (0x00400000) /* --mac-metadata */ /* Fake short equivalents for long options that otherwise lack them. */ enum { OPTION_ACLS = 1, OPTION_B64ENCODE, OPTION_CHECK_LINKS, OPTION_CHROOT, OPTION_CLEAR_NOCHANGE_FFLAGS, OPTION_EXCLUDE, OPTION_FFLAGS, OPTION_FORMAT, OPTION_GID, OPTION_GNAME, OPTION_GRZIP, OPTION_HELP, OPTION_HFS_COMPRESSION, OPTION_IGNORE_ZEROS, OPTION_INCLUDE, OPTION_KEEP_NEWER_FILES, OPTION_LRZIP, OPTION_LZ4, OPTION_LZIP, OPTION_LZMA, OPTION_LZOP, OPTION_MAC_METADATA, OPTION_NEWER_CTIME, OPTION_NEWER_CTIME_THAN, OPTION_NEWER_MTIME, OPTION_NEWER_MTIME_THAN, OPTION_NODUMP, OPTION_NOPRESERVE_HFS_COMPRESSION, OPTION_NO_ACLS, OPTION_NO_FFLAGS, OPTION_NO_MAC_METADATA, OPTION_NO_SAME_OWNER, OPTION_NO_SAME_PERMISSIONS, OPTION_NO_XATTRS, OPTION_NULL, OPTION_NUMERIC_OWNER, OPTION_OLDER_CTIME, OPTION_OLDER_CTIME_THAN, OPTION_OLDER_MTIME, OPTION_OLDER_MTIME_THAN, OPTION_ONE_FILE_SYSTEM, OPTION_OPTIONS, OPTION_PASSPHRASE, OPTION_POSIX, OPTION_SAME_OWNER, OPTION_STRIP_COMPONENTS, OPTION_TOTALS, OPTION_UID, OPTION_UNAME, OPTION_USE_COMPRESS_PROGRAM, OPTION_UUENCODE, OPTION_VERSION, - OPTION_XATTRS + OPTION_XATTRS, + OPTION_ZSTD, }; int bsdtar_getopt(struct bsdtar *); void do_chdir(struct bsdtar *); int edit_pathname(struct bsdtar *, struct archive_entry *); int need_report(void); int pathcmp(const char *a, const char *b); void safe_fprintf(FILE *, const char *fmt, ...) __LA_PRINTF(2, 3); void set_chdir(struct bsdtar *, const char *newdir); const char *tar_i64toa(int64_t); void tar_mode_c(struct bsdtar *bsdtar); void tar_mode_r(struct bsdtar *bsdtar); void tar_mode_t(struct bsdtar *bsdtar); void tar_mode_u(struct bsdtar *bsdtar); void tar_mode_x(struct bsdtar *bsdtar); void usage(void) __LA_DEAD; int yes(const char *fmt, ...) __LA_PRINTF(1, 2); #if defined(HAVE_REGEX_H) || defined(HAVE_PCREPOSIX_H) void add_substitution(struct bsdtar *, const char *); int apply_substitution(struct bsdtar *, const char *, char **, int, int); void cleanup_substitution(struct bsdtar *); #endif void cset_add_filter(struct creation_set *, const char *); void cset_add_filter_program(struct creation_set *, const char *); int cset_auto_compress(struct creation_set *, const char *); void cset_free(struct creation_set *); const char * cset_get_format(struct creation_set *); struct creation_set *cset_new(void); int cset_read_support_filter_program(struct creation_set *, struct archive *); void cset_set_format(struct creation_set *, const char *); int cset_write_add_filters(struct creation_set *, struct archive *, const void **); const char * passphrase_callback(struct archive *, void *); void passphrase_free(char *); void list_item_verbose(struct bsdtar *, FILE *, struct archive_entry *); Index: stable/10/contrib/libarchive/tar/cmdline.c =================================================================== --- stable/10/contrib/libarchive/tar/cmdline.c (revision 324417) +++ stable/10/contrib/libarchive/tar/cmdline.c (revision 324418) @@ -1,413 +1,414 @@ /*- * Copyright (c) 2003-2008 Tim Kientzle * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ /* * Command line parser for tar. */ #include "bsdtar_platform.h" __FBSDID("$FreeBSD$"); #ifdef HAVE_ERRNO_H #include #endif #ifdef HAVE_STDLIB_H #include #endif #ifdef HAVE_STRING_H #include #endif #include "bsdtar.h" #include "err.h" /* * Short options for tar. Please keep this sorted. */ static const char *short_options = "aBb:C:cf:HhI:JjkLlmnOoPpqrSs:T:tUuvW:wX:xyZz"; /* * Long options for tar. Please keep this list sorted. * * The symbolic names for options that lack a short equivalent are * defined in bsdtar.h. Also note that so far I've found no need * to support optional arguments to long options. That would be * a small change to the code below. */ static const struct bsdtar_option { const char *name; int required; /* 1 if this option requires an argument. */ int equivalent; /* Equivalent short option. */ } tar_longopts[] = { { "absolute-paths", 0, 'P' }, { "append", 0, 'r' }, { "acls", 0, OPTION_ACLS }, { "auto-compress", 0, 'a' }, { "b64encode", 0, OPTION_B64ENCODE }, { "block-size", 1, 'b' }, { "blocking-factor", 1, 'b' }, { "bunzip2", 0, 'j' }, { "bzip", 0, 'j' }, { "bzip2", 0, 'j' }, { "cd", 1, 'C' }, { "check-links", 0, OPTION_CHECK_LINKS }, { "chroot", 0, OPTION_CHROOT }, { "clear-nochange-fflags", 0, OPTION_CLEAR_NOCHANGE_FFLAGS }, { "compress", 0, 'Z' }, { "confirmation", 0, 'w' }, { "create", 0, 'c' }, { "dereference", 0, 'L' }, { "directory", 1, 'C' }, { "disable-copyfile", 0, OPTION_NO_MAC_METADATA }, { "exclude", 1, OPTION_EXCLUDE }, { "exclude-from", 1, 'X' }, { "extract", 0, 'x' }, { "fast-read", 0, 'q' }, { "fflags", 0, OPTION_FFLAGS }, { "file", 1, 'f' }, { "files-from", 1, 'T' }, { "format", 1, OPTION_FORMAT }, { "gid", 1, OPTION_GID }, { "gname", 1, OPTION_GNAME }, { "grzip", 0, OPTION_GRZIP }, { "gunzip", 0, 'z' }, { "gzip", 0, 'z' }, { "help", 0, OPTION_HELP }, { "hfsCompression", 0, OPTION_HFS_COMPRESSION }, { "ignore-zeros", 0, OPTION_IGNORE_ZEROS }, { "include", 1, OPTION_INCLUDE }, { "insecure", 0, 'P' }, { "interactive", 0, 'w' }, { "keep-newer-files", 0, OPTION_KEEP_NEWER_FILES }, { "keep-old-files", 0, 'k' }, { "list", 0, 't' }, { "lrzip", 0, OPTION_LRZIP }, { "lz4", 0, OPTION_LZ4 }, { "lzip", 0, OPTION_LZIP }, { "lzma", 0, OPTION_LZMA }, { "lzop", 0, OPTION_LZOP }, { "mac-metadata", 0, OPTION_MAC_METADATA }, { "modification-time", 0, 'm' }, { "newer", 1, OPTION_NEWER_CTIME }, { "newer-ctime", 1, OPTION_NEWER_CTIME }, { "newer-ctime-than", 1, OPTION_NEWER_CTIME_THAN }, { "newer-mtime", 1, OPTION_NEWER_MTIME }, { "newer-mtime-than", 1, OPTION_NEWER_MTIME_THAN }, { "newer-than", 1, OPTION_NEWER_CTIME_THAN }, { "no-acls", 0, OPTION_NO_ACLS }, { "no-fflags", 0, OPTION_NO_FFLAGS }, { "no-mac-metadata", 0, OPTION_NO_MAC_METADATA }, { "no-recursion", 0, 'n' }, { "no-same-owner", 0, OPTION_NO_SAME_OWNER }, { "no-same-permissions", 0, OPTION_NO_SAME_PERMISSIONS }, { "no-xattr", 0, OPTION_NO_XATTRS }, { "no-xattrs", 0, OPTION_NO_XATTRS }, { "nodump", 0, OPTION_NODUMP }, { "nopreserveHFSCompression",0, OPTION_NOPRESERVE_HFS_COMPRESSION }, { "norecurse", 0, 'n' }, { "null", 0, OPTION_NULL }, { "numeric-owner", 0, OPTION_NUMERIC_OWNER }, { "older", 1, OPTION_OLDER_CTIME }, { "older-ctime", 1, OPTION_OLDER_CTIME }, { "older-ctime-than", 1, OPTION_OLDER_CTIME_THAN }, { "older-mtime", 1, OPTION_OLDER_MTIME }, { "older-mtime-than", 1, OPTION_OLDER_MTIME_THAN }, { "older-than", 1, OPTION_OLDER_CTIME_THAN }, { "one-file-system", 0, OPTION_ONE_FILE_SYSTEM }, { "options", 1, OPTION_OPTIONS }, { "passphrase", 1, OPTION_PASSPHRASE }, { "posix", 0, OPTION_POSIX }, { "preserve-permissions", 0, 'p' }, { "read-full-blocks", 0, 'B' }, { "same-owner", 0, OPTION_SAME_OWNER }, { "same-permissions", 0, 'p' }, { "strip-components", 1, OPTION_STRIP_COMPONENTS }, { "to-stdout", 0, 'O' }, { "totals", 0, OPTION_TOTALS }, { "uid", 1, OPTION_UID }, { "uname", 1, OPTION_UNAME }, { "uncompress", 0, 'Z' }, { "unlink", 0, 'U' }, { "unlink-first", 0, 'U' }, { "update", 0, 'u' }, { "use-compress-program", 1, OPTION_USE_COMPRESS_PROGRAM }, { "uuencode", 0, OPTION_UUENCODE }, { "verbose", 0, 'v' }, { "version", 0, OPTION_VERSION }, { "xattrs", 0, OPTION_XATTRS }, { "xz", 0, 'J' }, + { "zstd", 0, OPTION_ZSTD }, { NULL, 0, 0 } }; /* * This getopt implementation has two key features that common * getopt_long() implementations lack. Apart from those, it's a * straightforward option parser, considerably simplified by not * needing to support the wealth of exotic getopt_long() features. It * has, of course, been shamelessly tailored for bsdtar. (If you're * looking for a generic getopt_long() implementation for your * project, I recommend Gregory Pietsch's public domain getopt_long() * implementation.) The two additional features are: * * Old-style tar arguments: The original tar implementation treated * the first argument word as a list of single-character option * letters. All arguments follow as separate words. For example, * tar xbf 32 /dev/tape * Here, the "xbf" is three option letters, "32" is the argument for * "b" and "/dev/tape" is the argument for "f". We support this usage * if the first command-line argument does not begin with '-'. We * also allow regular short and long options to follow, e.g., * tar xbf 32 /dev/tape -P --format=pax * * -W long options: There's an obscure GNU convention (only rarely * supported even there) that allows "-W option=argument" as an * alternative way to support long options. This was supported in * early bsdtar as a way to access long options on platforms that did * not support getopt_long() and is preserved here for backwards * compatibility. (Of course, if I'd started with a custom * command-line parser from the beginning, I would have had normal * long option support on every platform so that hack wouldn't have * been necessary. Oh, well. Some mistakes you just have to live * with.) * * TODO: We should be able to use this to pull files and intermingled * options (such as -C) from the command line in write mode. That * will require a little rethinking of the argument handling in * bsdtar.c. * * TODO: If we want to support arbitrary command-line options from -T * input (as GNU tar does), we may need to extend this to handle option * words from sources other than argv/argc. I'm not really sure if I * like that feature of GNU tar, so it's certainly not a priority. */ int bsdtar_getopt(struct bsdtar *bsdtar) { enum { state_start = 0, state_old_tar, state_next_word, state_short, state_long }; const struct bsdtar_option *popt, *match = NULL, *match2 = NULL; const char *p, *long_prefix = "--"; size_t optlength; int opt = '?'; int required = 0; bsdtar->argument = NULL; /* First time through, initialize everything. */ if (bsdtar->getopt_state == state_start) { /* Skip program name. */ ++bsdtar->argv; --bsdtar->argc; if (*bsdtar->argv == NULL) return (-1); /* Decide between "new style" and "old style" arguments. */ if (bsdtar->argv[0][0] == '-') { bsdtar->getopt_state = state_next_word; } else { bsdtar->getopt_state = state_old_tar; bsdtar->getopt_word = *bsdtar->argv++; --bsdtar->argc; } } /* * We're parsing old-style tar arguments */ if (bsdtar->getopt_state == state_old_tar) { /* Get the next option character. */ opt = *bsdtar->getopt_word++; if (opt == '\0') { /* New-style args can follow old-style. */ bsdtar->getopt_state = state_next_word; } else { /* See if it takes an argument. */ p = strchr(short_options, opt); if (p == NULL) return ('?'); if (p[1] == ':') { bsdtar->argument = *bsdtar->argv; if (bsdtar->argument == NULL) { lafe_warnc(0, "Option %c requires an argument", opt); return ('?'); } ++bsdtar->argv; --bsdtar->argc; } } } /* * We're ready to look at the next word in argv. */ if (bsdtar->getopt_state == state_next_word) { /* No more arguments, so no more options. */ if (bsdtar->argv[0] == NULL) return (-1); /* Doesn't start with '-', so no more options. */ if (bsdtar->argv[0][0] != '-') return (-1); /* "--" marks end of options; consume it and return. */ if (strcmp(bsdtar->argv[0], "--") == 0) { ++bsdtar->argv; --bsdtar->argc; return (-1); } /* Get next word for parsing. */ bsdtar->getopt_word = *bsdtar->argv++; --bsdtar->argc; if (bsdtar->getopt_word[1] == '-') { /* Set up long option parser. */ bsdtar->getopt_state = state_long; bsdtar->getopt_word += 2; /* Skip leading '--' */ } else { /* Set up short option parser. */ bsdtar->getopt_state = state_short; ++bsdtar->getopt_word; /* Skip leading '-' */ } } /* * We're parsing a group of POSIX-style single-character options. */ if (bsdtar->getopt_state == state_short) { /* Peel next option off of a group of short options. */ opt = *bsdtar->getopt_word++; if (opt == '\0') { /* End of this group; recurse to get next option. */ bsdtar->getopt_state = state_next_word; return bsdtar_getopt(bsdtar); } /* Does this option take an argument? */ p = strchr(short_options, opt); if (p == NULL) return ('?'); if (p[1] == ':') required = 1; /* If it takes an argument, parse that. */ if (required) { /* If arg is run-in, bsdtar->getopt_word already points to it. */ if (bsdtar->getopt_word[0] == '\0') { /* Otherwise, pick up the next word. */ bsdtar->getopt_word = *bsdtar->argv; if (bsdtar->getopt_word == NULL) { lafe_warnc(0, "Option -%c requires an argument", opt); return ('?'); } ++bsdtar->argv; --bsdtar->argc; } if (opt == 'W') { bsdtar->getopt_state = state_long; long_prefix = "-W "; /* For clearer errors. */ } else { bsdtar->getopt_state = state_next_word; bsdtar->argument = bsdtar->getopt_word; } } } /* We're reading a long option, including -W long=arg convention. */ if (bsdtar->getopt_state == state_long) { /* After this long option, we'll be starting a new word. */ bsdtar->getopt_state = state_next_word; /* Option name ends at '=' if there is one. */ p = strchr(bsdtar->getopt_word, '='); if (p != NULL) { optlength = (size_t)(p - bsdtar->getopt_word); bsdtar->argument = (char *)(uintptr_t)(p + 1); } else { optlength = strlen(bsdtar->getopt_word); } /* Search the table for an unambiguous match. */ for (popt = tar_longopts; popt->name != NULL; popt++) { /* Short-circuit if first chars don't match. */ if (popt->name[0] != bsdtar->getopt_word[0]) continue; /* If option is a prefix of name in table, record it.*/ if (strncmp(bsdtar->getopt_word, popt->name, optlength) == 0) { match2 = match; /* Record up to two matches. */ match = popt; /* If it's an exact match, we're done. */ if (strlen(popt->name) == optlength) { match2 = NULL; /* Forget the others. */ break; } } } /* Fail if there wasn't a unique match. */ if (match == NULL) { lafe_warnc(0, "Option %s%s is not supported", long_prefix, bsdtar->getopt_word); return ('?'); } if (match2 != NULL) { lafe_warnc(0, "Ambiguous option %s%s (matches --%s and --%s)", long_prefix, bsdtar->getopt_word, match->name, match2->name); return ('?'); } /* We've found a unique match; does it need an argument? */ if (match->required) { /* Argument required: get next word if necessary. */ if (bsdtar->argument == NULL) { bsdtar->argument = *bsdtar->argv; if (bsdtar->argument == NULL) { lafe_warnc(0, "Option %s%s requires an argument", long_prefix, match->name); return ('?'); } ++bsdtar->argv; --bsdtar->argc; } } else { /* Argument forbidden: fail if there is one. */ if (bsdtar->argument != NULL) { lafe_warnc(0, "Option %s%s does not allow an argument", long_prefix, match->name); return ('?'); } } return (match->equivalent); } return (opt); } Index: stable/10/contrib/libarchive/tar/creation_set.c =================================================================== --- stable/10/contrib/libarchive/tar/creation_set.c (revision 324417) +++ stable/10/contrib/libarchive/tar/creation_set.c (revision 324418) @@ -1,316 +1,318 @@ /*- * Copyright (c) 2012 Michihiro NAKAJIMA * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "bsdtar_platform.h" __FBSDID("$FreeBSD$"); #ifdef HAVE_STDLIB_H #include #endif #ifdef HAVE_STRING_H #include #endif #include "bsdtar.h" #include "err.h" struct creation_set { char *create_format; struct filter_set { int program; /* Set 1 if filter is a program name */ char *filter_name; } *filters; int filter_count; }; struct suffix_code_t { const char *suffix; const char *form; }; static const char * get_suffix_code(const struct suffix_code_t *tbl, const char *suffix) { int i; if (suffix == NULL) return (NULL); for (i = 0; tbl[i].suffix != NULL; i++) { if (strcmp(tbl[i].suffix, suffix) == 0) return (tbl[i].form); } return (NULL); } static const char * get_filter_code(const char *suffix) { /* A pair of suffix and compression/filter. */ static const struct suffix_code_t filters[] = { { ".Z", "compress" }, { ".bz2", "bzip2" }, { ".gz", "gzip" }, { ".grz", "grzip" }, { ".lrz", "lrzip" }, { ".lz", "lzip" }, { ".lz4", "lz4" }, { ".lzo", "lzop" }, { ".lzma", "lzma" }, { ".uu", "uuencode" }, { ".xz", "xz" }, + { ".zst", "zstd"}, { NULL, NULL } }; - + return get_suffix_code(filters, suffix); } static const char * get_format_code(const char *suffix) { /* A pair of suffix and format. */ static const struct suffix_code_t formats[] = { { ".7z", "7zip" }, { ".ar", "arbsd" }, { ".cpio", "cpio" }, { ".iso", "iso9960" }, { ".mtree", "mtree" }, { ".shar", "shar" }, { ".tar", "paxr" }, { ".warc", "warc" }, { ".xar", "xar" }, { ".zip", "zip" }, { NULL, NULL } }; return get_suffix_code(formats, suffix); } static const char * decompose_alias(const char *suffix) { static const struct suffix_code_t alias[] = { { ".taz", ".tar.gz" }, { ".tgz", ".tar.gz" }, { ".tbz", ".tar.bz2" }, { ".tbz2", ".tar.bz2" }, { ".tz2", ".tar.bz2" }, { ".tlz", ".tar.lzma" }, { ".txz", ".tar.xz" }, { ".tzo", ".tar.lzo" }, { ".taZ", ".tar.Z" }, { ".tZ", ".tar.Z" }, + { ".tzst", ".tar.zst" }, { NULL, NULL } }; return get_suffix_code(alias, suffix); } static void _cset_add_filter(struct creation_set *cset, int program, const char *filter) { struct filter_set *new_ptr; char *new_filter; new_ptr = (struct filter_set *)realloc(cset->filters, sizeof(*cset->filters) * (cset->filter_count + 1)); if (new_ptr == NULL) lafe_errc(1, 0, "No memory"); new_filter = strdup(filter); if (new_filter == NULL) lafe_errc(1, 0, "No memory"); cset->filters = new_ptr; cset->filters[cset->filter_count].program = program; cset->filters[cset->filter_count].filter_name = new_filter; cset->filter_count++; } void cset_add_filter(struct creation_set *cset, const char *filter) { _cset_add_filter(cset, 0, filter); } void cset_add_filter_program(struct creation_set *cset, const char *filter) { _cset_add_filter(cset, 1, filter); } int cset_read_support_filter_program(struct creation_set *cset, struct archive *a) { int cnt = 0, i; for (i = 0; i < cset->filter_count; i++) { if (cset->filters[i].program) { archive_read_support_filter_program(a, cset->filters[i].filter_name); ++cnt; } } return (cnt); } int cset_write_add_filters(struct creation_set *cset, struct archive *a, const void **filter_name) { int cnt = 0, i, r; for (i = 0; i < cset->filter_count; i++) { if (cset->filters[i].program) r = archive_write_add_filter_program(a, cset->filters[i].filter_name); else r = archive_write_add_filter_by_name(a, cset->filters[i].filter_name); if (r < ARCHIVE_WARN) { *filter_name = cset->filters[i].filter_name; return (r); } ++cnt; } return (cnt); } void cset_set_format(struct creation_set *cset, const char *format) { char *f; f = strdup(format); if (f == NULL) lafe_errc(1, 0, "No memory"); free(cset->create_format); cset->create_format = f; } const char * cset_get_format(struct creation_set *cset) { return (cset->create_format); } static void _cleanup_filters(struct filter_set *filters, int count) { int i; for (i = 0; i < count; i++) free(filters[i].filter_name); free(filters); } /* * Clean up a creation set. */ void cset_free(struct creation_set *cset) { _cleanup_filters(cset->filters, cset->filter_count); free(cset->create_format); free(cset); } struct creation_set * cset_new(void) { return calloc(1, sizeof(struct creation_set)); } /* * Build a creation set by a file name suffix. */ int cset_auto_compress(struct creation_set *cset, const char *filename) { struct filter_set *old_filters; char *name, *p; const char *code; int old_filter_count; name = strdup(filename); if (name == NULL) lafe_errc(1, 0, "No memory"); /* Save previous filters. */ old_filters = cset->filters; old_filter_count = cset->filter_count; cset->filters = NULL; cset->filter_count = 0; for (;;) { /* Get the suffix. */ p = strrchr(name, '.'); if (p == NULL) break; /* Suppose it indicates compression/filter type * such as ".gz". */ code = get_filter_code(p); if (code != NULL) { cset_add_filter(cset, code); *p = '\0'; continue; } /* Suppose it indicates format type such as ".tar". */ code = get_format_code(p); if (code != NULL) { cset_set_format(cset, code); break; } /* Suppose it indicates alias such as ".tgz". */ code = decompose_alias(p); if (code == NULL) break; /* Replace the suffix. */ *p = '\0'; name = realloc(name, strlen(name) + strlen(code) + 1); if (name == NULL) lafe_errc(1, 0, "No memory"); strcat(name, code); } free(name); if (cset->filters) { struct filter_set *v; int i, r; /* Release previous filters. */ _cleanup_filters(old_filters, old_filter_count); v = malloc(sizeof(*v) * cset->filter_count); if (v == NULL) lafe_errc(1, 0, "No memory"); /* Reverse filter sequence. */ for (i = 0, r = cset->filter_count; r > 0; ) v[i++] = cset->filters[--r]; free(cset->filters); cset->filters = v; return (1); } else { /* Put previous filters back. */ cset->filters = old_filters; cset->filter_count = old_filter_count; return (0); } } Index: stable/10/contrib/libarchive/tar/test/test_extract.tar.zst.uu =================================================================== --- stable/10/contrib/libarchive/tar/test/test_extract.tar.zst.uu (nonexistent) +++ stable/10/contrib/libarchive/tar/test/test_extract.tar.zst.uu (revision 324418) @@ -0,0 +1,6 @@ +begin 644 test_extract.tar.zst +M*+4O_010S0,`Y+>!0*5E/PM"$7^K^1VI3SS-AX&_W0KQWY!-Z1(`_4$%[$"]test.out 2>test.err", testprog, reffile); + if (f == 0 || canZstd()) { + assertEqualInt(0, systemf("%s -xf %s >test.out 2>test.err", + testprog, reffile)); + + assertFileExists("file1"); + assertTextFileContents("contents of file1.\n", "file1"); + assertFileExists("file2"); + assertTextFileContents("contents of file2.\n", "file2"); + assertEmptyFile("test.out"); + assertEmptyFile("test.err"); + } else { + skipping("It seems zstd is not supported on this platform"); + } +} Property changes on: stable/10/contrib/libarchive/tar/test/test_extract_tar_zstd.c ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: stable/10/contrib/libarchive/tar/test/test_option_acls.c =================================================================== --- stable/10/contrib/libarchive/tar/test/test_option_acls.c (revision 324417) +++ stable/10/contrib/libarchive/tar/test/test_option_acls.c (revision 324418) @@ -1,508 +1,508 @@ /*- * Copyright (c) 2017 Martin Matuska * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "test.h" __FBSDID("$FreeBSD$"); #if ARCHIVE_ACL_FREEBSD || ARCHIVE_ACL_DARWIN || ARCHIVE_ACL_LIBACL static const acl_perm_t acl_perms[] = { #if ARCHIVE_ACL_DARWIN ACL_READ_DATA, ACL_LIST_DIRECTORY, ACL_WRITE_DATA, ACL_ADD_FILE, ACL_EXECUTE, ACL_SEARCH, ACL_DELETE, ACL_APPEND_DATA, ACL_ADD_SUBDIRECTORY, ACL_DELETE_CHILD, ACL_READ_ATTRIBUTES, ACL_WRITE_ATTRIBUTES, ACL_READ_EXTATTRIBUTES, ACL_WRITE_EXTATTRIBUTES, ACL_READ_SECURITY, ACL_WRITE_SECURITY, ACL_CHANGE_OWNER, ACL_SYNCHRONIZE #else /* !ARCHIVE_ACL_DARWIN */ ACL_EXECUTE, ACL_WRITE, ACL_READ, #if ARCHIVE_ACL_FREEBSD_NFS4 ACL_READ_DATA, ACL_LIST_DIRECTORY, ACL_WRITE_DATA, ACL_ADD_FILE, ACL_APPEND_DATA, ACL_ADD_SUBDIRECTORY, ACL_READ_NAMED_ATTRS, ACL_WRITE_NAMED_ATTRS, ACL_DELETE_CHILD, ACL_READ_ATTRIBUTES, ACL_WRITE_ATTRIBUTES, ACL_DELETE, ACL_READ_ACL, ACL_WRITE_ACL, ACL_WRITE_OWNER, ACL_SYNCHRONIZE #endif /* ARCHIVE_ACL_FREEBSD_NFS4 */ #endif /* !ARCHIVE_ACL_DARWIN */ }; #if ARCHIVE_ACL_DARWIN || ARCHIVE_ACL_FREEBSD_NFS4 static const acl_flag_t acl_flags[] = { #if ARCHIVE_ACL_DARWIN ACL_ENTRY_INHERITED, ACL_ENTRY_FILE_INHERIT, ACL_ENTRY_DIRECTORY_INHERIT, ACL_ENTRY_LIMIT_INHERIT, ACL_ENTRY_ONLY_INHERIT #else /* ARCHIVE_ACL_FREEBSD_NFS4 */ ACL_ENTRY_FILE_INHERIT, ACL_ENTRY_DIRECTORY_INHERIT, ACL_ENTRY_NO_PROPAGATE_INHERIT, ACL_ENTRY_INHERIT_ONLY, ACL_ENTRY_SUCCESSFUL_ACCESS, ACL_ENTRY_FAILED_ACCESS, ACL_ENTRY_INHERITED #endif /* ARCHIVE_ACL_FREEBSD_NFS4 */ }; #endif /* ARCHIVE_ACL_DARWIN || ARCHIVE_ACL_FREEBSD_NFS4 */ /* * Compare two ACL entries on FreeBSD or on Mac OS X */ static int compare_acl_entry(acl_entry_t ae_a, acl_entry_t ae_b, int is_nfs4) { acl_tag_t tag_a, tag_b; acl_permset_t permset_a, permset_b; int perm_a, perm_b, perm_start, perm_end; void *qual_a, *qual_b; #if ARCHIVE_ACL_FREEBSD_NFS4 acl_entry_type_t type_a, type_b; #endif #if ARCHIVE_ACL_FREEBSD_NFS4 || ARCHIVE_ACL_DARWIN acl_flagset_t flagset_a, flagset_b; int flag_a, flag_b; #endif int i, r; /* Compare ACL tag */ r = acl_get_tag_type(ae_a, &tag_a); failure("acl_get_tag_type() error: %s", strerror(errno)); if (assertEqualInt(r, 0) == 0) return (-1); r = acl_get_tag_type(ae_b, &tag_b); failure("acl_get_tag_type() error: %s", strerror(errno)); if (assertEqualInt(r, 0) == 0) return (-1); if (tag_a != tag_b) return (0); /* Compare ACL qualifier */ #if ARCHIVE_ACL_DARWIN if (tag_a == ACL_EXTENDED_ALLOW || tag_b == ACL_EXTENDED_DENY) #else if (tag_a == ACL_USER || tag_a == ACL_GROUP) #endif { qual_a = acl_get_qualifier(ae_a); failure("acl_get_qualifier() error: %s", strerror(errno)); if (assert(qual_a != NULL) == 0) return (-1); qual_b = acl_get_qualifier(ae_b); failure("acl_get_qualifier() error: %s", strerror(errno)); if (assert(qual_b != NULL) == 0) { acl_free(qual_a); return (-1); } #if ARCHIVE_ACL_DARWIN if (memcmp(((guid_t *)qual_a)->g_guid, ((guid_t *)qual_b)->g_guid, KAUTH_GUID_SIZE) != 0) #else if ((tag_a == ACL_USER && (*(uid_t *)qual_a != *(uid_t *)qual_b)) || (tag_a == ACL_GROUP && (*(gid_t *)qual_a != *(gid_t *)qual_b))) #endif { acl_free(qual_a); acl_free(qual_b); return (0); } acl_free(qual_a); acl_free(qual_b); } #if ARCHIVE_ACL_FREEBSD_NFS4 if (is_nfs4) { /* Compare NFS4 ACL type */ r = acl_get_entry_type_np(ae_a, &type_a); failure("acl_get_entry_type_np() error: %s", strerror(errno)); if (assertEqualInt(r, 0) == 0) return (-1); r = acl_get_entry_type_np(ae_b, &type_b); failure("acl_get_entry_type_np() error: %s", strerror(errno)); if (assertEqualInt(r, 0) == 0) return (-1); if (type_a != type_b) return (0); } #endif /* Compare ACL perms */ r = acl_get_permset(ae_a, &permset_a); failure("acl_get_permset() error: %s", strerror(errno)); if (assertEqualInt(r, 0) == 0) return (-1); r = acl_get_permset(ae_b, &permset_b); failure("acl_get_permset() error: %s", strerror(errno)); if (assertEqualInt(r, 0) == 0) return (-1); perm_start = 0; perm_end = (int)(sizeof(acl_perms) / sizeof(acl_perms[0])); #if ARCHIVE_ACL_FREEBSD_NFS4 if (is_nfs4) perm_start = 3; else perm_end = 3; #endif /* Cycle through all perms and compare their value */ for (i = perm_start; i < perm_end; i++) { #if ARCHIVE_ACL_LIBACL perm_a = acl_get_perm(permset_a, acl_perms[i]); perm_b = acl_get_perm(permset_b, acl_perms[i]); #else perm_a = acl_get_perm_np(permset_a, acl_perms[i]); perm_b = acl_get_perm_np(permset_b, acl_perms[i]); #endif if (perm_a == -1 || perm_b == -1) return (-1); if (perm_a != perm_b) return (0); } #if ARCHIVE_ACL_FREEBSD_NFS4 || ARCHIVE_ACL_DARWIN if (is_nfs4) { r = acl_get_flagset_np(ae_a, &flagset_a); failure("acl_get_flagset_np() error: %s", strerror(errno)); if (assertEqualInt(r, 0) == 0) return (-1); r = acl_get_flagset_np(ae_b, &flagset_b); failure("acl_get_flagset_np() error: %s", strerror(errno)); if (assertEqualInt(r, 0) == 0) return (-1); /* Cycle through all flags and compare their status */ for (i = 0; i < (int)(sizeof(acl_flags) / sizeof(acl_flags[0])); i++) { flag_a = acl_get_flag_np(flagset_a, acl_flags[i]); flag_b = acl_get_flag_np(flagset_b, acl_flags[i]); if (flag_a == -1 || flag_b == -1) return (-1); if (flag_a != flag_b) return (0); } } #else /* ARCHIVE_ACL_FREEBSD_NFS4 || ARCHIVE_ACL_DARWIN */ (void)is_nfs4; /* UNUSED */ #endif return (1); } #endif /* ARCHIVE_ACL_FREEBSD || ARCHIVE_ACL_DARWIN || ARCHIVE_ACL_LIBACL */ #if ARCHIVE_ACL_SUPPORT /* * Clear default ACLs or inheritance flags */ static void clear_inheritance_flags(const char *path, int type) { switch (type) { case ARCHIVE_TEST_ACL_TYPE_POSIX1E: #if ARCHIVE_ACL_POSIX1E #if !ARCHIVE_ACL_SUNOS acl_delete_def_file(path); #else /* Solaris */ setTestAcl(path); #endif #endif /* ARCHIVE_ACL_POSIX1E */ break; case ARCHIVE_TEST_ACL_TYPE_NFS4: #if ARCHIVE_ACL_NFS4 setTestAcl(path); #endif break; default: (void)path; /* UNUSED */ break; } } static int compare_acls(const char *path_a, const char *path_b) { int ret = 1; int is_nfs4 = 0; #if ARCHIVE_ACL_SUNOS void *acl_a, *acl_b; int aclcnt_a, aclcnt_b; aclent_t *aclent_a, *aclent_b; ace_t *ace_a, *ace_b; int e; #elif ARCHIVE_ACL_DARWIN || ARCHIVE_ACL_FREEBSD || ARCHIVE_ACL_LIBACL acl_t acl_a, acl_b; acl_entry_t aclent_a, aclent_b; int a, b, r; #endif #if ARCHIVE_ACL_LIBRICHACL struct richacl *richacl_a, *richacl_b; richacl_a = NULL; richacl_b = NULL; #endif #if ARCHIVE_ACL_DARWIN || ARCHIVE_ACL_FREEBSD || ARCHIVE_ACL_LIBACL || \ ARCHIVE_ACL_SUNOS acl_a = NULL; acl_b = NULL; #endif #if ARCHIVE_ACL_SUNOS acl_a = sunacl_get(GETACL, &aclcnt_a, 0, path_a); if (acl_a == NULL) { #if ARCHIVE_ACL_SUNOS_NFS4 is_nfs4 = 1; acl_a = sunacl_get(ACE_GETACL, &aclcnt_a, 0, path_a); #endif failure("acl_get() error: %s", strerror(errno)); if (assert(acl_a != NULL) == 0) return (-1); #if ARCHIVE_ACL_SUNOS_NFS4 acl_b = sunacl_get(ACE_GETACL, &aclcnt_b, 0, path_b); #endif } else acl_b = sunacl_get(GETACL, &aclcnt_b, 0, path_b); if (acl_b == NULL && (errno == ENOSYS || errno == ENOTSUP)) { free(acl_a); return (0); } failure("acl_get() error: %s", strerror(errno)); if (assert(acl_b != NULL) == 0) { free(acl_a); return (-1); } if (aclcnt_a != aclcnt_b) { ret = 0; goto exit_free; } for (e = 0; e < aclcnt_a; e++) { if (!is_nfs4) { aclent_a = &((aclent_t *)acl_a)[e]; aclent_b = &((aclent_t *)acl_b)[e]; if (aclent_a->a_type != aclent_b->a_type || aclent_a->a_id != aclent_b->a_id || aclent_a->a_perm != aclent_b->a_perm) { ret = 0; goto exit_free; } } #if ARCHIVE_ACL_SUNOS_NFS4 else { ace_a = &((ace_t *)acl_a)[e]; ace_b = &((ace_t *)acl_b)[e]; if (ace_a->a_who != ace_b->a_who || ace_a->a_access_mask != ace_b->a_access_mask || ace_a->a_flags != ace_b->a_flags || ace_a->a_type != ace_b->a_type) { ret = 0; goto exit_free; } } #endif } #else /* !ARCHIVE_ACL_SUNOS */ #if ARCHIVE_ACL_LIBRICHACL richacl_a = richacl_get_file(path_a); #if !ARCHIVE_ACL_LIBACL if (richacl_a == NULL && (errno == ENODATA || errno == ENOTSUP || errno == ENOSYS)) return (0); failure("richacl_get_file() error: %s (%s)", path_a, strerror(errno)); if (assert(richacl_a != NULL) == 0) return (-1); #endif if (richacl_a != NULL) { richacl_b = richacl_get_file(path_b); if (richacl_b == NULL && (errno == ENODATA || errno == ENOTSUP || errno == ENOSYS)) return (0); failure("richacl_get_file() error: %s (%s)", path_b, strerror(errno)); if (assert(richacl_b != NULL) == 0) { richacl_free(richacl_a); return (-1); } if (richacl_compare(richacl_a, richacl_b) == 0) ret = 0; richacl_free(richacl_a); richacl_free(richacl_b); return (ret); } #endif /* ARCHIVE_ACL_LIBRICHACL */ #if ARCHIVE_ACL_DARWIN || ARCHIVE_ACL_FREEBSD || ARCHIVE_ACL_LIBACL #if ARCHIVE_ACL_DARWIN is_nfs4 = 1; acl_a = acl_get_file(path_a, ACL_TYPE_EXTENDED); #elif ARCHIVE_ACL_FREEBSD_NFS4 acl_a = acl_get_file(path_a, ACL_TYPE_NFS4); if (acl_a != NULL) is_nfs4 = 1; #endif if (acl_a == NULL) acl_a = acl_get_file(path_a, ACL_TYPE_ACCESS); failure("acl_get_file() error: %s (%s)", path_a, strerror(errno)); if (assert(acl_a != NULL) == 0) return (-1); #if ARCHIVE_ACL_DARWIN acl_b = acl_get_file(path_b, ACL_TYPE_EXTENDED); #elif ARCHIVE_ACL_FREEBSD_NFS4 acl_b = acl_get_file(path_b, ACL_TYPE_NFS4); #endif #if !ARCHIVE_ACL_DARWIN if (acl_b == NULL) { #if ARCHIVE_ACL_FREEBSD_NFS4 if (is_nfs4) { acl_free(acl_a); return (0); } #endif acl_b = acl_get_file(path_b, ACL_TYPE_ACCESS); } failure("acl_get_file() error: %s (%s)", path_b, strerror(errno)); if (assert(acl_b != NULL) == 0) { acl_free(acl_a); return (-1); } #endif a = acl_get_entry(acl_a, ACL_FIRST_ENTRY, &aclent_a); if (a == -1) { ret = 0; goto exit_free; } b = acl_get_entry(acl_b, ACL_FIRST_ENTRY, &aclent_b); if (b == -1) { ret = 0; goto exit_free; } #if ARCHIVE_ACL_DARWIN while (a == 0 && b == 0) #else /* FreeBSD, Linux */ while (a == 1 && b == 1) #endif { r = compare_acl_entry(aclent_a, aclent_b, is_nfs4); if (r != 1) { ret = r; goto exit_free; } a = acl_get_entry(acl_a, ACL_NEXT_ENTRY, &aclent_a); b = acl_get_entry(acl_b, ACL_NEXT_ENTRY, &aclent_b); } /* Entry count must match */ if (a != b) ret = 0; #endif /* ARCHIVE_ACL_DARWIN || ARCHIVE_ACL_FREEBSD || ARCHIVE_ACL_LIBACL */ #endif /* !ARCHIVE_ACL_SUNOS */ exit_free: #if ARCHIVE_ACL_SUNOS free(acl_a); free(acl_b); #else acl_free(acl_a); acl_free(acl_b); #endif return (ret); } #endif /* ARCHIVE_ACL_SUPPORT */ DEFINE_TEST(test_option_acls) { #if !ARCHIVE_ACL_SUPPORT skipping("ACLs are not supported on this platform"); #else /* ARCHIVE_ACL_SUPPORT */ int acltype, r; assertMakeFile("f", 0644, "a"); acltype = setTestAcl("f"); if (acltype == 0) { skipping("Can't write ACLs on the filesystem"); return; } /* Archive it with acls */ r = systemf("%s -c --no-mac-metadata --acls -f acls.tar f >acls.out 2>acls.err", testprog); assertEqualInt(r, 0); /* Archive it without acls */ r = systemf("%s -c --no-mac-metadata --no-acls -f noacls.tar f >noacls.out 2>noacls.err", testprog); assertEqualInt(r, 0); /* Extract acls with acls */ assertMakeDir("acls_acls", 0755); clear_inheritance_flags("acls_acls", acltype); r = systemf("%s -x -C acls_acls --no-same-permissions --acls -f acls.tar >acls_acls.out 2>acls_acls.err", testprog); assertEqualInt(r, 0); r = compare_acls("f", "acls_acls/f"); assertEqualInt(r, 1); - /* Extractl acls without acls */ + /* Extract acls without acls */ assertMakeDir("acls_noacls", 0755); clear_inheritance_flags("acls_noacls", acltype); r = systemf("%s -x -C acls_noacls -p --no-acls -f acls.tar >acls_noacls.out 2>acls_noacls.err", testprog); assertEqualInt(r, 0); r = compare_acls("f", "acls_noacls/f"); assertEqualInt(r, 0); /* Extract noacls with acls flag */ assertMakeDir("noacls_acls", 0755); clear_inheritance_flags("noacls_acls", acltype); r = systemf("%s -x -C noacls_acls --no-same-permissions --acls -f noacls.tar >noacls_acls.out 2>noacls_acls.err", testprog); assertEqualInt(r, 0); r = compare_acls("f", "noacls_acls/f"); assertEqualInt(r, 0); /* Extract noacls with noacls */ assertMakeDir("noacls_noacls", 0755); clear_inheritance_flags("noacls_noacls", acltype); r = systemf("%s -x -C noacls_noacls -p --no-acls -f noacls.tar >noacls_noacls.out 2>noacls_noacls.err", testprog); assertEqualInt(r, 0); r = compare_acls("f", "noacls_noacls/f"); assertEqualInt(r, 0); #endif /* ARCHIVE_ACL_SUPPORT */ } Index: stable/10/contrib/libarchive/tar/test/test_option_zstd.c =================================================================== --- stable/10/contrib/libarchive/tar/test/test_option_zstd.c (nonexistent) +++ stable/10/contrib/libarchive/tar/test/test_option_zstd.c (revision 324418) @@ -0,0 +1,85 @@ +/*- + * Copyright (c) 2017 Sean Purcell + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES + * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. + * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, + * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT + * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF + * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ +#include "test.h" +__FBSDID("$FreeBSD$"); + +DEFINE_TEST(test_option_zstd) +{ + char *p; + int r; + size_t s; + + /* Create a file. */ + assertMakeFile("f", 0644, "a"); + + /* Archive it with lz4 compression. */ + r = systemf("%s -cf - --zstd f >archive.out 2>archive.err", + testprog); + p = slurpfile(&s, "archive.err"); + p[s] = '\0'; + if (r != 0) { + if (strstr(p, "Unsupported compression") != NULL) { + skipping("This version of bsdtar was compiled " + "without zstd support"); + goto done; + } + /* POSIX permits different handling of the spawnp + * system call used to launch the subsidiary + * program: */ + /* Some systems fail immediately to spawn the new process. */ + if (strstr(p, "Can't launch") != NULL && !canZstd()) { + skipping("This version of bsdtar uses an external zstd program " + "but no such program is available on this system."); + goto done; + } + /* Some systems successfully spawn the new process, + * but fail to exec a program within that process. + * This results in failure at the first attempt to + * write. */ + if (strstr(p, "Can't write") != NULL && !canZstd()) { + skipping("This version of bsdtar uses an external zstd program " + "but no such program is available on this system."); + goto done; + } + /* On some systems the error won't be detected until closing + time, by a 127 exit error returned by waitpid. */ + if (strstr(p, "Error closing") != NULL && !canZstd()) { + skipping("This version of bsdcpio uses an external zstd program " + "but no such program is available on this system."); + return; + } + failure("--zstd option is broken: %s", p); + assertEqualInt(r, 0); + goto done; + } + free(p); + /* Check that the archive file has an lz4 signature. */ + p = slurpfile(&s, "archive.out"); + assert(s > 2); + assertEqualMem(p, "\x28\xb5\x2f\xfd", 4); + +done: + free(p); +} Property changes on: stable/10/contrib/libarchive/tar/test/test_option_zstd.c ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: stable/10/contrib/libarchive/test_utils/test_common.h =================================================================== --- stable/10/contrib/libarchive/test_utils/test_common.h (revision 324417) +++ stable/10/contrib/libarchive/test_utils/test_common.h (revision 324418) @@ -1,445 +1,448 @@ /* * Copyright (c) 2003-2017 Tim Kientzle * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. * * $FreeBSD$ */ #ifndef TEST_COMMON_H #define TEST_COMMON_H /* * The goal of this file (and the matching test.c) is to * simplify the very repetitive test-*.c test programs. */ #if defined(HAVE_CONFIG_H) /* Most POSIX platforms use the 'configure' script to build config.h */ #include "config.h" #elif defined(__FreeBSD__) /* Building as part of FreeBSD system requires a pre-built config.h. */ #include "config_freebsd.h" #elif defined(_WIN32) && !defined(__CYGWIN__) /* Win32 can't run the 'configure' script. */ #include "config_windows.h" #else /* Warn if the library hasn't been (automatically or manually) configured. */ #error Oops: No config.h and no pre-built configuration in test.h. #endif #include /* Windows requires this before sys/stat.h */ #include #if HAVE_DIRENT_H #include #endif #ifdef HAVE_DIRECT_H #include #define dirent direct #endif #include #include #ifdef HAVE_IO_H #include #endif #ifdef HAVE_STDINT_H #include #endif #include #include #include #include #include #ifdef HAVE_UNISTD_H #include #endif #include #ifdef HAVE_ACL_LIBACL_H #include #endif #ifdef HAVE_SYS_ACL_H #include #endif #ifdef HAVE_SYS_RICHACL_H #include #endif #ifdef HAVE_WINDOWS_H #include #endif /* * System-specific tweaks. We really want to minimize these * as much as possible, since they make it harder to understand * the mainline code. */ /* Windows (including Visual Studio and MinGW but not Cygwin) */ #if defined(_WIN32) && !defined(__CYGWIN__) #if !defined(__BORLANDC__) #undef chdir #define chdir _chdir #define strdup _strdup #endif #endif /* Visual Studio */ #if defined(_MSC_VER) && _MSC_VER < 1900 #define snprintf sprintf_s #endif #if defined(__BORLANDC__) #pragma warn -8068 /* Constant out of range in comparison. */ #endif /* Haiku OS and QNX */ #if defined(__HAIKU__) || defined(__QNXNTO__) /* Haiku and QNX have typedefs in stdint.h (needed for int64_t) */ #include #endif /* Get a real definition for __FBSDID if we can */ #if HAVE_SYS_CDEFS_H #include #endif /* If not, define it so as to avoid dangling semicolons. */ #ifndef __FBSDID #define __FBSDID(a) struct _undefined_hack #endif #ifndef O_BINARY #define O_BINARY 0 #endif #include "archive_platform_acl.h" #define ARCHIVE_TEST_ACL_TYPE_POSIX1E 1 #define ARCHIVE_TEST_ACL_TYPE_NFS4 2 #include "archive_platform_xattr.h" /* * Redefine DEFINE_TEST for use in defining the test functions. */ #undef DEFINE_TEST #define DEFINE_TEST(name) void name(void); void name(void) /* An implementation of the standard assert() macro */ #define assert(e) assertion_assert(__FILE__, __LINE__, (e), #e, NULL) /* chdir() and error if it fails */ #define assertChdir(path) \ assertion_chdir(__FILE__, __LINE__, path) /* Assert two files have the same file flags */ #define assertEqualFflags(patha, pathb) \ assertion_compare_fflags(__FILE__, __LINE__, patha, pathb, 0) /* Assert two integers are the same. Reports value of each one if not. */ #define assertEqualInt(v1,v2) \ assertion_equal_int(__FILE__, __LINE__, (v1), #v1, (v2), #v2, NULL) /* Assert two strings are the same. Reports value of each one if not. */ #define assertEqualString(v1,v2) \ assertion_equal_string(__FILE__, __LINE__, (v1), #v1, (v2), #v2, NULL, 0) #define assertEqualUTF8String(v1,v2) \ assertion_equal_string(__FILE__, __LINE__, (v1), #v1, (v2), #v2, NULL, 1) /* As above, but v1 and v2 are wchar_t * */ #define assertEqualWString(v1,v2) \ assertion_equal_wstring(__FILE__, __LINE__, (v1), #v1, (v2), #v2, NULL) /* As above, but raw blocks of bytes. */ #define assertEqualMem(v1, v2, l) \ assertion_equal_mem(__FILE__, __LINE__, (v1), #v1, (v2), #v2, (l), #l, NULL) /* Assert that memory is full of a specified byte */ #define assertMemoryFilledWith(v1, l, b) \ assertion_memory_filled_with(__FILE__, __LINE__, (v1), #v1, (l), #l, (b), #b, NULL) /* Assert two files are the same. */ #define assertEqualFile(f1, f2) \ assertion_equal_file(__FILE__, __LINE__, (f1), (f2)) /* Assert that a file is empty. */ #define assertEmptyFile(pathname) \ assertion_empty_file(__FILE__, __LINE__, (pathname)) /* Assert that a file is not empty. */ #define assertNonEmptyFile(pathname) \ assertion_non_empty_file(__FILE__, __LINE__, (pathname)) #define assertFileAtime(pathname, sec, nsec) \ assertion_file_atime(__FILE__, __LINE__, pathname, sec, nsec) #define assertFileAtimeRecent(pathname) \ assertion_file_atime_recent(__FILE__, __LINE__, pathname) #define assertFileBirthtime(pathname, sec, nsec) \ assertion_file_birthtime(__FILE__, __LINE__, pathname, sec, nsec) #define assertFileBirthtimeRecent(pathname) \ assertion_file_birthtime_recent(__FILE__, __LINE__, pathname) /* Assert that a file exists; supports printf-style arguments. */ #define assertFileExists(pathname) \ assertion_file_exists(__FILE__, __LINE__, pathname) /* Assert that a file exists. */ #define assertFileNotExists(pathname) \ assertion_file_not_exists(__FILE__, __LINE__, pathname) /* Assert that file contents match a string. */ #define assertFileContents(data, data_size, pathname) \ assertion_file_contents(__FILE__, __LINE__, data, data_size, pathname) /* Verify that a file does not contain invalid strings */ #define assertFileContainsNoInvalidStrings(pathname, strings) \ assertion_file_contains_no_invalid_strings(__FILE__, __LINE__, pathname, strings) #define assertFileMtime(pathname, sec, nsec) \ assertion_file_mtime(__FILE__, __LINE__, pathname, sec, nsec) #define assertFileMtimeRecent(pathname) \ assertion_file_mtime_recent(__FILE__, __LINE__, pathname) #define assertFileNLinks(pathname, nlinks) \ assertion_file_nlinks(__FILE__, __LINE__, pathname, nlinks) #define assertFileSize(pathname, size) \ assertion_file_size(__FILE__, __LINE__, pathname, size) #define assertFileMode(pathname, mode) \ assertion_file_mode(__FILE__, __LINE__, pathname, mode) #define assertTextFileContents(text, pathname) \ assertion_text_file_contents(__FILE__, __LINE__, text, pathname) #define assertFileContainsLinesAnyOrder(pathname, lines) \ assertion_file_contains_lines_any_order(__FILE__, __LINE__, pathname, lines) #define assertIsDir(pathname, mode) \ assertion_is_dir(__FILE__, __LINE__, pathname, mode) #define assertIsHardlink(path1, path2) \ assertion_is_hardlink(__FILE__, __LINE__, path1, path2) #define assertIsNotHardlink(path1, path2) \ assertion_is_not_hardlink(__FILE__, __LINE__, path1, path2) #define assertIsReg(pathname, mode) \ assertion_is_reg(__FILE__, __LINE__, pathname, mode) #define assertIsSymlink(pathname, contents) \ assertion_is_symlink(__FILE__, __LINE__, pathname, contents) /* Create a directory, report error if it fails. */ #define assertMakeDir(dirname, mode) \ assertion_make_dir(__FILE__, __LINE__, dirname, mode) #define assertMakeFile(path, mode, contents) \ assertion_make_file(__FILE__, __LINE__, path, mode, -1, contents) #define assertMakeBinFile(path, mode, csize, contents) \ assertion_make_file(__FILE__, __LINE__, path, mode, csize, contents) #define assertMakeHardlink(newfile, oldfile) \ assertion_make_hardlink(__FILE__, __LINE__, newfile, oldfile) #define assertMakeSymlink(newfile, linkto) \ assertion_make_symlink(__FILE__, __LINE__, newfile, linkto) #define assertSetNodump(path) \ assertion_set_nodump(__FILE__, __LINE__, path) #define assertUmask(mask) \ assertion_umask(__FILE__, __LINE__, mask) /* Assert that two files have unequal file flags */ #define assertUnequalFflags(patha, pathb) \ assertion_compare_fflags(__FILE__, __LINE__, patha, pathb, 1) #define assertUtimes(pathname, atime, atime_nsec, mtime, mtime_nsec) \ assertion_utimes(__FILE__, __LINE__, pathname, atime, atime_nsec, mtime, mtime_nsec) #ifndef PROGRAM #define assertEntrySetAcls(entry, acls, count) \ assertion_entry_set_acls(__FILE__, __LINE__, entry, acls, count) #define assertEntryCompareAcls(entry, acls, count, type, mode) \ assertion_entry_compare_acls(__FILE__, __LINE__, entry, acls, count, type, mode) #endif /* * This would be simple with C99 variadic macros, but I don't want to * require that. Instead, I insert a function call before each * skipping() call to pass the file and line information down. Crude, * but effective. */ #define skipping \ skipping_setup(__FILE__, __LINE__);test_skipping /* Function declarations. These are defined in test_utility.c. */ void failure(const char *fmt, ...); int assertion_assert(const char *, int, int, const char *, void *); int assertion_chdir(const char *, int, const char *); int assertion_compare_fflags(const char *, int, const char *, const char *, int); int assertion_empty_file(const char *, int, const char *); int assertion_equal_file(const char *, int, const char *, const char *); int assertion_equal_int(const char *, int, long long, const char *, long long, const char *, void *); int assertion_equal_mem(const char *, int, const void *, const char *, const void *, const char *, size_t, const char *, void *); int assertion_memory_filled_with(const char *, int, const void *, const char *, size_t, const char *, char, const char *, void *); int assertion_equal_string(const char *, int, const char *v1, const char *, const char *v2, const char *, void *, int); int assertion_equal_wstring(const char *, int, const wchar_t *v1, const char *, const wchar_t *v2, const char *, void *); int assertion_file_atime(const char *, int, const char *, long, long); int assertion_file_atime_recent(const char *, int, const char *); int assertion_file_birthtime(const char *, int, const char *, long, long); int assertion_file_birthtime_recent(const char *, int, const char *); int assertion_file_contains_lines_any_order(const char *, int, const char *, const char **); int assertion_file_contains_no_invalid_strings(const char *, int, const char *, const char **); int assertion_file_contents(const char *, int, const void *, int, const char *); int assertion_file_exists(const char *, int, const char *); int assertion_file_mode(const char *, int, const char *, int); int assertion_file_mtime(const char *, int, const char *, long, long); int assertion_file_mtime_recent(const char *, int, const char *); int assertion_file_nlinks(const char *, int, const char *, int); int assertion_file_not_exists(const char *, int, const char *); int assertion_file_size(const char *, int, const char *, long); int assertion_is_dir(const char *, int, const char *, int); int assertion_is_hardlink(const char *, int, const char *, const char *); int assertion_is_not_hardlink(const char *, int, const char *, const char *); int assertion_is_reg(const char *, int, const char *, int); int assertion_is_symlink(const char *, int, const char *, const char *); int assertion_make_dir(const char *, int, const char *, int); int assertion_make_file(const char *, int, const char *, int, int, const void *); int assertion_make_hardlink(const char *, int, const char *newpath, const char *); int assertion_make_symlink(const char *, int, const char *newpath, const char *); int assertion_non_empty_file(const char *, int, const char *); int assertion_set_nodump(const char *, int, const char *); int assertion_text_file_contents(const char *, int, const char *buff, const char *f); int assertion_umask(const char *, int, int); int assertion_utimes(const char *, int, const char *, long, long, long, long ); int assertion_version(const char*, int, const char *, const char *); void skipping_setup(const char *, int); void test_skipping(const char *fmt, ...); /* Like sprintf, then system() */ int systemf(const char * fmt, ...); /* Delay until time() returns a value after this. */ void sleepUntilAfter(time_t); /* Return true if this platform can create symlinks. */ int canSymlink(void); /* Return true if this platform can run the "bzip2" program. */ int canBzip2(void); /* Return true if this platform can run the "grzip" program. */ int canGrzip(void); /* Return true if this platform can run the "gzip" program. */ int canGzip(void); /* Return true if this platform can run the specified command. */ int canRunCommand(const char *); /* Return true if this platform can run the "lrzip" program. */ int canLrzip(void); /* Return true if this platform can run the "lz4" program. */ int canLz4(void); +/* Return true if this platform can run the "zstd" program. */ +int canZstd(void); + /* Return true if this platform can run the "lzip" program. */ int canLzip(void); /* Return true if this platform can run the "lzma" program. */ int canLzma(void); /* Return true if this platform can run the "lzop" program. */ int canLzop(void); /* Return true if this platform can run the "xz" program. */ int canXz(void); /* Return true if this filesystem can handle nodump flags. */ int canNodump(void); /* Set test ACLs */ int setTestAcl(const char *path); /* Get extended attribute */ const void *getXattr(const char *, const char *, size_t *); /* Set extended attribute */ int setXattr(const char *, const char *, const void *, size_t); /* Return true if the file has large i-node number(>0xffffffff). */ int is_LargeInode(const char *); #if ARCHIVE_ACL_SUNOS /* Fetch ACLs on Solaris using acl() or facl() */ void *sunacl_get(int cmd, int *aclcnt, int fd, const char *path); #endif /* Suck file into string allocated via malloc(). Call free() when done. */ /* Supports printf-style args: slurpfile(NULL, "%s/myfile", refdir); */ char *slurpfile(size_t *, const char *fmt, ...); /* Dump block of bytes to a file. */ void dumpfile(const char *filename, void *, size_t); /* Extracts named reference file to the current directory. */ void extract_reference_file(const char *); /* Copies named reference file to the current directory. */ void copy_reference_file(const char *); /* Extracts a list of files to the current directory. * List must be NULL terminated. */ void extract_reference_files(const char **); /* Subtract umask from mode */ mode_t umasked(mode_t expected_mode); /* Path to working directory for current test */ extern const char *testworkdir; #ifndef PROGRAM /* * Special interfaces for libarchive test harness. */ #include "archive.h" #include "archive_entry.h" /* ACL structure */ struct archive_test_acl_t { int type; /* Type of ACL */ int permset; /* Permissions for this class of users. */ int tag; /* Owner, User, Owning group, group, other, etc. */ int qual; /* GID or UID of user/group, depending on tag. */ const char *name; /* Name of user/group, depending on tag. */ }; /* Set ACLs */ int assertion_entry_set_acls(const char *, int, struct archive_entry *, struct archive_test_acl_t *, int); /* Compare ACLs */ int assertion_entry_compare_acls(const char *, int, struct archive_entry *, struct archive_test_acl_t *, int, int, int); /* Special customized read-from-memory interface. */ int read_open_memory(struct archive *, const void *, size_t, size_t); /* _minimal version exercises a slightly different set of libarchive APIs. */ int read_open_memory_minimal(struct archive *, const void *, size_t, size_t); /* _seek version produces a seekable file. */ int read_open_memory_seek(struct archive *, const void *, size_t, size_t); /* Versions of above that accept an archive argument for additional info. */ #define assertA(e) assertion_assert(__FILE__, __LINE__, (e), #e, (a)) #define assertEqualIntA(a,v1,v2) \ assertion_equal_int(__FILE__, __LINE__, (v1), #v1, (v2), #v2, (a)) #define assertEqualStringA(a,v1,v2) \ assertion_equal_string(__FILE__, __LINE__, (v1), #v1, (v2), #v2, (a), 0) #else /* defined(PROGRAM) */ /* * Special interfaces for program test harness. */ /* Pathname of exe to be tested. */ extern const char *testprogfile; /* Name of exe to use in printf-formatted command strings. */ /* On Windows, this includes leading/trailing quotes. */ extern const char *testprog; void assertVersion(const char *prog, const char *base); #endif /* defined(PROGRAM) */ #ifdef USE_DMALLOC #include #endif #endif /* TEST_COMMON_H */ Index: stable/10/contrib/libarchive/test_utils/test_main.c =================================================================== --- stable/10/contrib/libarchive/test_utils/test_main.c (revision 324417) +++ stable/10/contrib/libarchive/test_utils/test_main.c (revision 324418) @@ -1,3834 +1,3849 @@ /* * Copyright (c) 2003-2009 Tim Kientzle * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include "test.h" #include "test_utils.h" #ifdef HAVE_SYS_IOCTL_H #include #endif #ifdef HAVE_SYS_TIME_H #include #endif #include #ifdef HAVE_ICONV_H #include #endif /* * Some Linux distributions have both linux/ext2_fs.h and ext2fs/ext2_fs.h. * As the include guards don't agree, the order of include is important. */ #ifdef HAVE_LINUX_EXT2_FS_H #include /* for Linux file flags */ #endif #if defined(HAVE_EXT2FS_EXT2_FS_H) && !defined(__CYGWIN__) #include /* Linux file flags, broken on Cygwin */ #endif #ifdef HAVE_LINUX_FS_H #include #endif #include #include #ifdef HAVE_SIGNAL_H #include #endif #include #include #ifdef HAVE_SIGNAL_H #endif #ifdef HAVE_ACL_LIBACL_H #include #endif #ifdef HAVE_SYS_TYPES_H #include #endif #ifdef HAVE_SYS_ACL_H #include #endif #ifdef HAVE_SYS_EA_H #include #endif #ifdef HAVE_SYS_EXTATTR_H #include #endif #if HAVE_SYS_XATTR_H #include #elif HAVE_ATTR_XATTR_H #include #endif #ifdef HAVE_SYS_RICHACL_H #include #endif #if HAVE_MEMBERSHIP_H #include #endif /* * * Windows support routines * * Note: Configuration is a tricky issue. Using HAVE_* feature macros * in the test harness is dangerous because they cover up * configuration errors. The classic example of this is omitting a * configure check. If libarchive and libarchive_test both look for * the same feature macro, such errors are hard to detect. Platform * macros (e.g., _WIN32 or __GNUC__) are a little better, but can * easily lead to very messy code. It's best to limit yourself * to only the most generic programming techniques in the test harness * and thus avoid conditionals altogether. Where that's not possible, * try to minimize conditionals by grouping platform-specific tests in * one place (e.g., test_acl_freebsd) or by adding new assert() * functions (e.g., assertMakeHardlink()) to cover up platform * differences. Platform-specific coding in libarchive_test is often * a symptom that some capability is missing from libarchive itself. */ #if defined(_WIN32) && !defined(__CYGWIN__) #include #include #include #ifndef F_OK #define F_OK (0) #endif #ifndef S_ISDIR #define S_ISDIR(m) ((m) & _S_IFDIR) #endif #ifndef S_ISREG #define S_ISREG(m) ((m) & _S_IFREG) #endif #if !defined(__BORLANDC__) #define access _access #undef chdir #define chdir _chdir #endif #ifndef fileno #define fileno _fileno #endif /*#define fstat _fstat64*/ #if !defined(__BORLANDC__) #define getcwd _getcwd #endif #define lstat stat /*#define lstat _stat64*/ /*#define stat _stat64*/ #define rmdir _rmdir #if !defined(__BORLANDC__) #define strdup _strdup #define umask _umask #endif #define int64_t __int64 #endif #if defined(HAVE__CrtSetReportMode) # include #endif mode_t umasked(mode_t expected_mode) { mode_t mode = umask(0); umask(mode); return expected_mode & ~mode; } /* Path to working directory for current test */ const char *testworkdir; #ifdef PROGRAM /* Pathname of exe to be tested. */ const char *testprogfile; /* Name of exe to use in printf-formatted command strings. */ /* On Windows, this includes leading/trailing quotes. */ const char *testprog; #endif #if defined(_WIN32) && !defined(__CYGWIN__) static void *GetFunctionKernel32(const char *); static int my_CreateSymbolicLinkA(const char *, const char *, int); static int my_CreateHardLinkA(const char *, const char *); static int my_GetFileInformationByName(const char *, BY_HANDLE_FILE_INFORMATION *); static void * GetFunctionKernel32(const char *name) { static HINSTANCE lib; static int set; if (!set) { set = 1; lib = LoadLibrary("kernel32.dll"); } if (lib == NULL) { fprintf(stderr, "Can't load kernel32.dll?!\n"); exit(1); } return (void *)GetProcAddress(lib, name); } static int my_CreateSymbolicLinkA(const char *linkname, const char *target, int flags) { static BOOLEAN (WINAPI *f)(LPCSTR, LPCSTR, DWORD); static int set; if (!set) { set = 1; f = GetFunctionKernel32("CreateSymbolicLinkA"); } return f == NULL ? 0 : (*f)(linkname, target, flags); } static int my_CreateHardLinkA(const char *linkname, const char *target) { static BOOLEAN (WINAPI *f)(LPCSTR, LPCSTR, LPSECURITY_ATTRIBUTES); static int set; if (!set) { set = 1; f = GetFunctionKernel32("CreateHardLinkA"); } return f == NULL ? 0 : (*f)(linkname, target, NULL); } static int my_GetFileInformationByName(const char *path, BY_HANDLE_FILE_INFORMATION *bhfi) { HANDLE h; int r; memset(bhfi, 0, sizeof(*bhfi)); h = CreateFile(path, FILE_READ_ATTRIBUTES, 0, NULL, OPEN_EXISTING, FILE_FLAG_BACKUP_SEMANTICS, NULL); if (h == INVALID_HANDLE_VALUE) return (0); r = GetFileInformationByHandle(h, bhfi); CloseHandle(h); return (r); } #endif #if defined(HAVE__CrtSetReportMode) && !defined(__WATCOMC__) static void invalid_parameter_handler(const wchar_t * expression, const wchar_t * function, const wchar_t * file, unsigned int line, uintptr_t pReserved) { /* nop */ // Silence unused-parameter compiler warnings. (void)expression; (void)function; (void)file; (void)line; (void)pReserved; } #endif /* * * OPTIONS FLAGS * */ /* Enable core dump on failure. */ static int dump_on_failure = 0; /* Default is to remove temp dirs and log data for successful tests. */ static int keep_temp_files = 0; /* Default is to run the specified tests once and report errors. */ static int until_failure = 0; /* Default is to just report pass/fail for each test. */ static int verbosity = 0; #define VERBOSITY_SUMMARY_ONLY -1 /* -q */ #define VERBOSITY_PASSFAIL 0 /* Default */ #define VERBOSITY_LIGHT_REPORT 1 /* -v */ #define VERBOSITY_FULL 2 /* -vv */ /* A few places generate even more output for verbosity > VERBOSITY_FULL, * mostly for debugging the test harness itself. */ /* Cumulative count of assertion failures. */ static int failures = 0; /* Cumulative count of reported skips. */ static int skips = 0; /* Cumulative count of assertions checked. */ static int assertions = 0; /* Directory where uuencoded reference files can be found. */ static const char *refdir; /* * Report log information selectively to console and/or disk log. */ static int log_console = 0; static FILE *logfile; static void vlogprintf(const char *fmt, va_list ap) { #ifdef va_copy va_list lfap; va_copy(lfap, ap); #endif if (log_console) vfprintf(stdout, fmt, ap); if (logfile != NULL) #ifdef va_copy vfprintf(logfile, fmt, lfap); va_end(lfap); #else vfprintf(logfile, fmt, ap); #endif } static void logprintf(const char *fmt, ...) { va_list ap; va_start(ap, fmt); vlogprintf(fmt, ap); va_end(ap); } /* Set up a message to display only if next assertion fails. */ static char msgbuff[4096]; static const char *msg, *nextmsg; void failure(const char *fmt, ...) { va_list ap; if (fmt == NULL) { nextmsg = NULL; } else { va_start(ap, fmt); vsprintf(msgbuff, fmt, ap); va_end(ap); nextmsg = msgbuff; } } /* * Copy arguments into file-local variables. * This was added to permit vararg assert() functions without needing * variadic wrapper macros. Turns out that the vararg capability is almost * never used, so almost all of the vararg assertions can be simplified * by removing the vararg capability and reworking the wrapper macro to * pass __FILE__, __LINE__ directly into the function instead of using * this hook. I suspect this machinery is used so rarely that we * would be better off just removing it entirely. That would simplify * the code here noticeably. */ static const char *skipping_filename; static int skipping_line; void skipping_setup(const char *filename, int line) { skipping_filename = filename; skipping_line = line; } /* Called at the beginning of each assert() function. */ static void assertion_count(const char *file, int line) { (void)file; /* UNUSED */ (void)line; /* UNUSED */ ++assertions; /* Proper handling of "failure()" message. */ msg = nextmsg; nextmsg = NULL; /* Uncomment to print file:line after every assertion. * Verbose, but occasionally useful in tracking down crashes. */ /* printf("Checked %s:%d\n", file, line); */ } /* * For each test source file, we remember how many times each * assertion was reported. Cleared before each new test, * used by test_summarize(). */ static struct line { int count; int skip; } failed_lines[10000]; const char *failed_filename; /* Count this failure, setup up log destination and handle initial report. */ static void failure_start(const char *filename, int line, const char *fmt, ...) { va_list ap; /* Record another failure for this line. */ ++failures; failed_filename = filename; failed_lines[line].count++; /* Determine whether to log header to console. */ switch (verbosity) { case VERBOSITY_LIGHT_REPORT: log_console = (failed_lines[line].count < 2); break; default: log_console = (verbosity >= VERBOSITY_FULL); } /* Log file:line header for this failure */ va_start(ap, fmt); #if _MSC_VER logprintf("%s(%d): ", filename, line); #else logprintf("%s:%d: ", filename, line); #endif vlogprintf(fmt, ap); va_end(ap); logprintf("\n"); if (msg != NULL && msg[0] != '\0') { logprintf(" Description: %s\n", msg); msg = NULL; } /* Determine whether to log details to console. */ if (verbosity == VERBOSITY_LIGHT_REPORT) log_console = 0; } /* Complete reporting of failed tests. */ /* * The 'extra' hook here is used by libarchive to include libarchive * error messages with assertion failures. It could also be used * to add strerror() output, for example. Just define the EXTRA_DUMP() * macro appropriately. */ static void failure_finish(void *extra) { (void)extra; /* UNUSED (maybe) */ #ifdef EXTRA_DUMP if (extra != NULL) { logprintf(" errno: %d\n", EXTRA_ERRNO(extra)); logprintf(" detail: %s\n", EXTRA_DUMP(extra)); } #endif if (dump_on_failure) { fprintf(stderr, " *** forcing core dump so failure can be debugged ***\n"); abort(); } } /* Inform user that we're skipping some checks. */ void test_skipping(const char *fmt, ...) { char buff[1024]; va_list ap; va_start(ap, fmt); vsprintf(buff, fmt, ap); va_end(ap); /* Use failure() message if set. */ msg = nextmsg; nextmsg = NULL; /* failure_start() isn't quite right, but is awfully convenient. */ failure_start(skipping_filename, skipping_line, "SKIPPING: %s", buff); --failures; /* Undo failures++ in failure_start() */ /* Don't failure_finish() here. */ /* Mark as skip, so doesn't count as failed test. */ failed_lines[skipping_line].skip = 1; ++skips; } /* * * ASSERTIONS * */ /* Generic assert() just displays the failed condition. */ int assertion_assert(const char *file, int line, int value, const char *condition, void *extra) { assertion_count(file, line); if (!value) { failure_start(file, line, "Assertion failed: %s", condition); failure_finish(extra); } return (value); } /* chdir() and report any errors */ int assertion_chdir(const char *file, int line, const char *pathname) { assertion_count(file, line); if (chdir(pathname) == 0) return (1); failure_start(file, line, "chdir(\"%s\")", pathname); failure_finish(NULL); return (0); } /* Verify two integers are equal. */ int assertion_equal_int(const char *file, int line, long long v1, const char *e1, long long v2, const char *e2, void *extra) { assertion_count(file, line); if (v1 == v2) return (1); failure_start(file, line, "%s != %s", e1, e2); logprintf(" %s=%lld (0x%llx, 0%llo)\n", e1, v1, v1, v1); logprintf(" %s=%lld (0x%llx, 0%llo)\n", e2, v2, v2, v2); failure_finish(extra); return (0); } /* * Utility to convert a single UTF-8 sequence. */ static int _utf8_to_unicode(uint32_t *pwc, const char *s, size_t n) { static const char utf8_count[256] = { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,/* 00 - 0F */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,/* 10 - 1F */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,/* 20 - 2F */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,/* 30 - 3F */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,/* 40 - 4F */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,/* 50 - 5F */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,/* 60 - 6F */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,/* 70 - 7F */ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,/* 80 - 8F */ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,/* 90 - 9F */ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,/* A0 - AF */ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,/* B0 - BF */ 0, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,/* C0 - CF */ 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,/* D0 - DF */ 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,/* E0 - EF */ 4, 4, 4, 4, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 /* F0 - FF */ }; int ch; int cnt; uint32_t wc; *pwc = 0; /* Sanity check. */ if (n == 0) return (0); /* * Decode 1-4 bytes depending on the value of the first byte. */ ch = (unsigned char)*s; if (ch == 0) return (0); /* Standard: return 0 for end-of-string. */ cnt = utf8_count[ch]; /* Invalid sequence or there are not plenty bytes. */ if (n < (size_t)cnt) return (-1); /* Make a Unicode code point from a single UTF-8 sequence. */ switch (cnt) { case 1: /* 1 byte sequence. */ *pwc = ch & 0x7f; return (cnt); case 2: /* 2 bytes sequence. */ if ((s[1] & 0xc0) != 0x80) return (-1); *pwc = ((ch & 0x1f) << 6) | (s[1] & 0x3f); return (cnt); case 3: /* 3 bytes sequence. */ if ((s[1] & 0xc0) != 0x80) return (-1); if ((s[2] & 0xc0) != 0x80) return (-1); wc = ((ch & 0x0f) << 12) | ((s[1] & 0x3f) << 6) | (s[2] & 0x3f); if (wc < 0x800) return (-1);/* Overlong sequence. */ break; case 4: /* 4 bytes sequence. */ if (n < 4) return (-1); if ((s[1] & 0xc0) != 0x80) return (-1); if ((s[2] & 0xc0) != 0x80) return (-1); if ((s[3] & 0xc0) != 0x80) return (-1); wc = ((ch & 0x07) << 18) | ((s[1] & 0x3f) << 12) | ((s[2] & 0x3f) << 6) | (s[3] & 0x3f); if (wc < 0x10000) return (-1);/* Overlong sequence. */ break; default: return (-1); } /* The code point larger than 0x10FFFF is not legal * Unicode values. */ if (wc > 0x10FFFF) return (-1); /* Correctly gets a Unicode, returns used bytes. */ *pwc = wc; return (cnt); } static void strdump(const char *e, const char *p, int ewidth, int utf8) { const char *q = p; logprintf(" %*s = ", ewidth, e); if (p == NULL) { logprintf("NULL\n"); return; } logprintf("\""); while (*p != '\0') { unsigned int c = 0xff & *p++; switch (c) { case '\a': logprintf("\\a"); break; case '\b': logprintf("\\b"); break; case '\n': logprintf("\\n"); break; case '\r': logprintf("\\r"); break; default: if (c >= 32 && c < 127) logprintf("%c", c); else logprintf("\\x%02X", c); } } logprintf("\""); logprintf(" (length %d)", q == NULL ? -1 : (int)strlen(q)); /* * If the current string is UTF-8, dump its code points. */ if (utf8) { size_t len; uint32_t uc; int n; int cnt = 0; p = q; len = strlen(p); logprintf(" ["); while ((n = _utf8_to_unicode(&uc, p, len)) > 0) { if (p != q) logprintf(" "); logprintf("%04X", uc); p += n; len -= n; cnt++; } logprintf("]"); logprintf(" (count %d", cnt); if (n < 0) { logprintf(",unknown %d bytes", len); } logprintf(")"); } logprintf("\n"); } /* Verify two strings are equal, dump them if not. */ int assertion_equal_string(const char *file, int line, const char *v1, const char *e1, const char *v2, const char *e2, void *extra, int utf8) { int l1, l2; assertion_count(file, line); if (v1 == v2 || (v1 != NULL && v2 != NULL && strcmp(v1, v2) == 0)) return (1); failure_start(file, line, "%s != %s", e1, e2); l1 = (int)strlen(e1); l2 = (int)strlen(e2); if (l1 < l2) l1 = l2; strdump(e1, v1, l1, utf8); strdump(e2, v2, l1, utf8); failure_finish(extra); return (0); } static void wcsdump(const char *e, const wchar_t *w) { logprintf(" %s = ", e); if (w == NULL) { logprintf("(null)"); return; } logprintf("\""); while (*w != L'\0') { unsigned int c = *w++; if (c >= 32 && c < 127) logprintf("%c", c); else if (c < 256) logprintf("\\x%02X", c); else if (c < 0x10000) logprintf("\\u%04X", c); else logprintf("\\U%08X", c); } logprintf("\"\n"); } #ifndef HAVE_WCSCMP static int wcscmp(const wchar_t *s1, const wchar_t *s2) { while (*s1 == *s2++) { if (*s1++ == L'\0') return 0; } if (*s1 > *--s2) return 1; else return -1; } #endif /* Verify that two wide strings are equal, dump them if not. */ int assertion_equal_wstring(const char *file, int line, const wchar_t *v1, const char *e1, const wchar_t *v2, const char *e2, void *extra) { assertion_count(file, line); if (v1 == v2) return (1); if (v1 != NULL && v2 != NULL && wcscmp(v1, v2) == 0) return (1); failure_start(file, line, "%s != %s", e1, e2); wcsdump(e1, v1); wcsdump(e2, v2); failure_finish(extra); return (0); } /* * Pretty standard hexdump routine. As a bonus, if ref != NULL, then * any bytes in p that differ from ref will be highlighted with '_' * before and after the hex value. */ static void hexdump(const char *p, const char *ref, size_t l, size_t offset) { size_t i, j; char sep; if (p == NULL) { logprintf("(null)\n"); return; } for(i=0; i < l; i+=16) { logprintf("%04x", (unsigned)(i + offset)); sep = ' '; for (j = 0; j < 16 && i + j < l; j++) { if (ref != NULL && p[i + j] != ref[i + j]) sep = '_'; logprintf("%c%02x", sep, 0xff & (int)p[i+j]); if (ref != NULL && p[i + j] == ref[i + j]) sep = ' '; } for (; j < 16; j++) { logprintf("%c ", sep); sep = ' '; } logprintf("%c", sep); for (j=0; j < 16 && i + j < l; j++) { int c = p[i + j]; if (c >= ' ' && c <= 126) logprintf("%c", c); else logprintf("."); } logprintf("\n"); } } /* Verify that two blocks of memory are the same, display the first * block of differences if they're not. */ int assertion_equal_mem(const char *file, int line, const void *_v1, const char *e1, const void *_v2, const char *e2, size_t l, const char *ld, void *extra) { const char *v1 = (const char *)_v1; const char *v2 = (const char *)_v2; size_t offset; assertion_count(file, line); if (v1 == v2 || (v1 != NULL && v2 != NULL && memcmp(v1, v2, l) == 0)) return (1); if (v1 == NULL || v2 == NULL) return (0); failure_start(file, line, "%s != %s", e1, e2); logprintf(" size %s = %d\n", ld, (int)l); /* Dump 48 bytes (3 lines) so that the first difference is * in the second line. */ offset = 0; while (l > 64 && memcmp(v1, v2, 32) == 0) { /* Two lines agree, so step forward one line. */ v1 += 16; v2 += 16; l -= 16; offset += 16; } logprintf(" Dump of %s\n", e1); hexdump(v1, v2, l < 128 ? l : 128, offset); logprintf(" Dump of %s\n", e2); hexdump(v2, v1, l < 128 ? l : 128, offset); logprintf("\n"); failure_finish(extra); return (0); } /* Verify that a block of memory is filled with the specified byte. */ int assertion_memory_filled_with(const char *file, int line, const void *_v1, const char *vd, size_t l, const char *ld, char b, const char *bd, void *extra) { const char *v1 = (const char *)_v1; size_t c = 0; size_t i; (void)ld; /* UNUSED */ assertion_count(file, line); for (i = 0; i < l; ++i) { if (v1[i] == b) { ++c; } } if (c == l) return (1); failure_start(file, line, "%s (size %d) not filled with %s", vd, (int)l, bd); logprintf(" Only %d bytes were correct\n", (int)c); failure_finish(extra); return (0); } /* Verify that the named file exists and is empty. */ int assertion_empty_file(const char *filename, int line, const char *f1) { char buff[1024]; struct stat st; ssize_t s; FILE *f; assertion_count(filename, line); if (stat(f1, &st) != 0) { failure_start(filename, line, "Stat failed: %s", f1); failure_finish(NULL); return (0); } if (st.st_size == 0) return (1); failure_start(filename, line, "File should be empty: %s", f1); logprintf(" File size: %d\n", (int)st.st_size); logprintf(" Contents:\n"); f = fopen(f1, "rb"); if (f == NULL) { logprintf(" Unable to open %s\n", f1); } else { s = ((off_t)sizeof(buff) < st.st_size) ? (ssize_t)sizeof(buff) : (ssize_t)st.st_size; s = fread(buff, 1, s, f); hexdump(buff, NULL, s, 0); fclose(f); } failure_finish(NULL); return (0); } /* Verify that the named file exists and is not empty. */ int assertion_non_empty_file(const char *filename, int line, const char *f1) { struct stat st; assertion_count(filename, line); if (stat(f1, &st) != 0) { failure_start(filename, line, "Stat failed: %s", f1); failure_finish(NULL); return (0); } if (st.st_size == 0) { failure_start(filename, line, "File empty: %s", f1); failure_finish(NULL); return (0); } return (1); } /* Verify that two files have the same contents. */ /* TODO: hexdump the first bytes that actually differ. */ int assertion_equal_file(const char *filename, int line, const char *fn1, const char *fn2) { char buff1[1024]; char buff2[1024]; FILE *f1, *f2; int n1, n2; assertion_count(filename, line); f1 = fopen(fn1, "rb"); f2 = fopen(fn2, "rb"); if (f1 == NULL || f2 == NULL) { if (f1) fclose(f1); if (f2) fclose(f2); return (0); } for (;;) { n1 = (int)fread(buff1, 1, sizeof(buff1), f1); n2 = (int)fread(buff2, 1, sizeof(buff2), f2); if (n1 != n2) break; if (n1 == 0 && n2 == 0) { fclose(f1); fclose(f2); return (1); } if (memcmp(buff1, buff2, n1) != 0) break; } fclose(f1); fclose(f2); failure_start(filename, line, "Files not identical"); logprintf(" file1=\"%s\"\n", fn1); logprintf(" file2=\"%s\"\n", fn2); failure_finish(NULL); return (0); } /* Verify that the named file does exist. */ int assertion_file_exists(const char *filename, int line, const char *f) { assertion_count(filename, line); #if defined(_WIN32) && !defined(__CYGWIN__) if (!_access(f, 0)) return (1); #else if (!access(f, F_OK)) return (1); #endif failure_start(filename, line, "File should exist: %s", f); failure_finish(NULL); return (0); } /* Verify that the named file doesn't exist. */ int assertion_file_not_exists(const char *filename, int line, const char *f) { assertion_count(filename, line); #if defined(_WIN32) && !defined(__CYGWIN__) if (_access(f, 0)) return (1); #else if (access(f, F_OK)) return (1); #endif failure_start(filename, line, "File should not exist: %s", f); failure_finish(NULL); return (0); } /* Compare the contents of a file to a block of memory. */ int assertion_file_contents(const char *filename, int line, const void *buff, int s, const char *fn) { char *contents; FILE *f; int n; assertion_count(filename, line); f = fopen(fn, "rb"); if (f == NULL) { failure_start(filename, line, "File should exist: %s", fn); failure_finish(NULL); return (0); } contents = malloc(s * 2); n = (int)fread(contents, 1, s * 2, f); fclose(f); if (n == s && memcmp(buff, contents, s) == 0) { free(contents); return (1); } failure_start(filename, line, "File contents don't match"); logprintf(" file=\"%s\"\n", fn); if (n > 0) hexdump(contents, buff, n > 512 ? 512 : n, 0); else { logprintf(" File empty, contents should be:\n"); hexdump(buff, NULL, s > 512 ? 512 : s, 0); } failure_finish(NULL); free(contents); return (0); } /* Check the contents of a text file, being tolerant of line endings. */ int assertion_text_file_contents(const char *filename, int line, const char *buff, const char *fn) { char *contents; const char *btxt, *ftxt; FILE *f; int n, s; assertion_count(filename, line); f = fopen(fn, "r"); if (f == NULL) { failure_start(filename, line, "File doesn't exist: %s", fn); failure_finish(NULL); return (0); } s = (int)strlen(buff); contents = malloc(s * 2 + 128); n = (int)fread(contents, 1, s * 2 + 128 - 1, f); if (n >= 0) contents[n] = '\0'; fclose(f); /* Compare texts. */ btxt = buff; ftxt = (const char *)contents; while (*btxt != '\0' && *ftxt != '\0') { if (*btxt == *ftxt) { ++btxt; ++ftxt; continue; } if (btxt[0] == '\n' && ftxt[0] == '\r' && ftxt[1] == '\n') { /* Pass over different new line characters. */ ++btxt; ftxt += 2; continue; } break; } if (*btxt == '\0' && *ftxt == '\0') { free(contents); return (1); } failure_start(filename, line, "Contents don't match"); logprintf(" file=\"%s\"\n", fn); if (n > 0) { hexdump(contents, buff, n, 0); logprintf(" expected\n", fn); hexdump(buff, contents, s, 0); } else { logprintf(" File empty, contents should be:\n"); hexdump(buff, NULL, s, 0); } failure_finish(NULL); free(contents); return (0); } /* Verify that a text file contains the specified lines, regardless of order */ /* This could be more efficient if we sorted both sets of lines, etc, but * since this is used only for testing and only ever deals with a dozen or so * lines at a time, this relatively crude approach is just fine. */ int assertion_file_contains_lines_any_order(const char *file, int line, const char *pathname, const char *lines[]) { char *buff; size_t buff_size; size_t expected_count, actual_count, i, j; char **expected = NULL; char *p, **actual = NULL; char c; int expected_failure = 0, actual_failure = 0; assertion_count(file, line); buff = slurpfile(&buff_size, "%s", pathname); if (buff == NULL) { failure_start(pathname, line, "Can't read file: %s", pathname); failure_finish(NULL); return (0); } /* Make a copy of the provided lines and count up the expected * file size. */ for (i = 0; lines[i] != NULL; ++i) { } expected_count = i; if (expected_count) { expected = malloc(sizeof(char *) * expected_count); if (expected == NULL) { failure_start(pathname, line, "Can't allocate memory"); failure_finish(NULL); free(expected); free(buff); return (0); } for (i = 0; lines[i] != NULL; ++i) { expected[i] = strdup(lines[i]); } } /* Break the file into lines */ actual_count = 0; for (c = '\0', p = buff; p < buff + buff_size; ++p) { if (*p == '\x0d' || *p == '\x0a') *p = '\0'; if (c == '\0' && *p != '\0') ++actual_count; c = *p; } if (actual_count) { actual = calloc(sizeof(char *), actual_count); if (actual == NULL) { failure_start(pathname, line, "Can't allocate memory"); failure_finish(NULL); free(expected); free(buff); return (0); } for (j = 0, p = buff; p < buff + buff_size; p += 1 + strlen(p)) { if (*p != '\0') { actual[j] = p; ++j; } } } /* Erase matching lines from both lists */ for (i = 0; i < expected_count; ++i) { if (expected[i] == NULL) continue; for (j = 0; j < actual_count; ++j) { if (actual[j] == NULL) continue; if (strcmp(expected[i], actual[j]) == 0) { free(expected[i]); expected[i] = NULL; actual[j] = NULL; break; } } } /* If there's anything left, it's a failure */ for (i = 0; i < expected_count; ++i) { if (expected[i] != NULL) ++expected_failure; } for (j = 0; j < actual_count; ++j) { if (actual[j] != NULL) ++actual_failure; } if (expected_failure == 0 && actual_failure == 0) { free(buff); free(expected); free(actual); return (1); } failure_start(file, line, "File doesn't match: %s", pathname); for (i = 0; i < expected_count; ++i) { if (expected[i] != NULL) { logprintf(" Expected but not present: %s\n", expected[i]); free(expected[i]); } } for (j = 0; j < actual_count; ++j) { if (actual[j] != NULL) logprintf(" Present but not expected: %s\n", actual[j]); } failure_finish(NULL); free(buff); free(expected); free(actual); return (0); } /* Verify that a text file does not contains the specified strings */ int assertion_file_contains_no_invalid_strings(const char *file, int line, const char *pathname, const char *strings[]) { char *buff; int i; buff = slurpfile(NULL, "%s", pathname); if (buff == NULL) { failure_start(file, line, "Can't read file: %s", pathname); failure_finish(NULL); return (0); } for (i = 0; strings[i] != NULL; ++i) { if (strstr(buff, strings[i]) != NULL) { failure_start(file, line, "Invalid string in %s: %s", pathname, strings[i]); failure_finish(NULL); free(buff); return(0); } } free(buff); return (0); } /* Test that two paths point to the same file. */ /* As a side-effect, asserts that both files exist. */ static int is_hardlink(const char *file, int line, const char *path1, const char *path2) { #if defined(_WIN32) && !defined(__CYGWIN__) BY_HANDLE_FILE_INFORMATION bhfi1, bhfi2; int r; assertion_count(file, line); r = my_GetFileInformationByName(path1, &bhfi1); if (r == 0) { failure_start(file, line, "File %s can't be inspected?", path1); failure_finish(NULL); return (0); } r = my_GetFileInformationByName(path2, &bhfi2); if (r == 0) { failure_start(file, line, "File %s can't be inspected?", path2); failure_finish(NULL); return (0); } return (bhfi1.dwVolumeSerialNumber == bhfi2.dwVolumeSerialNumber && bhfi1.nFileIndexHigh == bhfi2.nFileIndexHigh && bhfi1.nFileIndexLow == bhfi2.nFileIndexLow); #else struct stat st1, st2; int r; assertion_count(file, line); r = lstat(path1, &st1); if (r != 0) { failure_start(file, line, "File should exist: %s", path1); failure_finish(NULL); return (0); } r = lstat(path2, &st2); if (r != 0) { failure_start(file, line, "File should exist: %s", path2); failure_finish(NULL); return (0); } return (st1.st_ino == st2.st_ino && st1.st_dev == st2.st_dev); #endif } int assertion_is_hardlink(const char *file, int line, const char *path1, const char *path2) { if (is_hardlink(file, line, path1, path2)) return (1); failure_start(file, line, "Files %s and %s are not hardlinked", path1, path2); failure_finish(NULL); return (0); } int assertion_is_not_hardlink(const char *file, int line, const char *path1, const char *path2) { if (!is_hardlink(file, line, path1, path2)) return (1); failure_start(file, line, "Files %s and %s should not be hardlinked", path1, path2); failure_finish(NULL); return (0); } /* Verify a/b/mtime of 'pathname'. */ /* If 'recent', verify that it's within last 10 seconds. */ static int assertion_file_time(const char *file, int line, const char *pathname, long t, long nsec, char type, int recent) { long long filet, filet_nsec; int r; #if defined(_WIN32) && !defined(__CYGWIN__) #define EPOC_TIME (116444736000000000ULL) FILETIME fxtime, fbirthtime, fatime, fmtime; ULARGE_INTEGER wintm; HANDLE h; fxtime.dwLowDateTime = 0; fxtime.dwHighDateTime = 0; assertion_count(file, line); /* Note: FILE_FLAG_BACKUP_SEMANTICS applies to open * a directory file. If not, CreateFile() will fail when * the pathname is a directory. */ h = CreateFile(pathname, FILE_READ_ATTRIBUTES, 0, NULL, OPEN_EXISTING, FILE_FLAG_BACKUP_SEMANTICS, NULL); if (h == INVALID_HANDLE_VALUE) { failure_start(file, line, "Can't access %s\n", pathname); failure_finish(NULL); return (0); } r = GetFileTime(h, &fbirthtime, &fatime, &fmtime); switch (type) { case 'a': fxtime = fatime; break; case 'b': fxtime = fbirthtime; break; case 'm': fxtime = fmtime; break; } CloseHandle(h); if (r == 0) { failure_start(file, line, "Can't GetFileTime %s\n", pathname); failure_finish(NULL); return (0); } wintm.LowPart = fxtime.dwLowDateTime; wintm.HighPart = fxtime.dwHighDateTime; filet = (wintm.QuadPart - EPOC_TIME) / 10000000; filet_nsec = ((wintm.QuadPart - EPOC_TIME) % 10000000) * 100; nsec = (nsec / 100) * 100; /* Round the request */ #else struct stat st; assertion_count(file, line); r = lstat(pathname, &st); if (r != 0) { failure_start(file, line, "Can't stat %s\n", pathname); failure_finish(NULL); return (0); } switch (type) { case 'a': filet = st.st_atime; break; case 'm': filet = st.st_mtime; break; case 'b': filet = 0; break; default: fprintf(stderr, "INTERNAL: Bad type %c for file time", type); exit(1); } #if defined(__FreeBSD__) switch (type) { case 'a': filet_nsec = st.st_atimespec.tv_nsec; break; case 'b': filet = st.st_birthtime; /* FreeBSD filesystems that don't support birthtime * (e.g., UFS1) always return -1 here. */ if (filet == -1) { return (1); } filet_nsec = st.st_birthtimespec.tv_nsec; break; case 'm': filet_nsec = st.st_mtimespec.tv_nsec; break; default: fprintf(stderr, "INTERNAL: Bad type %c for file time", type); exit(1); } /* FreeBSD generally only stores to microsecond res, so round. */ filet_nsec = (filet_nsec / 1000) * 1000; nsec = (nsec / 1000) * 1000; #else filet_nsec = nsec = 0; /* Generic POSIX only has whole seconds. */ if (type == 'b') return (1); /* Generic POSIX doesn't have birthtime */ #if defined(__HAIKU__) if (type == 'a') return (1); /* Haiku doesn't have atime. */ #endif #endif #endif if (recent) { /* Check that requested time is up-to-date. */ time_t now = time(NULL); if (filet < now - 10 || filet > now + 1) { failure_start(file, line, "File %s has %ctime %lld, %lld seconds ago\n", pathname, type, filet, now - filet); failure_finish(NULL); return (0); } } else if (filet != t || filet_nsec != nsec) { failure_start(file, line, "File %s has %ctime %lld.%09lld, expected %lld.%09lld", pathname, type, filet, filet_nsec, t, nsec); failure_finish(NULL); return (0); } return (1); } /* Verify atime of 'pathname'. */ int assertion_file_atime(const char *file, int line, const char *pathname, long t, long nsec) { return assertion_file_time(file, line, pathname, t, nsec, 'a', 0); } /* Verify atime of 'pathname' is up-to-date. */ int assertion_file_atime_recent(const char *file, int line, const char *pathname) { return assertion_file_time(file, line, pathname, 0, 0, 'a', 1); } /* Verify birthtime of 'pathname'. */ int assertion_file_birthtime(const char *file, int line, const char *pathname, long t, long nsec) { return assertion_file_time(file, line, pathname, t, nsec, 'b', 0); } /* Verify birthtime of 'pathname' is up-to-date. */ int assertion_file_birthtime_recent(const char *file, int line, const char *pathname) { return assertion_file_time(file, line, pathname, 0, 0, 'b', 1); } /* Verify mode of 'pathname'. */ int assertion_file_mode(const char *file, int line, const char *pathname, int expected_mode) { int mode; int r; assertion_count(file, line); #if defined(_WIN32) && !defined(__CYGWIN__) failure_start(file, line, "assertFileMode not yet implemented for Windows"); (void)mode; /* UNUSED */ (void)r; /* UNUSED */ (void)pathname; /* UNUSED */ (void)expected_mode; /* UNUSED */ #else { struct stat st; r = lstat(pathname, &st); mode = (int)(st.st_mode & 0777); } if (r == 0 && mode == expected_mode) return (1); failure_start(file, line, "File %s has mode %o, expected %o", pathname, mode, expected_mode); #endif failure_finish(NULL); return (0); } /* Verify mtime of 'pathname'. */ int assertion_file_mtime(const char *file, int line, const char *pathname, long t, long nsec) { return assertion_file_time(file, line, pathname, t, nsec, 'm', 0); } /* Verify mtime of 'pathname' is up-to-date. */ int assertion_file_mtime_recent(const char *file, int line, const char *pathname) { return assertion_file_time(file, line, pathname, 0, 0, 'm', 1); } /* Verify number of links to 'pathname'. */ int assertion_file_nlinks(const char *file, int line, const char *pathname, int nlinks) { #if defined(_WIN32) && !defined(__CYGWIN__) BY_HANDLE_FILE_INFORMATION bhfi; int r; assertion_count(file, line); r = my_GetFileInformationByName(pathname, &bhfi); if (r != 0 && bhfi.nNumberOfLinks == (DWORD)nlinks) return (1); failure_start(file, line, "File %s has %d links, expected %d", pathname, bhfi.nNumberOfLinks, nlinks); failure_finish(NULL); return (0); #else struct stat st; int r; assertion_count(file, line); r = lstat(pathname, &st); if (r == 0 && (int)st.st_nlink == nlinks) return (1); failure_start(file, line, "File %s has %d links, expected %d", pathname, st.st_nlink, nlinks); failure_finish(NULL); return (0); #endif } /* Verify size of 'pathname'. */ int assertion_file_size(const char *file, int line, const char *pathname, long size) { int64_t filesize; int r; assertion_count(file, line); #if defined(_WIN32) && !defined(__CYGWIN__) { BY_HANDLE_FILE_INFORMATION bhfi; r = !my_GetFileInformationByName(pathname, &bhfi); filesize = ((int64_t)bhfi.nFileSizeHigh << 32) + bhfi.nFileSizeLow; } #else { struct stat st; r = lstat(pathname, &st); filesize = st.st_size; } #endif if (r == 0 && filesize == size) return (1); failure_start(file, line, "File %s has size %ld, expected %ld", pathname, (long)filesize, (long)size); failure_finish(NULL); return (0); } /* Assert that 'pathname' is a dir. If mode >= 0, verify that too. */ int assertion_is_dir(const char *file, int line, const char *pathname, int mode) { struct stat st; int r; #if defined(_WIN32) && !defined(__CYGWIN__) (void)mode; /* UNUSED */ #endif assertion_count(file, line); r = lstat(pathname, &st); if (r != 0) { failure_start(file, line, "Dir should exist: %s", pathname); failure_finish(NULL); return (0); } if (!S_ISDIR(st.st_mode)) { failure_start(file, line, "%s is not a dir", pathname); failure_finish(NULL); return (0); } #if !defined(_WIN32) || defined(__CYGWIN__) /* Windows doesn't handle permissions the same way as POSIX, * so just ignore the mode tests. */ /* TODO: Can we do better here? */ if (mode >= 0 && (mode_t)mode != (st.st_mode & 07777)) { failure_start(file, line, "Dir %s has wrong mode", pathname); logprintf(" Expected: 0%3o\n", mode); logprintf(" Found: 0%3o\n", st.st_mode & 07777); failure_finish(NULL); return (0); } #endif return (1); } /* Verify that 'pathname' is a regular file. If 'mode' is >= 0, * verify that too. */ int assertion_is_reg(const char *file, int line, const char *pathname, int mode) { struct stat st; int r; #if defined(_WIN32) && !defined(__CYGWIN__) (void)mode; /* UNUSED */ #endif assertion_count(file, line); r = lstat(pathname, &st); if (r != 0 || !S_ISREG(st.st_mode)) { failure_start(file, line, "File should exist: %s", pathname); failure_finish(NULL); return (0); } #if !defined(_WIN32) || defined(__CYGWIN__) /* Windows doesn't handle permissions the same way as POSIX, * so just ignore the mode tests. */ /* TODO: Can we do better here? */ if (mode >= 0 && (mode_t)mode != (st.st_mode & 07777)) { failure_start(file, line, "File %s has wrong mode", pathname); logprintf(" Expected: 0%3o\n", mode); logprintf(" Found: 0%3o\n", st.st_mode & 07777); failure_finish(NULL); return (0); } #endif return (1); } /* Check whether 'pathname' is a symbolic link. If 'contents' is * non-NULL, verify that the symlink has those contents. */ static int is_symlink(const char *file, int line, const char *pathname, const char *contents) { #if defined(_WIN32) && !defined(__CYGWIN__) (void)pathname; /* UNUSED */ (void)contents; /* UNUSED */ assertion_count(file, line); /* Windows sort-of has real symlinks, but they're only usable * by privileged users and are crippled even then, so there's * really not much point in bothering with this. */ return (0); #else char buff[300]; struct stat st; ssize_t linklen; int r; assertion_count(file, line); r = lstat(pathname, &st); if (r != 0) { failure_start(file, line, "Symlink should exist: %s", pathname); failure_finish(NULL); return (0); } if (!S_ISLNK(st.st_mode)) return (0); if (contents == NULL) return (1); linklen = readlink(pathname, buff, sizeof(buff)); if (linklen < 0) { failure_start(file, line, "Can't read symlink %s", pathname); failure_finish(NULL); return (0); } buff[linklen] = '\0'; if (strcmp(buff, contents) != 0) return (0); return (1); #endif } /* Assert that path is a symlink that (optionally) contains contents. */ int assertion_is_symlink(const char *file, int line, const char *path, const char *contents) { if (is_symlink(file, line, path, contents)) return (1); if (contents) failure_start(file, line, "File %s is not a symlink to %s", path, contents); else failure_start(file, line, "File %s is not a symlink", path); failure_finish(NULL); return (0); } /* Create a directory and report any errors. */ int assertion_make_dir(const char *file, int line, const char *dirname, int mode) { assertion_count(file, line); #if defined(_WIN32) && !defined(__CYGWIN__) (void)mode; /* UNUSED */ if (0 == _mkdir(dirname)) return (1); #else if (0 == mkdir(dirname, mode)) { if (0 == chmod(dirname, mode)) { assertion_file_mode(file, line, dirname, mode); return (1); } } #endif failure_start(file, line, "Could not create directory %s", dirname); failure_finish(NULL); return(0); } /* Create a file with the specified contents and report any failures. */ int assertion_make_file(const char *file, int line, const char *path, int mode, int csize, const void *contents) { #if defined(_WIN32) && !defined(__CYGWIN__) /* TODO: Rework this to set file mode as well. */ FILE *f; (void)mode; /* UNUSED */ assertion_count(file, line); f = fopen(path, "wb"); if (f == NULL) { failure_start(file, line, "Could not create file %s", path); failure_finish(NULL); return (0); } if (contents != NULL) { size_t wsize; if (csize < 0) wsize = strlen(contents); else wsize = (size_t)csize; if (wsize != fwrite(contents, 1, wsize, f)) { fclose(f); failure_start(file, line, "Could not write file %s", path); failure_finish(NULL); return (0); } } fclose(f); return (1); #else int fd; assertion_count(file, line); fd = open(path, O_CREAT | O_WRONLY, mode >= 0 ? mode : 0644); if (fd < 0) { failure_start(file, line, "Could not create %s", path); failure_finish(NULL); return (0); } if (0 != chmod(path, mode)) { failure_start(file, line, "Could not chmod %s", path); failure_finish(NULL); close(fd); return (0); } if (contents != NULL) { ssize_t wsize; if (csize < 0) wsize = (ssize_t)strlen(contents); else wsize = (ssize_t)csize; if (wsize != write(fd, contents, wsize)) { close(fd); failure_start(file, line, "Could not write to %s", path); failure_finish(NULL); close(fd); return (0); } } close(fd); assertion_file_mode(file, line, path, mode); return (1); #endif } /* Create a hardlink and report any failures. */ int assertion_make_hardlink(const char *file, int line, const char *newpath, const char *linkto) { int succeeded; assertion_count(file, line); #if defined(_WIN32) && !defined(__CYGWIN__) succeeded = my_CreateHardLinkA(newpath, linkto); #elif HAVE_LINK succeeded = !link(linkto, newpath); #else succeeded = 0; #endif if (succeeded) return (1); failure_start(file, line, "Could not create hardlink"); logprintf(" New link: %s\n", newpath); logprintf(" Old name: %s\n", linkto); failure_finish(NULL); return(0); } /* Create a symlink and report any failures. */ int assertion_make_symlink(const char *file, int line, const char *newpath, const char *linkto) { #if defined(_WIN32) && !defined(__CYGWIN__) int targetIsDir = 0; /* TODO: Fix this */ assertion_count(file, line); if (my_CreateSymbolicLinkA(newpath, linkto, targetIsDir)) return (1); #elif HAVE_SYMLINK assertion_count(file, line); if (0 == symlink(linkto, newpath)) return (1); #endif failure_start(file, line, "Could not create symlink"); logprintf(" New link: %s\n", newpath); logprintf(" Old name: %s\n", linkto); failure_finish(NULL); return(0); } /* Set umask, report failures. */ int assertion_umask(const char *file, int line, int mask) { assertion_count(file, line); (void)file; /* UNUSED */ (void)line; /* UNUSED */ umask(mask); return (1); } /* Set times, report failures. */ int assertion_utimes(const char *file, int line, const char *pathname, long at, long at_nsec, long mt, long mt_nsec) { int r; #if defined(_WIN32) && !defined(__CYGWIN__) #define WINTIME(sec, nsec) ((Int32x32To64(sec, 10000000) + EPOC_TIME)\ + (((nsec)/1000)*10)) HANDLE h; ULARGE_INTEGER wintm; FILETIME fatime, fmtime; FILETIME *pat, *pmt; assertion_count(file, line); h = CreateFileA(pathname,GENERIC_READ | GENERIC_WRITE, FILE_SHARE_READ | FILE_SHARE_WRITE, NULL, OPEN_EXISTING, FILE_FLAG_BACKUP_SEMANTICS, NULL); if (h == INVALID_HANDLE_VALUE) { failure_start(file, line, "Can't access %s\n", pathname); failure_finish(NULL); return (0); } if (at > 0 || at_nsec > 0) { wintm.QuadPart = WINTIME(at, at_nsec); fatime.dwLowDateTime = wintm.LowPart; fatime.dwHighDateTime = wintm.HighPart; pat = &fatime; } else pat = NULL; if (mt > 0 || mt_nsec > 0) { wintm.QuadPart = WINTIME(mt, mt_nsec); fmtime.dwLowDateTime = wintm.LowPart; fmtime.dwHighDateTime = wintm.HighPart; pmt = &fmtime; } else pmt = NULL; if (pat != NULL || pmt != NULL) r = SetFileTime(h, NULL, pat, pmt); else r = 1; CloseHandle(h); if (r == 0) { failure_start(file, line, "Can't SetFileTime %s\n", pathname); failure_finish(NULL); return (0); } return (1); #else /* defined(_WIN32) && !defined(__CYGWIN__) */ struct stat st; struct timeval times[2]; #if !defined(__FreeBSD__) mt_nsec = at_nsec = 0; /* Generic POSIX only has whole seconds. */ #endif if (mt == 0 && mt_nsec == 0 && at == 0 && at_nsec == 0) return (1); r = lstat(pathname, &st); if (r < 0) { failure_start(file, line, "Can't stat %s\n", pathname); failure_finish(NULL); return (0); } if (mt == 0 && mt_nsec == 0) { mt = st.st_mtime; #if defined(__FreeBSD__) mt_nsec = st.st_mtimespec.tv_nsec; /* FreeBSD generally only stores to microsecond res, so round. */ mt_nsec = (mt_nsec / 1000) * 1000; #endif } if (at == 0 && at_nsec == 0) { at = st.st_atime; #if defined(__FreeBSD__) at_nsec = st.st_atimespec.tv_nsec; /* FreeBSD generally only stores to microsecond res, so round. */ at_nsec = (at_nsec / 1000) * 1000; #endif } times[1].tv_sec = mt; times[1].tv_usec = mt_nsec / 1000; times[0].tv_sec = at; times[0].tv_usec = at_nsec / 1000; #ifdef HAVE_LUTIMES r = lutimes(pathname, times); #else r = utimes(pathname, times); #endif if (r < 0) { failure_start(file, line, "Can't utimes %s\n", pathname); failure_finish(NULL); return (0); } return (1); #endif /* defined(_WIN32) && !defined(__CYGWIN__) */ } /* Compare file flags */ int assertion_compare_fflags(const char *file, int line, const char *patha, const char *pathb, int nomatch) { #if defined(HAVE_STRUCT_STAT_ST_FLAGS) && defined(UF_NODUMP) struct stat sa, sb; assertion_count(file, line); if (stat(patha, &sa) < 0) return (0); if (stat(pathb, &sb) < 0) return (0); if (!nomatch && sa.st_flags != sb.st_flags) { failure_start(file, line, "File flags should be identical: " "%s=%#010x %s=%#010x", patha, sa.st_flags, pathb, sb.st_flags); failure_finish(NULL); return (0); } if (nomatch && sa.st_flags == sb.st_flags) { failure_start(file, line, "File flags should be different: " "%s=%#010x %s=%#010x", patha, sa.st_flags, pathb, sb.st_flags); failure_finish(NULL); return (0); } #elif (defined(FS_IOC_GETFLAGS) && defined(HAVE_WORKING_FS_IOC_GETFLAGS) && \ defined(FS_NODUMP_FL)) || \ (defined(EXT2_IOC_GETFLAGS) && defined(HAVE_WORKING_EXT2_IOC_GETFLAGS) \ && defined(EXT2_NODUMP_FL)) int fd, r, flagsa, flagsb; assertion_count(file, line); fd = open(patha, O_RDONLY | O_NONBLOCK); if (fd < 0) { failure_start(file, line, "Can't open %s\n", patha); failure_finish(NULL); return (0); } r = ioctl(fd, #ifdef FS_IOC_GETFLAGS FS_IOC_GETFLAGS, #else EXT2_IOC_GETFLAGS, #endif &flagsa); close(fd); if (r < 0) { failure_start(file, line, "Can't get flags %s\n", patha); failure_finish(NULL); return (0); } fd = open(pathb, O_RDONLY | O_NONBLOCK); if (fd < 0) { failure_start(file, line, "Can't open %s\n", pathb); failure_finish(NULL); return (0); } r = ioctl(fd, #ifdef FS_IOC_GETFLAGS FS_IOC_GETFLAGS, #else EXT2_IOC_GETFLAGS, #endif &flagsb); close(fd); if (r < 0) { failure_start(file, line, "Can't get flags %s\n", pathb); failure_finish(NULL); return (0); } if (!nomatch && flagsa != flagsb) { failure_start(file, line, "File flags should be identical: " "%s=%#010x %s=%#010x", patha, flagsa, pathb, flagsb); failure_finish(NULL); return (0); } if (nomatch && flagsa == flagsb) { failure_start(file, line, "File flags should be different: " "%s=%#010x %s=%#010x", patha, flagsa, pathb, flagsb); failure_finish(NULL); return (0); } #else (void)patha; /* UNUSED */ (void)pathb; /* UNUSED */ (void)nomatch; /* UNUSED */ assertion_count(file, line); #endif return (1); } /* Set nodump, report failures. */ int assertion_set_nodump(const char *file, int line, const char *pathname) { #if defined(HAVE_STRUCT_STAT_ST_FLAGS) && defined(UF_NODUMP) int r; assertion_count(file, line); r = chflags(pathname, UF_NODUMP); if (r < 0) { failure_start(file, line, "Can't set nodump %s\n", pathname); failure_finish(NULL); return (0); } #elif (defined(FS_IOC_GETFLAGS) && defined(HAVE_WORKING_FS_IOC_GETFLAGS) && \ defined(FS_NODUMP_FL)) || \ (defined(EXT2_IOC_GETFLAGS) && defined(HAVE_WORKING_EXT2_IOC_GETFLAGS) \ && defined(EXT2_NODUMP_FL)) int fd, r, flags; assertion_count(file, line); fd = open(pathname, O_RDONLY | O_NONBLOCK); if (fd < 0) { failure_start(file, line, "Can't open %s\n", pathname); failure_finish(NULL); return (0); } r = ioctl(fd, #ifdef FS_IOC_GETFLAGS FS_IOC_GETFLAGS, #else EXT2_IOC_GETFLAGS, #endif &flags); if (r < 0) { failure_start(file, line, "Can't get flags %s\n", pathname); failure_finish(NULL); return (0); } #ifdef FS_NODUMP_FL flags |= FS_NODUMP_FL; #else flags |= EXT2_NODUMP_FL; #endif r = ioctl(fd, #ifdef FS_IOC_SETFLAGS FS_IOC_SETFLAGS, #else EXT2_IOC_SETFLAGS, #endif &flags); if (r < 0) { failure_start(file, line, "Can't set nodump %s\n", pathname); failure_finish(NULL); return (0); } close(fd); #else (void)pathname; /* UNUSED */ assertion_count(file, line); #endif return (1); } #ifdef PROGRAM static void assert_version_id(char **qq, size_t *ss) { char *q = *qq; size_t s = *ss; /* Version number is a series of digits and periods. */ while (s > 0 && (*q == '.' || (*q >= '0' && *q <= '9'))) { ++q; --s; } if (q[0] == 'd' && q[1] == 'e' && q[2] == 'v') { q += 3; s -= 3; } /* Skip a single trailing a,b,c, or d. */ if (*q == 'a' || *q == 'b' || *q == 'c' || *q == 'd') ++q; /* Version number terminated by space. */ failure("No space after version: ``%s''", q); assert(s > 1); failure("No space after version: ``%s''", q); assert(*q == ' '); ++q; --s; *qq = q; *ss = s; } /* * Check program version */ void assertVersion(const char *prog, const char *base) { int r; char *p, *q; size_t s; unsigned int prog_len = strlen(base); r = systemf("%s --version >version.stdout 2>version.stderr", prog); if (r != 0) r = systemf("%s -W version >version.stdout 2>version.stderr", prog); failure("Unable to run either %s --version or %s -W version", prog, prog); if (!assert(r == 0)) return; /* --version should generate nothing to stdout. */ assertEmptyFile("version.stderr"); /* Verify format of version message. */ q = p = slurpfile(&s, "version.stdout"); /* Version message should start with name of program, then space. */ assert(s > prog_len + 1); failure("Version must start with '%s': ``%s''", base, p); if (!assertEqualMem(q, base, prog_len)) { free(p); return; } q += prog_len; s -= prog_len; assert(*q == ' '); q++; s--; assert_version_id(&q, &s); /* Separator. */ failure("No `-' between program name and versions: ``%s''", p); assertEqualMem(q, "- ", 2); q += 2; s -= 2; failure("Not long enough for libarchive version: ``%s''", p); assert(s > 11); failure("Libarchive version must start with `libarchive': ``%s''", p); assertEqualMem(q, "libarchive ", 11); q += 11; s -= 11; assert_version_id(&q, &s); /* Skip arbitrary third-party version numbers. */ while (s > 0 && (*q == ' ' || *q == '-' || *q == '/' || *q == '.' || isalnum(*q))) { ++q; --s; } /* All terminated by end-of-line. */ assert(s >= 1); /* Skip an optional CR character (e.g., Windows) */ failure("Version output must end with \\n or \\r\\n"); if (*q == '\r') { ++q; --s; } assertEqualMem(q, "\n", 1); free(p); } #endif /* PROGRAM */ /* * * UTILITIES for use by tests. * */ /* * Check whether platform supports symlinks. This is intended * for tests to use in deciding whether to bother testing symlink * support; if the platform doesn't support symlinks, there's no point * in checking whether the program being tested can create them. * * Note that the first time this test is called, we actually go out to * disk to create and verify a symlink. This is necessary because * symlink support is actually a property of a particular filesystem * and can thus vary between directories on a single system. After * the first call, this returns the cached result from memory, so it's * safe to call it as often as you wish. */ int canSymlink(void) { /* Remember the test result */ static int value = 0, tested = 0; if (tested) return (value); ++tested; assertion_make_file(__FILE__, __LINE__, "canSymlink.0", 0644, 1, "a"); /* Note: Cygwin has its own symlink() emulation that does not * use the Win32 CreateSymbolicLink() function. */ #if defined(_WIN32) && !defined(__CYGWIN__) value = my_CreateSymbolicLinkA("canSymlink.1", "canSymlink.0", 0) && is_symlink(__FILE__, __LINE__, "canSymlink.1", "canSymlink.0"); #elif HAVE_SYMLINK value = (0 == symlink("canSymlink.0", "canSymlink.1")) && is_symlink(__FILE__, __LINE__, "canSymlink.1","canSymlink.0"); #endif return (value); } /* Platform-dependent options for hiding the output of a subcommand. */ #if defined(_WIN32) && !defined(__CYGWIN__) static const char *redirectArgs = ">NUL 2>NUL"; /* Win32 cmd.exe */ #else static const char *redirectArgs = ">/dev/null 2>/dev/null"; /* POSIX 'sh' */ #endif /* * Can this platform run the bzip2 program? */ int canBzip2(void) { static int tested = 0, value = 0; if (!tested) { tested = 1; if (systemf("bzip2 -d -V %s", redirectArgs) == 0) value = 1; } return (value); } /* * Can this platform run the grzip program? */ int canGrzip(void) { static int tested = 0, value = 0; if (!tested) { tested = 1; if (systemf("grzip -V %s", redirectArgs) == 0) value = 1; } return (value); } /* * Can this platform run the gzip program? */ int canGzip(void) { static int tested = 0, value = 0; if (!tested) { tested = 1; if (systemf("gzip -V %s", redirectArgs) == 0) value = 1; } return (value); } /* * Can this platform run the lrzip program? */ int canRunCommand(const char *cmd) { static int tested = 0, value = 0; if (!tested) { tested = 1; if (systemf("%s %s", cmd, redirectArgs) == 0) value = 1; } return (value); } int canLrzip(void) { static int tested = 0, value = 0; if (!tested) { tested = 1; if (systemf("lrzip -V %s", redirectArgs) == 0) value = 1; } return (value); } /* * Can this platform run the lz4 program? */ int canLz4(void) { static int tested = 0, value = 0; if (!tested) { tested = 1; if (systemf("lz4 -V %s", redirectArgs) == 0) value = 1; } return (value); } /* + * Can this platform run the zstd program? + */ +int +canZstd(void) +{ + static int tested = 0, value = 0; + if (!tested) { + tested = 1; + if (systemf("zstd -V %s", redirectArgs) == 0) + value = 1; + } + return (value); +} + +/* * Can this platform run the lzip program? */ int canLzip(void) { static int tested = 0, value = 0; if (!tested) { tested = 1; if (systemf("lzip -V %s", redirectArgs) == 0) value = 1; } return (value); } /* * Can this platform run the lzma program? */ int canLzma(void) { static int tested = 0, value = 0; if (!tested) { tested = 1; if (systemf("lzma -V %s", redirectArgs) == 0) value = 1; } return (value); } /* * Can this platform run the lzop program? */ int canLzop(void) { static int tested = 0, value = 0; if (!tested) { tested = 1; if (systemf("lzop -V %s", redirectArgs) == 0) value = 1; } return (value); } /* * Can this platform run the xz program? */ int canXz(void) { static int tested = 0, value = 0; if (!tested) { tested = 1; if (systemf("xz -V %s", redirectArgs) == 0) value = 1; } return (value); } /* * Can this filesystem handle nodump flags. */ int canNodump(void) { #if defined(HAVE_STRUCT_STAT_ST_FLAGS) && defined(UF_NODUMP) const char *path = "cannodumptest"; struct stat sb; assertion_make_file(__FILE__, __LINE__, path, 0644, 0, NULL); if (chflags(path, UF_NODUMP) < 0) return (0); if (stat(path, &sb) < 0) return (0); if (sb.st_flags & UF_NODUMP) return (1); #elif (defined(FS_IOC_GETFLAGS) && defined(HAVE_WORKING_FS_IOC_GETFLAGS) \ && defined(FS_NODUMP_FL)) || \ (defined(EXT2_IOC_GETFLAGS) && defined(HAVE_WORKING_EXT2_IOC_GETFLAGS) \ && defined(EXT2_NODUMP_FL)) const char *path = "cannodumptest"; int fd, r, flags; assertion_make_file(__FILE__, __LINE__, path, 0644, 0, NULL); fd = open(path, O_RDONLY | O_NONBLOCK); if (fd < 0) return (0); r = ioctl(fd, #ifdef FS_IOC_GETFLAGS FS_IOC_GETFLAGS, #else EXT2_IOC_GETFLAGS, #endif &flags); if (r < 0) return (0); #ifdef FS_NODUMP_FL flags |= FS_NODUMP_FL; #else flags |= EXT2_NODUMP_FL; #endif r = ioctl(fd, #ifdef FS_IOC_SETFLAGS FS_IOC_SETFLAGS, #else EXT2_IOC_SETFLAGS, #endif &flags); if (r < 0) return (0); close(fd); fd = open(path, O_RDONLY | O_NONBLOCK); if (fd < 0) return (0); r = ioctl(fd, #ifdef FS_IOC_GETFLAGS FS_IOC_GETFLAGS, #else EXT2_IOC_GETFLAGS, #endif &flags); if (r < 0) return (0); close(fd); #ifdef FS_NODUMP_FL if (flags & FS_NODUMP_FL) #else if (flags & EXT2_NODUMP_FL) #endif return (1); #endif return (0); } /* Get extended attribute from a path */ const void * getXattr(const char *path, const char *name, size_t *sizep) { void *value = NULL; #if ARCHIVE_XATTR_SUPPORT ssize_t size; #if ARCHIVE_XATTR_LINUX size = lgetxattr(path, name, NULL, 0); #elif ARCHIVE_XATTR_DARWIN size = getxattr(path, name, NULL, 0, 0, XATTR_NOFOLLOW); #elif ARCHIVE_XATTR_AIX size = lgetea(path, name, NULL, 0); #elif ARCHIVE_XATTR_FREEBSD size = extattr_get_link(path, EXTATTR_NAMESPACE_USER, name + 5, NULL, 0); #endif if (size >= 0) { value = malloc(size); #if ARCHIVE_XATTR_LINUX size = lgetxattr(path, name, value, size); #elif ARCHIVE_XATTR_DARWIN size = getxattr(path, name, value, size, 0, XATTR_NOFOLLOW); #elif ARCHIVE_XATTR_AIX size = lgetea(path, name, value, size); #elif ARCHIVE_XATTR_FREEBSD size = extattr_get_link(path, EXTATTR_NAMESPACE_USER, name + 5, value, size); #endif if (size < 0) { free(value); value = NULL; } } if (size < 0) *sizep = 0; else *sizep = (size_t)size; #else /* !ARCHIVE_XATTR_SUPPORT */ (void)path; /* UNUSED */ (void)name; /* UNUSED */ *sizep = 0; #endif /* !ARCHIVE_XATTR_SUPPORT */ return (value); } /* * Set extended attribute on a path * Returns 0 on error, 1 on success */ int setXattr(const char *path, const char *name, const void *value, size_t size) { #if ARCHIVE_XATTR_SUPPORT #if ARCHIVE_XATTR_LINUX if (lsetxattr(path, name, value, size, 0) == 0) #elif ARCHIVE_XATTR_DARWIN if (setxattr(path, name, value, size, 0, XATTR_NOFOLLOW) == 0) #elif ARCHIVE_XATTR_AIX if (lsetea(path, name, value, size, 0) == 0) #elif ARCHIVE_XATTR_FREEBSD if (extattr_set_link(path, EXTATTR_NAMESPACE_USER, name + 5, value, size) > -1) #else if (0) #endif return (1); #else /* !ARCHIVE_XATTR_SUPPORT */ (void)path; /* UNUSED */ (void)name; /* UNUSED */ (void)value; /* UNUSED */ (void)size; /* UNUSED */ #endif /* !ARCHIVE_XATTR_SUPPORT */ return (0); } #if ARCHIVE_ACL_SUNOS /* Fetch ACLs on Solaris using acl() or facl() */ void * sunacl_get(int cmd, int *aclcnt, int fd, const char *path) { int cnt, cntcmd; size_t size; void *aclp; if (cmd == GETACL) { cntcmd = GETACLCNT; size = sizeof(aclent_t); } #if ARCHIVE_ACL_SUNOS_NFS4 else if (cmd == ACE_GETACL) { cntcmd = ACE_GETACLCNT; size = sizeof(ace_t); } #endif else { errno = EINVAL; *aclcnt = -1; return (NULL); } aclp = NULL; cnt = -2; while (cnt == -2 || (cnt == -1 && errno == ENOSPC)) { if (path != NULL) cnt = acl(path, cntcmd, 0, NULL); else cnt = facl(fd, cntcmd, 0, NULL); if (cnt > 0) { if (aclp == NULL) aclp = malloc(cnt * size); else aclp = realloc(NULL, cnt * size); if (aclp != NULL) { if (path != NULL) cnt = acl(path, cmd, cnt, aclp); else cnt = facl(fd, cmd, cnt, aclp); } } else { if (aclp != NULL) { free(aclp); aclp = NULL; } break; } } *aclcnt = cnt; return (aclp); } #endif /* ARCHIVE_ACL_SUNOS */ /* * Set test ACLs on a path * Return values: * 0: error setting ACLs * ARCHIVE_TEST_ACL_TYPE_POSIX1E: POSIX.1E ACLs have been set * ARCHIVE_TEST_ACL_TYPE_NFS4: NFSv4 or extended ACLs have been set */ int setTestAcl(const char *path) { #if ARCHIVE_ACL_SUPPORT int r = 1; #if ARCHIVE_ACL_LIBACL || ARCHIVE_ACL_FREEBSD || ARCHIVE_ACL_DARWIN acl_t acl; #endif #if ARCHIVE_ACL_LIBRICHACL struct richacl *richacl; #endif #if ARCHIVE_ACL_LIBACL || ARCHIVE_ACL_FREEBSD const char *acltext_posix1e = "user:1:rw-," "group:15:r-x," "user::rwx," "group::rwx," "other::r-x," "mask::rwx"; #elif ARCHIVE_ACL_SUNOS /* Solaris POSIX.1e */ aclent_t aclp_posix1e[] = { { USER_OBJ, -1, 4 | 2 | 1 }, { USER, 1, 4 | 2 }, { GROUP_OBJ, -1, 4 | 2 | 1 }, { GROUP, 15, 4 | 1 }, { CLASS_OBJ, -1, 4 | 2 | 1 }, { OTHER_OBJ, -1, 4 | 2 | 1 } }; #endif #if ARCHIVE_ACL_FREEBSD /* FreeBSD NFS4 */ const char *acltext_nfs4 = "user:1:rwpaRcs::allow:1," "group:15:rxaRcs::allow:15," "owner@:rwpxaARWcCos::allow," "group@:rwpxaRcs::allow," "everyone@:rxaRcs::allow"; #elif ARCHIVE_ACL_LIBRICHACL const char *acltext_nfs4 = "owner:rwpxaARWcCoS::mask," "group:rwpxaRcS::mask," "other:rxaRcS::mask," "user:1:rwpaRcS::allow," "group:15:rxaRcS::allow," "owner@:rwpxaARWcCoS::allow," "group@:rwpxaRcS::allow," "everyone@:rxaRcS::allow"; #elif ARCHIVE_ACL_SUNOS_NFS4 /* Solaris NFS4 */ ace_t aclp_nfs4[] = { { 1, ACE_READ_DATA | ACE_WRITE_DATA | ACE_APPEND_DATA | ACE_READ_ATTRIBUTES | ACE_READ_NAMED_ATTRS | ACE_READ_ACL | ACE_SYNCHRONIZE, 0, ACE_ACCESS_ALLOWED_ACE_TYPE }, { 15, ACE_READ_DATA | ACE_EXECUTE | ACE_READ_ATTRIBUTES | ACE_READ_NAMED_ATTRS | ACE_READ_ACL | ACE_SYNCHRONIZE, ACE_IDENTIFIER_GROUP, ACE_ACCESS_ALLOWED_ACE_TYPE }, { -1, ACE_READ_DATA | ACE_WRITE_DATA | ACE_APPEND_DATA | ACE_EXECUTE | ACE_READ_ATTRIBUTES | ACE_WRITE_ATTRIBUTES | ACE_READ_NAMED_ATTRS | ACE_WRITE_NAMED_ATTRS | ACE_READ_ACL | ACE_WRITE_ACL | ACE_WRITE_OWNER | ACE_SYNCHRONIZE, ACE_OWNER, ACE_ACCESS_ALLOWED_ACE_TYPE }, { -1, ACE_READ_DATA | ACE_WRITE_DATA | ACE_APPEND_DATA | ACE_EXECUTE | ACE_READ_ATTRIBUTES | ACE_READ_NAMED_ATTRS | ACE_READ_ACL | ACE_SYNCHRONIZE, ACE_GROUP | ACE_IDENTIFIER_GROUP, ACE_ACCESS_ALLOWED_ACE_TYPE }, { -1, ACE_READ_DATA | ACE_EXECUTE | ACE_READ_ATTRIBUTES | ACE_READ_NAMED_ATTRS | ACE_READ_ACL | ACE_SYNCHRONIZE, ACE_EVERYONE, ACE_ACCESS_ALLOWED_ACE_TYPE } }; #elif ARCHIVE_ACL_DARWIN /* Mac OS X */ acl_entry_t aclent; acl_permset_t permset; const uid_t uid = 1; uuid_t uuid; int i; const acl_perm_t acl_perms[] = { ACL_READ_DATA, ACL_WRITE_DATA, ACL_APPEND_DATA, ACL_EXECUTE, ACL_READ_ATTRIBUTES, ACL_READ_EXTATTRIBUTES, ACL_READ_SECURITY, #if HAVE_DECL_ACL_SYNCHRONIZE ACL_SYNCHRONIZE #endif }; #endif /* ARCHIVE_ACL_DARWIN */ #if ARCHIVE_ACL_FREEBSD acl = acl_from_text(acltext_nfs4); failure("acl_from_text() error: %s", strerror(errno)); if (assert(acl != NULL) == 0) return (0); #elif ARCHIVE_ACL_LIBRICHACL richacl = richacl_from_text(acltext_nfs4, NULL, NULL); failure("richacl_from_text() error: %s", strerror(errno)); if (assert(richacl != NULL) == 0) return (0); #elif ARCHIVE_ACL_DARWIN acl = acl_init(1); failure("acl_init() error: %s", strerror(errno)); if (assert(acl != NULL) == 0) return (0); r = acl_create_entry(&acl, &aclent); failure("acl_create_entry() error: %s", strerror(errno)); if (assertEqualInt(r, 0) == 0) goto testacl_free; r = acl_set_tag_type(aclent, ACL_EXTENDED_ALLOW); failure("acl_set_tag_type() error: %s", strerror(errno)); if (assertEqualInt(r, 0) == 0) goto testacl_free; r = acl_get_permset(aclent, &permset); failure("acl_get_permset() error: %s", strerror(errno)); if (assertEqualInt(r, 0) == 0) goto testacl_free; for (i = 0; i < (int)(sizeof(acl_perms) / sizeof(acl_perms[0])); i++) { r = acl_add_perm(permset, acl_perms[i]); failure("acl_add_perm() error: %s", strerror(errno)); if (assertEqualInt(r, 0) == 0) goto testacl_free; } r = acl_set_permset(aclent, permset); failure("acl_set_permset() error: %s", strerror(errno)); if (assertEqualInt(r, 0) == 0) goto testacl_free; r = mbr_uid_to_uuid(uid, uuid); failure("mbr_uid_to_uuid() error: %s", strerror(errno)); if (assertEqualInt(r, 0) == 0) goto testacl_free; r = acl_set_qualifier(aclent, uuid); failure("acl_set_qualifier() error: %s", strerror(errno)); if (assertEqualInt(r, 0) == 0) goto testacl_free; #endif /* ARCHIVE_ACL_DARWIN */ #if ARCHIVE_ACL_NFS4 #if ARCHIVE_ACL_FREEBSD r = acl_set_file(path, ACL_TYPE_NFS4, acl); acl_free(acl); #elif ARCHIVE_ACL_LIBRICHACL r = richacl_set_file(path, richacl); richacl_free(richacl); #elif ARCHIVE_ACL_SUNOS_NFS4 r = acl(path, ACE_SETACL, (int)(sizeof(aclp_nfs4)/sizeof(aclp_nfs4[0])), aclp_nfs4); #elif ARCHIVE_ACL_DARWIN r = acl_set_file(path, ACL_TYPE_EXTENDED, acl); acl_free(acl); #endif if (r == 0) return (ARCHIVE_TEST_ACL_TYPE_NFS4); #endif /* ARCHIVE_ACL_NFS4 */ #if ARCHIVE_ACL_POSIX1E #if ARCHIVE_ACL_FREEBSD || ARCHIVE_ACL_LIBACL acl = acl_from_text(acltext_posix1e); failure("acl_from_text() error: %s", strerror(errno)); if (assert(acl != NULL) == 0) return (0); r = acl_set_file(path, ACL_TYPE_ACCESS, acl); acl_free(acl); #elif ARCHIVE_ACL_SUNOS r = acl(path, SETACL, (int)(sizeof(aclp_posix1e)/sizeof(aclp_posix1e[0])), aclp_posix1e); #endif if (r == 0) return (ARCHIVE_TEST_ACL_TYPE_POSIX1E); else return (0); #endif /* ARCHIVE_ACL_POSIX1E */ #if ARCHIVE_ACL_DARWIN testacl_free: acl_free(acl); #endif #endif /* ARCHIVE_ACL_SUPPORT */ (void)path; /* UNUSED */ return (0); } /* * Sleep as needed; useful for verifying disk timestamp changes by * ensuring that the wall-clock time has actually changed before we * go back to re-read something from disk. */ void sleepUntilAfter(time_t t) { while (t >= time(NULL)) #if defined(_WIN32) && !defined(__CYGWIN__) Sleep(500); #else sleep(1); #endif } /* * Call standard system() call, but build up the command line using * sprintf() conventions. */ int systemf(const char *fmt, ...) { char buff[8192]; va_list ap; int r; va_start(ap, fmt); vsprintf(buff, fmt, ap); if (verbosity > VERBOSITY_FULL) logprintf("Cmd: %s\n", buff); r = system(buff); va_end(ap); return (r); } /* * Slurp a file into memory for ease of comparison and testing. * Returns size of file in 'sizep' if non-NULL, null-terminates * data in memory for ease of use. */ char * slurpfile(size_t * sizep, const char *fmt, ...) { char filename[8192]; struct stat st; va_list ap; char *p; ssize_t bytes_read; FILE *f; int r; va_start(ap, fmt); vsprintf(filename, fmt, ap); va_end(ap); f = fopen(filename, "rb"); if (f == NULL) { /* Note: No error; non-existent file is okay here. */ return (NULL); } r = fstat(fileno(f), &st); if (r != 0) { logprintf("Can't stat file %s\n", filename); fclose(f); return (NULL); } p = malloc((size_t)st.st_size + 1); if (p == NULL) { logprintf("Can't allocate %ld bytes of memory to read file %s\n", (long int)st.st_size, filename); fclose(f); return (NULL); } bytes_read = fread(p, 1, (size_t)st.st_size, f); if (bytes_read < st.st_size) { logprintf("Can't read file %s\n", filename); fclose(f); free(p); return (NULL); } p[st.st_size] = '\0'; if (sizep != NULL) *sizep = (size_t)st.st_size; fclose(f); return (p); } /* * Slurp a file into memory for ease of comparison and testing. * Returns size of file in 'sizep' if non-NULL, null-terminates * data in memory for ease of use. */ void dumpfile(const char *filename, void *data, size_t len) { ssize_t bytes_written; FILE *f; f = fopen(filename, "wb"); if (f == NULL) { logprintf("Can't open file %s for writing\n", filename); return; } bytes_written = fwrite(data, 1, len, f); if (bytes_written < (ssize_t)len) logprintf("Can't write file %s\n", filename); fclose(f); } /* Read a uuencoded file from the reference directory, decode, and * write the result into the current directory. */ #define VALID_UUDECODE(c) (c >= 32 && c <= 96) #define UUDECODE(c) (((c) - 0x20) & 0x3f) void extract_reference_file(const char *name) { char buff[1024]; FILE *in, *out; sprintf(buff, "%s/%s.uu", refdir, name); in = fopen(buff, "r"); failure("Couldn't open reference file %s", buff); assert(in != NULL); if (in == NULL) return; /* Read up to and including the 'begin' line. */ for (;;) { if (fgets(buff, sizeof(buff), in) == NULL) { /* TODO: This is a failure. */ return; } if (memcmp(buff, "begin ", 6) == 0) break; } /* Now, decode the rest and write it. */ out = fopen(name, "wb"); while (fgets(buff, sizeof(buff), in) != NULL) { char *p = buff; int bytes; if (memcmp(buff, "end", 3) == 0) break; bytes = UUDECODE(*p++); while (bytes > 0) { int n = 0; /* Write out 1-3 bytes from that. */ if (bytes > 0) { assert(VALID_UUDECODE(p[0])); assert(VALID_UUDECODE(p[1])); n = UUDECODE(*p++) << 18; n |= UUDECODE(*p++) << 12; fputc(n >> 16, out); --bytes; } if (bytes > 0) { assert(VALID_UUDECODE(p[0])); n |= UUDECODE(*p++) << 6; fputc((n >> 8) & 0xFF, out); --bytes; } if (bytes > 0) { assert(VALID_UUDECODE(p[0])); n |= UUDECODE(*p++); fputc(n & 0xFF, out); --bytes; } } } fclose(out); fclose(in); } void copy_reference_file(const char *name) { char buff[1024]; FILE *in, *out; size_t rbytes; sprintf(buff, "%s/%s", refdir, name); in = fopen(buff, "rb"); failure("Couldn't open reference file %s", buff); assert(in != NULL); if (in == NULL) return; /* Now, decode the rest and write it. */ /* Not a lot of error checking here; the input better be right. */ out = fopen(name, "wb"); while ((rbytes = fread(buff, 1, sizeof(buff), in)) > 0) { if (fwrite(buff, 1, rbytes, out) != rbytes) { logprintf("Error: fwrite\n"); break; } } fclose(out); fclose(in); } int is_LargeInode(const char *file) { #if defined(_WIN32) && !defined(__CYGWIN__) BY_HANDLE_FILE_INFORMATION bhfi; int r; r = my_GetFileInformationByName(file, &bhfi); if (r != 0) return (0); return (bhfi.nFileIndexHigh & 0x0000FFFFUL); #else struct stat st; int64_t ino; if (stat(file, &st) < 0) return (0); ino = (int64_t)st.st_ino; return (ino > 0xffffffff); #endif } void extract_reference_files(const char **names) { while (names && *names) extract_reference_file(*names++); } #ifndef PROGRAM /* Set ACLs */ int assertion_entry_set_acls(const char *file, int line, struct archive_entry *ae, struct archive_test_acl_t *acls, int n) { int i, r, ret; assertion_count(file, line); ret = 0; archive_entry_acl_clear(ae); for (i = 0; i < n; i++) { r = archive_entry_acl_add_entry(ae, acls[i].type, acls[i].permset, acls[i].tag, acls[i].qual, acls[i].name); if (r != 0) { ret = 1; failure_start(file, line, "type=%#010x, ", "permset=%#010x, tag=%d, qual=%d name=%s", acls[i].type, acls[i].permset, acls[i].tag, acls[i].qual, acls[i].name); failure_finish(NULL); } } return (ret); } static int archive_test_acl_match(struct archive_test_acl_t *acl, int type, int permset, int tag, int qual, const char *name) { if (type != acl->type) return (0); if (permset != acl->permset) return (0); if (tag != acl->tag) return (0); if (tag == ARCHIVE_ENTRY_ACL_USER_OBJ) return (1); if (tag == ARCHIVE_ENTRY_ACL_GROUP_OBJ) return (1); if (tag == ARCHIVE_ENTRY_ACL_EVERYONE) return (1); if (tag == ARCHIVE_ENTRY_ACL_OTHER) return (1); if (qual != acl->qual) return (0); if (name == NULL) { if (acl->name == NULL || acl->name[0] == '\0') return (1); return (0); } if (acl->name == NULL) { if (name[0] == '\0') return (1); return (0); } return (0 == strcmp(name, acl->name)); } /* Compare ACLs */ int assertion_entry_compare_acls(const char *file, int line, struct archive_entry *ae, struct archive_test_acl_t *acls, int cnt, int want_type, int mode) { int *marker; int i, r, n, ret; int type, permset, tag, qual; int matched; const char *name; assertion_count(file, line); ret = 0; n = 0; marker = malloc(sizeof(marker[0]) * cnt); for (i = 0; i < cnt; i++) { if ((acls[i].type & want_type) != 0) { marker[n] = i; n++; } } if (n == 0) { failure_start(file, line, "No ACL's to compare, type mask: %d", want_type); return (1); } while (0 == (r = archive_entry_acl_next(ae, want_type, &type, &permset, &tag, &qual, &name))) { for (i = 0, matched = 0; i < n && !matched; i++) { if (archive_test_acl_match(&acls[marker[i]], type, permset, tag, qual, name)) { /* We found a match; remove it. */ marker[i] = marker[n - 1]; n--; matched = 1; } } if (type == ARCHIVE_ENTRY_ACL_TYPE_ACCESS && tag == ARCHIVE_ENTRY_ACL_USER_OBJ) { if (!matched) { failure_start(file, line, "No match for " "user_obj perm"); failure_finish(NULL); ret = 1; } if ((permset << 6) != (mode & 0700)) { failure_start(file, line, "USER_OBJ permset " "(%02o) != user mode (%02o)", permset, 07 & (mode >> 6)); failure_finish(NULL); ret = 1; } } else if (type == ARCHIVE_ENTRY_ACL_TYPE_ACCESS && tag == ARCHIVE_ENTRY_ACL_GROUP_OBJ) { if (!matched) { failure_start(file, line, "No match for " "group_obj perm"); failure_finish(NULL); ret = 1; } if ((permset << 3) != (mode & 0070)) { failure_start(file, line, "GROUP_OBJ permset " "(%02o) != group mode (%02o)", permset, 07 & (mode >> 3)); failure_finish(NULL); ret = 1; } } else if (type == ARCHIVE_ENTRY_ACL_TYPE_ACCESS && tag == ARCHIVE_ENTRY_ACL_OTHER) { if (!matched) { failure_start(file, line, "No match for " "other perm"); failure_finish(NULL); ret = 1; } if ((permset << 0) != (mode & 0007)) { failure_start(file, line, "OTHER permset " "(%02o) != other mode (%02o)", permset, mode & 07); failure_finish(NULL); ret = 1; } } else if (matched != 1) { failure_start(file, line, "Could not find match for " "ACL (type=%#010x,permset=%#010x,tag=%d,qual=%d," "name=``%s'')", type, permset, tag, qual, name); failure_finish(NULL); ret = 1; } } if (r != ARCHIVE_EOF) { failure_start(file, line, "Should not exit before EOF"); failure_finish(NULL); ret = 1; } if ((want_type & ARCHIVE_ENTRY_ACL_TYPE_ACCESS) != 0 && (mode_t)(mode & 0777) != (archive_entry_mode(ae) & 0777)) { failure_start(file, line, "Mode (%02o) and entry mode (%02o) " "mismatch", mode, archive_entry_mode(ae)); failure_finish(NULL); ret = 1; } if (n != 0) { failure_start(file, line, "Could not find match for ACL " "(type=%#010x,permset=%#010x,tag=%d,qual=%d,name=``%s'')", acls[marker[0]].type, acls[marker[0]].permset, acls[marker[0]].tag, acls[marker[0]].qual, acls[marker[0]].name); failure_finish(NULL); ret = 1; /* Number of ACLs not matched should == 0 */ } free(marker); return (ret); } #endif /* !defined(PROGRAM) */ /* * * TEST management * */ /* * "list.h" is simply created by "grep DEFINE_TEST test_*.c"; it has * a line like * DEFINE_TEST(test_function) * for each test. */ /* Use "list.h" to declare all of the test functions. */ #undef DEFINE_TEST #define DEFINE_TEST(name) void name(void); #include "list.h" /* Use "list.h" to create a list of all tests (functions and names). */ #undef DEFINE_TEST #define DEFINE_TEST(n) { n, #n, 0 }, struct test_list_t tests[] = { #include "list.h" }; /* * Summarize repeated failures in the just-completed test. */ static void test_summarize(int failed, int skips_num) { unsigned int i; switch (verbosity) { case VERBOSITY_SUMMARY_ONLY: printf(failed ? "E" : "."); fflush(stdout); break; case VERBOSITY_PASSFAIL: printf(failed ? "FAIL\n" : skips_num ? "ok (S)\n" : "ok\n"); break; } log_console = (verbosity == VERBOSITY_LIGHT_REPORT); for (i = 0; i < sizeof(failed_lines)/sizeof(failed_lines[0]); i++) { if (failed_lines[i].count > 1 && !failed_lines[i].skip) logprintf("%s:%d: Summary: Failed %d times\n", failed_filename, i, failed_lines[i].count); } /* Clear the failure history for the next file. */ failed_filename = NULL; memset(failed_lines, 0, sizeof(failed_lines)); } /* * Actually run a single test, with appropriate setup and cleanup. */ static int test_run(int i, const char *tmpdir) { char workdir[1024]; char logfilename[64]; int failures_before = failures; int skips_before = skips; int oldumask; switch (verbosity) { case VERBOSITY_SUMMARY_ONLY: /* No per-test reports at all */ break; case VERBOSITY_PASSFAIL: /* rest of line will include ok/FAIL marker */ printf("%3d: %-64s", i, tests[i].name); fflush(stdout); break; default: /* Title of test, details will follow */ printf("%3d: %s\n", i, tests[i].name); } /* Chdir to the top-level work directory. */ if (!assertChdir(tmpdir)) { fprintf(stderr, "ERROR: Can't chdir to top work dir %s\n", tmpdir); exit(1); } /* Create a log file for this test. */ sprintf(logfilename, "%s.log", tests[i].name); logfile = fopen(logfilename, "w"); fprintf(logfile, "%s\n\n", tests[i].name); /* Chdir() to a work dir for this specific test. */ snprintf(workdir, sizeof(workdir), "%s/%s", tmpdir, tests[i].name); testworkdir = workdir; if (!assertMakeDir(testworkdir, 0755) || !assertChdir(testworkdir)) { fprintf(stderr, "ERROR: Can't chdir to work dir %s\n", testworkdir); exit(1); } /* Explicitly reset the locale before each test. */ setlocale(LC_ALL, "C"); /* Record the umask before we run the test. */ umask(oldumask = umask(0)); /* * Run the actual test. */ (*tests[i].func)(); /* * Clean up and report afterwards. */ testworkdir = NULL; /* Restore umask */ umask(oldumask); /* Reset locale. */ setlocale(LC_ALL, "C"); /* Reset directory. */ if (!assertChdir(tmpdir)) { fprintf(stderr, "ERROR: Couldn't chdir to temp dir %s\n", tmpdir); exit(1); } /* Report per-test summaries. */ tests[i].failures = failures - failures_before; test_summarize(tests[i].failures, skips - skips_before); /* Close the per-test log file. */ fclose(logfile); logfile = NULL; /* If there were no failures, we can remove the work dir and logfile. */ if (tests[i].failures == 0) { if (!keep_temp_files && assertChdir(tmpdir)) { #if defined(_WIN32) && !defined(__CYGWIN__) /* Make sure not to leave empty directories. * Sometimes a processing of closing files used by tests * is not done, then rmdir will be failed and it will * leave a empty test directory. So we should wait a few * seconds and retry rmdir. */ int r, t; for (t = 0; t < 10; t++) { if (t > 0) Sleep(1000); r = systemf("rmdir /S /Q %s", tests[i].name); if (r == 0) break; } systemf("del %s", logfilename); #else systemf("rm -rf %s", tests[i].name); systemf("rm %s", logfilename); #endif } } /* Return appropriate status. */ return (tests[i].failures); } /* * * * MAIN and support routines. * * */ static void usage(const char *program) { static const int limit = sizeof(tests) / sizeof(tests[0]); int i; printf("Usage: %s [options] ...\n", program); printf("Default is to run all tests.\n"); printf("Otherwise, specify the numbers of the tests you wish to run.\n"); printf("Options:\n"); printf(" -d Dump core after any failure, for debugging.\n"); printf(" -k Keep all temp files.\n"); printf(" Default: temp files for successful tests deleted.\n"); #ifdef PROGRAM printf(" -p Path to executable to be tested.\n"); printf(" Default: path taken from " ENVBASE " environment variable.\n"); #endif printf(" -q Quiet.\n"); printf(" -r Path to dir containing reference files.\n"); printf(" Default: Current directory.\n"); printf(" -u Keep running specifies tests until one fails.\n"); printf(" -v Verbose.\n"); printf("Available tests:\n"); for (i = 0; i < limit; i++) printf(" %d: %s\n", i, tests[i].name); exit(1); } static char * get_refdir(const char *d) { size_t tried_size, buff_size; char *buff, *tried, *pwd = NULL, *p = NULL; #ifdef PATH_MAX buff_size = PATH_MAX; #else buff_size = 8192; #endif buff = calloc(buff_size, 1); if (buff == NULL) { fprintf(stderr, "Unable to allocate memory\n"); exit(1); } /* Allocate a buffer to hold the various directories we checked. */ tried_size = buff_size * 2; tried = calloc(tried_size, 1); if (tried == NULL) { fprintf(stderr, "Unable to allocate memory\n"); exit(1); } /* If a dir was specified, try that */ if (d != NULL) { pwd = NULL; snprintf(buff, buff_size, "%s", d); p = slurpfile(NULL, "%s/%s", buff, KNOWNREF); if (p != NULL) goto success; strncat(tried, buff, tried_size - strlen(tried) - 1); strncat(tried, "\n", tried_size - strlen(tried) - 1); goto failure; } /* Get the current dir. */ #ifdef PATH_MAX pwd = getcwd(NULL, PATH_MAX);/* Solaris getcwd needs the size. */ #else pwd = getcwd(NULL, 0); #endif while (pwd[strlen(pwd) - 1] == '\n') pwd[strlen(pwd) - 1] = '\0'; /* Look for a known file. */ snprintf(buff, buff_size, "%s", pwd); p = slurpfile(NULL, "%s/%s", buff, KNOWNREF); if (p != NULL) goto success; strncat(tried, buff, tried_size - strlen(tried) - 1); strncat(tried, "\n", tried_size - strlen(tried) - 1); snprintf(buff, buff_size, "%s/test", pwd); p = slurpfile(NULL, "%s/%s", buff, KNOWNREF); if (p != NULL) goto success; strncat(tried, buff, tried_size - strlen(tried) - 1); strncat(tried, "\n", tried_size - strlen(tried) - 1); #if defined(LIBRARY) snprintf(buff, buff_size, "%s/%s/test", pwd, LIBRARY); #else snprintf(buff, buff_size, "%s/%s/test", pwd, PROGRAM); #endif p = slurpfile(NULL, "%s/%s", buff, KNOWNREF); if (p != NULL) goto success; strncat(tried, buff, tried_size - strlen(tried) - 1); strncat(tried, "\n", tried_size - strlen(tried) - 1); #if defined(PROGRAM_ALIAS) snprintf(buff, buff_size, "%s/%s/test", pwd, PROGRAM_ALIAS); p = slurpfile(NULL, "%s/%s", buff, KNOWNREF); if (p != NULL) goto success; strncat(tried, buff, tried_size - strlen(tried) - 1); strncat(tried, "\n", tried_size - strlen(tried) - 1); #endif if (memcmp(pwd, "/usr/obj", 8) == 0) { snprintf(buff, buff_size, "%s", pwd + 8); p = slurpfile(NULL, "%s/%s", buff, KNOWNREF); if (p != NULL) goto success; strncat(tried, buff, tried_size - strlen(tried) - 1); strncat(tried, "\n", tried_size - strlen(tried) - 1); snprintf(buff, buff_size, "%s/test", pwd + 8); p = slurpfile(NULL, "%s/%s", buff, KNOWNREF); if (p != NULL) goto success; strncat(tried, buff, tried_size - strlen(tried) - 1); strncat(tried, "\n", tried_size - strlen(tried) - 1); } failure: printf("Unable to locate known reference file %s\n", KNOWNREF); printf(" Checked following directories:\n%s\n", tried); printf("Use -r option to specify full path to reference directory\n"); #if defined(_WIN32) && !defined(__CYGWIN__) && defined(_DEBUG) DebugBreak(); #endif exit(1); success: free(p); free(pwd); free(tried); /* Copy result into a fresh buffer to reduce memory usage. */ p = strdup(buff); free(buff); return p; } int main(int argc, char **argv) { static const int limit = sizeof(tests) / sizeof(tests[0]); int test_set[sizeof(tests) / sizeof(tests[0])]; int i = 0, j = 0, tests_run = 0, tests_failed = 0, option; time_t now; char *refdir_alloc = NULL; const char *progname; char **saved_argv; const char *tmp, *option_arg, *p; char tmpdir[256], *pwd, *testprogdir, *tmp2 = NULL, *vlevel = NULL; char tmpdir_timestamp[256]; (void)argc; /* UNUSED */ /* Get the current dir. */ #ifdef PATH_MAX pwd = getcwd(NULL, PATH_MAX);/* Solaris getcwd needs the size. */ #else pwd = getcwd(NULL, 0); #endif while (pwd[strlen(pwd) - 1] == '\n') pwd[strlen(pwd) - 1] = '\0'; #if defined(HAVE__CrtSetReportMode) && !defined(__WATCOMC__) /* To stop to run the default invalid parameter handler. */ _set_invalid_parameter_handler(invalid_parameter_handler); /* Disable annoying assertion message box. */ _CrtSetReportMode(_CRT_ASSERT, 0); #endif /* * Name of this program, used to build root of our temp directory * tree. */ progname = p = argv[0]; if ((testprogdir = (char *)malloc(strlen(progname) + 1)) == NULL) { fprintf(stderr, "ERROR: Out of memory."); exit(1); } strcpy(testprogdir, progname); while (*p != '\0') { /* Support \ or / dir separators for Windows compat. */ if (*p == '/' || *p == '\\') { progname = p + 1; i = j; } ++p; j++; } testprogdir[i] = '\0'; #if defined(_WIN32) && !defined(__CYGWIN__) if (testprogdir[0] != '/' && testprogdir[0] != '\\' && !(((testprogdir[0] >= 'a' && testprogdir[0] <= 'z') || (testprogdir[0] >= 'A' && testprogdir[0] <= 'Z')) && testprogdir[1] == ':' && (testprogdir[2] == '/' || testprogdir[2] == '\\'))) #else if (testprogdir[0] != '/') #endif { /* Fixup path for relative directories. */ if ((testprogdir = (char *)realloc(testprogdir, strlen(pwd) + 1 + strlen(testprogdir) + 1)) == NULL) { fprintf(stderr, "ERROR: Out of memory."); exit(1); } memmove(testprogdir + strlen(pwd) + 1, testprogdir, strlen(testprogdir) + 1); memcpy(testprogdir, pwd, strlen(pwd)); testprogdir[strlen(pwd)] = '/'; } #ifdef PROGRAM /* Get the target program from environment, if available. */ testprogfile = getenv(ENVBASE); #endif if (getenv("TMPDIR") != NULL) tmp = getenv("TMPDIR"); else if (getenv("TMP") != NULL) tmp = getenv("TMP"); else if (getenv("TEMP") != NULL) tmp = getenv("TEMP"); else if (getenv("TEMPDIR") != NULL) tmp = getenv("TEMPDIR"); else tmp = "/tmp"; /* Allow -d to be controlled through the environment. */ if (getenv(ENVBASE "_DEBUG") != NULL) dump_on_failure = 1; /* Allow -v to be controlled through the environment. */ if (getenv("_VERBOSITY_LEVEL") != NULL) { vlevel = getenv("_VERBOSITY_LEVEL"); verbosity = atoi(vlevel); if (verbosity < VERBOSITY_SUMMARY_ONLY || verbosity > VERBOSITY_FULL) { /* Unsupported verbosity levels are silently ignored */ vlevel = NULL; verbosity = VERBOSITY_PASSFAIL; } } /* Get the directory holding test files from environment. */ refdir = getenv(ENVBASE "_TEST_FILES"); /* * Parse options, without using getopt(), which isn't available * on all platforms. */ ++argv; /* Skip program name */ while (*argv != NULL) { if (**argv != '-') break; p = *argv++; ++p; /* Skip '-' */ while (*p != '\0') { option = *p++; option_arg = NULL; /* If 'opt' takes an argument, parse that. */ if (option == 'p' || option == 'r') { if (*p != '\0') option_arg = p; else if (*argv == NULL) { fprintf(stderr, "Option -%c requires argument.\n", option); usage(progname); } else option_arg = *argv++; p = ""; /* End of this option word. */ } /* Now, handle the option. */ switch (option) { case 'd': dump_on_failure = 1; break; case 'k': keep_temp_files = 1; break; case 'p': #ifdef PROGRAM testprogfile = option_arg; #else fprintf(stderr, "-p option not permitted\n"); usage(progname); #endif break; case 'q': if (!vlevel) verbosity--; break; case 'r': refdir = option_arg; break; case 'u': until_failure++; break; case 'v': if (!vlevel) verbosity++; break; default: fprintf(stderr, "Unrecognized option '%c'\n", option); usage(progname); } } } /* * Sanity-check that our options make sense. */ #ifdef PROGRAM if (testprogfile == NULL) { if ((tmp2 = (char *)malloc(strlen(testprogdir) + 1 + strlen(PROGRAM) + 1)) == NULL) { fprintf(stderr, "ERROR: Out of memory."); exit(1); } strcpy(tmp2, testprogdir); strcat(tmp2, "/"); strcat(tmp2, PROGRAM); testprogfile = tmp2; } { char *testprg; #if defined(_WIN32) && !defined(__CYGWIN__) /* Command.com sometimes rejects '/' separators. */ testprg = strdup(testprogfile); for (i = 0; testprg[i] != '\0'; i++) { if (testprg[i] == '/') testprg[i] = '\\'; } testprogfile = testprg; #endif /* Quote the name that gets put into shell command lines. */ testprg = malloc(strlen(testprogfile) + 3); strcpy(testprg, "\""); strcat(testprg, testprogfile); strcat(testprg, "\""); testprog = testprg; } #endif #if !defined(_WIN32) && defined(SIGPIPE) { /* Ignore SIGPIPE signals */ struct sigaction sa; sa.sa_handler = SIG_IGN; sigemptyset(&sa.sa_mask); sa.sa_flags = 0; sigaction(SIGPIPE, &sa, NULL); } #endif /* * Create a temp directory for the following tests. * Include the time the tests started as part of the name, * to make it easier to track the results of multiple tests. */ now = time(NULL); for (i = 0; ; i++) { strftime(tmpdir_timestamp, sizeof(tmpdir_timestamp), "%Y-%m-%dT%H.%M.%S", localtime(&now)); sprintf(tmpdir, "%s/%s.%s-%03d", tmp, progname, tmpdir_timestamp, i); if (assertMakeDir(tmpdir,0755)) break; if (i >= 999) { fprintf(stderr, "ERROR: Unable to create temp directory %s\n", tmpdir); exit(1); } } /* * If the user didn't specify a directory for locating * reference files, try to find the reference files in * the "usual places." */ refdir = refdir_alloc = get_refdir(refdir); /* * Banner with basic information. */ printf("\n"); printf("If tests fail or crash, details will be in:\n"); printf(" %s\n", tmpdir); printf("\n"); if (verbosity > VERBOSITY_SUMMARY_ONLY) { printf("Reference files will be read from: %s\n", refdir); #ifdef PROGRAM printf("Running tests on: %s\n", testprog); #endif printf("Exercising: "); fflush(stdout); printf("%s\n", EXTRA_VERSION); } else { printf("Running "); fflush(stdout); } /* * Run some or all of the individual tests. */ saved_argv = argv; do { argv = saved_argv; do { int test_num; test_num = get_test_set(test_set, limit, *argv, tests); if (test_num < 0) { printf("*** INVALID Test %s\n", *argv); free(refdir_alloc); free(testprogdir); usage(progname); return (1); } for (i = 0; i < test_num; i++) { tests_run++; if (test_run(test_set[i], tmpdir)) { tests_failed++; if (until_failure) goto finish; } } if (*argv != NULL) argv++; } while (*argv != NULL); } while (until_failure); finish: /* Must be freed after all tests run */ free(tmp2); free(testprogdir); free(pwd); /* * Report summary statistics. */ if (verbosity > VERBOSITY_SUMMARY_ONLY) { printf("\n"); printf("Totals:\n"); printf(" Tests run: %8d\n", tests_run); printf(" Tests failed: %8d\n", tests_failed); printf(" Assertions checked:%8d\n", assertions); printf(" Assertions failed: %8d\n", failures); printf(" Skips reported: %8d\n", skips); } if (failures) { printf("\n"); printf("Failing tests:\n"); for (i = 0; i < limit; ++i) { if (tests[i].failures) printf(" %d: %s (%d failures)\n", i, tests[i].name, tests[i].failures); } printf("\n"); printf("Details for failing tests: %s\n", tmpdir); printf("\n"); } else { if (verbosity == VERBOSITY_SUMMARY_ONLY) printf("\n"); printf("%d tests passed, no failures\n", tests_run); } free(refdir_alloc); /* If the final tmpdir is empty, we can remove it. */ /* This should be the usual case when all tests succeed. */ assertChdir(".."); rmdir(tmpdir); return (tests_failed ? 1 : 0); } Index: stable/10/lib/libarchive/Makefile =================================================================== --- stable/10/lib/libarchive/Makefile (revision 324417) +++ stable/10/lib/libarchive/Makefile (revision 324418) @@ -1,421 +1,423 @@ # $FreeBSD$ .include LIBARCHIVEDIR= ${SRCTOP}/contrib/libarchive LIB= archive DPADD= ${LIBZ} ${LIBBZ2} ${LIBLZMA} ${LIBPTHREAD} ${LIBBSDXML} LDADD= -lz -lbz2 -llzma -lpthread -lbsdxml CFLAGS+= -DHAVE_BZLIB_H=1 -DHAVE_LIBLZMA=1 -DHAVE_LZMA_H=1 # FreeBSD SHLIB_MAJOR value is managed as part of the FreeBSD system. # It has no real relation to the libarchive version number. SHLIB_MAJOR= 6 CFLAGS+= -DPLATFORM_CONFIG_H=\"${.CURDIR}/config_freebsd.h\" CFLAGS+= -I${.OBJDIR} .if ${MK_OPENSSL} != "no" CFLAGS+= -DWITH_OPENSSL DPADD+= ${LIBCRYPTO} LDADD+= -lcrypto .else DPADD+= ${LIBMD} LDADD+= -lmd .endif .if ${MK_ICONV} != "no" # TODO: This can be changed back to CFLAGS once iconv works correctly # with statically linked binaries. SHARED_CFLAGS+= -DHAVE_ICONV=1 -DHAVE_ICONV_H=1 -DICONV_CONST= .endif .if ${MACHINE_ARCH:Marm*} != "" || ${MACHINE_ARCH:Mmips*} != "" || \ ${MACHINE_ARCH:Msparc64*} != "" || ${MACHINE_ARCH:Mia64*} != "" NO_WCAST_ALIGN= yes .if ${MACHINE_ARCH:M*64*} == "" CFLAGS+= -DPPMD_32BIT .endif .endif NO_WCAST_ALIGN.clang= .PATH: ${LIBARCHIVEDIR}/libarchive # Headers to be installed in /usr/include INCS= archive.h archive_entry.h # Sources to be compiled. SRCS= archive_acl.c \ archive_check_magic.c \ archive_cmdline.c \ archive_cryptor.c \ archive_disk_acl_freebsd.c \ archive_digest.c \ archive_entry.c \ archive_entry_copy_stat.c \ archive_entry_link_resolver.c \ archive_entry_sparse.c \ archive_entry_stat.c \ archive_entry_strmode.c \ archive_entry_xattr.c \ archive_getdate.c \ archive_hmac.c \ archive_match.c \ archive_options.c \ archive_pack_dev.c \ archive_pathmatch.c \ archive_ppmd7.c \ archive_random.c \ archive_rb.c \ archive_read.c \ archive_read_add_passphrase.c \ archive_read_append_filter.c \ archive_read_data_into_fd.c \ archive_read_disk_entry_from_file.c \ archive_read_disk_posix.c \ archive_read_disk_set_standard_lookup.c \ archive_read_extract.c \ archive_read_extract2.c \ archive_read_open_fd.c \ archive_read_open_file.c \ archive_read_open_filename.c \ archive_read_open_memory.c \ archive_read_set_format.c \ archive_read_set_options.c \ archive_read_support_filter_all.c \ archive_read_support_filter_bzip2.c \ archive_read_support_filter_compress.c \ archive_read_support_filter_gzip.c \ archive_read_support_filter_grzip.c \ archive_read_support_filter_lrzip.c \ archive_read_support_filter_lz4.c \ archive_read_support_filter_lzop.c \ archive_read_support_filter_none.c \ archive_read_support_filter_program.c \ archive_read_support_filter_rpm.c \ archive_read_support_filter_uu.c \ archive_read_support_filter_xz.c \ + archive_read_support_filter_zstd.c \ archive_read_support_format_7zip.c \ archive_read_support_format_all.c \ archive_read_support_format_ar.c \ archive_read_support_format_by_code.c \ archive_read_support_format_cab.c \ archive_read_support_format_cpio.c \ archive_read_support_format_empty.c \ archive_read_support_format_iso9660.c \ archive_read_support_format_lha.c \ archive_read_support_format_mtree.c \ archive_read_support_format_rar.c \ archive_read_support_format_raw.c \ archive_read_support_format_tar.c \ archive_read_support_format_warc.c \ archive_read_support_format_xar.c \ archive_read_support_format_zip.c \ archive_string.c \ archive_string_sprintf.c \ archive_util.c \ archive_version_details.c \ archive_virtual.c \ archive_write.c \ archive_write_add_filter.c \ archive_write_disk_set_standard_lookup.c \ archive_write_disk_posix.c \ archive_write_open_fd.c \ archive_write_open_file.c \ archive_write_open_filename.c \ archive_write_open_memory.c \ archive_write_add_filter_b64encode.c \ archive_write_add_filter_by_name.c \ archive_write_add_filter_bzip2.c \ archive_write_add_filter_compress.c \ archive_write_add_filter_grzip.c \ archive_write_add_filter_gzip.c \ archive_write_add_filter_lrzip.c \ archive_write_add_filter_lz4.c \ archive_write_add_filter_lzop.c \ archive_write_add_filter_none.c \ archive_write_add_filter_program.c \ archive_write_add_filter_uuencode.c \ archive_write_add_filter_xz.c \ + archive_write_add_filter_zstd.c \ archive_write_set_format.c \ archive_write_set_format_7zip.c \ archive_write_set_format_ar.c \ archive_write_set_format_by_name.c \ archive_write_set_format_cpio.c \ archive_write_set_format_cpio_newc.c \ archive_write_set_format_filter_by_ext.c \ archive_write_set_format_gnutar.c \ archive_write_set_format_iso9660.c \ archive_write_set_format_mtree.c \ archive_write_set_format_pax.c \ archive_write_set_format_raw.c \ archive_write_set_format_shar.c \ archive_write_set_format_ustar.c \ archive_write_set_format_v7tar.c \ archive_write_set_format_warc.c \ archive_write_set_format_xar.c \ archive_write_set_format_zip.c \ archive_write_set_passphrase.c \ archive_write_set_options.c \ filter_fork_posix.c # Man pages to be installed. MAN= archive_entry.3 \ archive_entry_acl.3 \ archive_entry_linkify.3 \ archive_entry_paths.3 \ archive_entry_perms.3 \ archive_entry_stat.3 \ archive_entry_time.3 \ archive_read.3 \ archive_read_data.3 \ archive_read_disk.3 \ archive_read_extract.3 \ archive_read_filter.3 \ archive_read_format.3 \ archive_read_free.3 \ archive_read_header.3 \ archive_read_new.3 \ archive_read_open.3 \ archive_read_set_options.3 \ archive_util.3 \ archive_write.3 \ archive_write_blocksize.3 \ archive_write_data.3 \ archive_write_disk.3 \ archive_write_filter.3 \ archive_write_finish_entry.3 \ archive_write_format.3 \ archive_write_free.3 \ archive_write_header.3 \ archive_write_new.3 \ archive_write_open.3 \ archive_write_set_options.3 \ cpio.5 \ libarchive.3 \ libarchive_changes.3 \ libarchive_internals.3 \ libarchive-formats.5 \ tar.5 # Symlink the man pages under each function name. MLINKS+= archive_entry.3 archive_entry_clear.3 MLINKS+= archive_entry.3 archive_entry_clone.3 MLINKS+= archive_entry.3 archive_entry_free.3 MLINKS+= archive_entry.3 archive_entry_new.3 MLINKS+= archive_entry_acl.3 archive_entry_acl_add_entry.3 MLINKS+= archive_entry_acl.3 archive_entry_acl_add_entry_w.3 MLINKS+= archive_entry_acl.3 archive_entry_acl_clear.3 MLINKS+= archive_entry_acl.3 archive_entry_acl_count.3 MLINKS+= archive_entry_acl.3 archive_entry_acl_next.3 MLINKS+= archive_entry_acl.3 archive_entry_acl_next_w.3 MLINKS+= archive_entry_acl.3 archive_entry_acl_reset.3 MLINKS+= archive_entry_acl.3 archive_entry_acl_text_w.3 MLINKS+= archive_entry_linkify.3 archive_entry_linkresolver.3 MLINKS+= archive_entry_linkify.3 archive_entry_linkresolver_new.3 MLINKS+= archive_entry_linkify.3 archive_entry_linkresolver_set_strategy.3 MLINKS+= archive_entry_linkify.3 archive_entry_linkresolver_free.3 MLINKS+= archive_entry_paths.3 archive_entry_copy_hardlink.3 MLINKS+= archive_entry_paths.3 archive_entry_copy_hardlink_w.3 MLINKS+= archive_entry_paths.3 archive_entry_copy_link.3 MLINKS+= archive_entry_paths.3 archive_entry_copy_link_w.3 MLINKS+= archive_entry_paths.3 archive_entry_copy_pathname.3 MLINKS+= archive_entry_paths.3 archive_entry_copy_pathname_w.3 MLINKS+= archive_entry_paths.3 archive_entry_copy_sourcepath.3 MLINKS+= archive_entry_paths.3 archive_entry_copy_symlink.3 MLINKS+= archive_entry_paths.3 archive_entry_copy_symlink_w.3 MLINKS+= archive_entry_paths.3 archive_entry_hardlink.3 MLINKS+= archive_entry_paths.3 archive_entry_hardlink_w.3 MLINKS+= archive_entry_paths.3 archive_entry_pathname.3 MLINKS+= archive_entry_paths.3 archive_entry_pathname_w.3 MLINKS+= archive_entry_paths.3 archive_entry_set_hardlink.3 MLINKS+= archive_entry_paths.3 archive_entry_set_link.3 MLINKS+= archive_entry_paths.3 archive_entry_set_pathname.3 MLINKS+= archive_entry_paths.3 archive_entry_set_symlink.3 MLINKS+= archive_entry_paths.3 archive_entry_symlink.3 MLINKS+= archive_entry_paths.3 archive_entry_symlink_w.3 MLINKS+= archive_entry_paths.3 archive_entry_update_symlink_utf8.3 MLINKS+= archive_entry_paths.3 archive_entry_update_hardlink_utf8.3 MLINKS+= archive_entry_perms.3 archive_entry_copy_fflags_text.3 MLINKS+= archive_entry_perms.3 archive_entry_copy_fflags_text_w.3 MLINKS+= archive_entry_perms.3 archive_entry_copy_gname.3 MLINKS+= archive_entry_perms.3 archive_entry_copy_gname_w.3 MLINKS+= archive_entry_perms.3 archive_entry_copy_uname.3 MLINKS+= archive_entry_perms.3 archive_entry_copy_uname_w.3 MLINKS+= archive_entry_perms.3 archive_entry_fflags.3 MLINKS+= archive_entry_perms.3 archive_entry_fflags_text.3 MLINKS+= archive_entry_perms.3 archive_entry_gid.3 MLINKS+= archive_entry_perms.3 archive_entry_gname.3 MLINKS+= archive_entry_perms.3 archive_entry_gname_w.3 MLINKS+= archive_entry_perms.3 archive_entry_set_fflags.3 MLINKS+= archive_entry_perms.3 archive_entry_set_gid.3 MLINKS+= archive_entry_perms.3 archive_entry_set_gname.3 MLINKS+= archive_entry_perms.3 archive_entry_perm.3 MLINKS+= archive_entry_perms.3 archive_entry_set_perm.3 MLINKS+= archive_entry_perms.3 archive_entry_set_uid.3 MLINKS+= archive_entry_perms.3 archive_entry_set_uname.3 MLINKS+= archive_entry_perms.3 archive_entry_strmode.3 MLINKS+= archive_entry_perms.3 archive_entry_uid.3 MLINKS+= archive_entry_perms.3 archive_entry_uname.3 MLINKS+= archive_entry_perms.3 archive_entry_uname_w.3 MLINKS+= archive_entry_perms.3 archive_entry_update_gname_utf8.3 MLINKS+= archive_entry_perms.3 archive_entry_update_uname_utf8.3 MLINKS+= archive_entry_stat.3 archive_entry_copy_stat.3 MLINKS+= archive_entry_stat.3 archive_entry_dev.3 MLINKS+= archive_entry_stat.3 archive_entry_dev_is_set.3 MLINKS+= archive_entry_stat.3 archive_entry_devmajor.3 MLINKS+= archive_entry_stat.3 archive_entry_devminor.3 MLINKS+= archive_entry_stat.3 archive_entry_filetype.3 MLINKS+= archive_entry_stat.3 archive_entry_ino.3 MLINKS+= archive_entry_stat.3 archive_entry_ino64.3 MLINKS+= archive_entry_stat.3 archive_entry_ino_is_set.3 MLINKS+= archive_entry_stat.3 archive_entry_mode.3 MLINKS+= archive_entry_stat.3 archive_entry_nlink.3 MLINKS+= archive_entry_stat.3 archive_entry_rdev.3 MLINKS+= archive_entry_stat.3 archive_entry_rdevmajor.3 MLINKS+= archive_entry_stat.3 archive_entry_rdevminor.3 MLINKS+= archive_entry_stat.3 archive_entry_set_dev.3 MLINKS+= archive_entry_stat.3 archive_entry_set_devmajor.3 MLINKS+= archive_entry_stat.3 archive_entry_set_devminor.3 MLINKS+= archive_entry_stat.3 archive_entry_set_filetype.3 MLINKS+= archive_entry_stat.3 archive_entry_set_ino.3 MLINKS+= archive_entry_stat.3 archive_entry_set_ino64.3 MLINKS+= archive_entry_stat.3 archive_entry_set_mode.3 MLINKS+= archive_entry_stat.3 archive_entry_set_nlink.3 MLINKS+= archive_entry_stat.3 archive_entry_set_rdev.3 MLINKS+= archive_entry_stat.3 archive_entry_set_rdevmajor.3 MLINKS+= archive_entry_stat.3 archive_entry_set_rdevminor.3 MLINKS+= archive_entry_stat.3 archive_entry_set_size.3 MLINKS+= archive_entry_stat.3 archive_entry_size.3 MLINKS+= archive_entry_stat.3 archive_entry_size_is_set.3 MLINKS+= archive_entry_stat.3 archive_entry_unset_size.3 MLINKS+= archive_entry_time.3 archive_entry_atime.3 MLINKS+= archive_entry_time.3 archive_entry_atime_is_set.3 MLINKS+= archive_entry_time.3 archive_entry_atime_nsec.3 MLINKS+= archive_entry_time.3 archive_entry_birthtime.3 MLINKS+= archive_entry_time.3 archive_entry_birthtime_is_set.3 MLINKS+= archive_entry_time.3 archive_entry_birthtime_nsec.3 MLINKS+= archive_entry_time.3 archive_entry_ctime.3 MLINKS+= archive_entry_time.3 archive_entry_ctime_is_set.3 MLINKS+= archive_entry_time.3 archive_entry_ctime_nsec.3 MLINKS+= archive_entry_time.3 archive_entry_mtime.3 MLINKS+= archive_entry_time.3 archive_entry_mtime_is_set.3 MLINKS+= archive_entry_time.3 archive_entry_mtime_nsec.3 MLINKS+= archive_entry_time.3 archive_entry_set_atime.3 MLINKS+= archive_entry_time.3 archive_entry_set_birthtime.3 MLINKS+= archive_entry_time.3 archive_entry_set_ctime.3 MLINKS+= archive_entry_time.3 archive_entry_set_mtime.3 MLINKS+= archive_entry_time.3 archive_entry_unset_atime.3 MLINKS+= archive_entry_time.3 archive_entry_unset_birthtime.3 MLINKS+= archive_entry_time.3 archive_entry_unset_ctime.3 MLINKS+= archive_entry_time.3 archive_entry_unset_mtime.3 MLINKS+= archive_read_data.3 archive_read_data_block.3 MLINKS+= archive_read_data.3 archive_read_data_into_fd.3 MLINKS+= archive_read_data.3 archive_read_data_skip.3 MLINKS+= archive_read_header.3 archive_read_next_header.3 MLINKS+= archive_read_header.3 archive_read_next_header2.3 MLINKS+= archive_read_extract.3 archive_read_extract2.3 MLINKS+= archive_read_extract.3 archive_read_extract_set_progress_callback.3 MLINKS+= archive_read_extract.3 archive_read_extract_set_skip_file.3 MLINKS+= archive_read_open.3 archive_read_open2.3 MLINKS+= archive_read_open.3 archive_read_open_FILE.3 MLINKS+= archive_read_open.3 archive_read_open_fd.3 MLINKS+= archive_read_open.3 archive_read_open_file.3 MLINKS+= archive_read_open.3 archive_read_open_filename.3 MLINKS+= archive_read_open.3 archive_read_open_memory.3 MLINKS+= archive_read_free.3 archive_read_close.3 MLINKS+= archive_read_free.3 archive_read_finish.3 MLINKS+= archive_read_filter.3 archive_read_support_filter_all.3 MLINKS+= archive_read_filter.3 archive_read_support_filter_bzip2.3 MLINKS+= archive_read_filter.3 archive_read_support_filter_compress.3 MLINKS+= archive_read_filter.3 archive_read_support_filter_gzip.3 MLINKS+= archive_read_filter.3 archive_read_support_filter_lzma.3 MLINKS+= archive_read_filter.3 archive_read_support_filter_none.3 MLINKS+= archive_read_filter.3 archive_read_support_filter_xz.3 MLINKS+= archive_read_filter.3 archive_read_support_filter_program.3 MLINKS+= archive_read_filter.3 archive_read_support_filter_program_signature.3 MLINKS+= archive_read_format.3 archive_read_support_format_7zip.3 MLINKS+= archive_read_format.3 archive_read_support_format_all.3 MLINKS+= archive_read_format.3 archive_read_support_format_ar.3 MLINKS+= archive_read_format.3 archive_read_support_format_by_code.3 MLINKS+= archive_read_format.3 archive_read_support_format_cab.3 MLINKS+= archive_read_format.3 archive_read_support_format_cpio.3 MLINKS+= archive_read_format.3 archive_read_support_format_empty.3 MLINKS+= archive_read_format.3 archive_read_support_format_iso9660.3 MLINKS+= archive_read_format.3 archive_read_support_format_lha.3 MLINKS+= archive_read_format.3 archive_read_support_format_mtree.3 MLINKS+= archive_read_format.3 archive_read_support_format_rar.3 MLINKS+= archive_read_format.3 archive_read_support_format_raw.3 MLINKS+= archive_read_format.3 archive_read_support_format_tar.3 MLINKS+= archive_read_format.3 archive_read_support_format_xar.3 MLINKS+= archive_read_format.3 archive_read_support_format_zip.3 MLINKS+= archive_read_disk.3 archive_read_disk_entry_from_file.3 MLINKS+= archive_read_disk.3 archive_read_disk_gname.3 MLINKS+= archive_read_disk.3 archive_read_disk_new.3 MLINKS+= archive_read_disk.3 archive_read_disk_set_gname_lookup.3 MLINKS+= archive_read_disk.3 archive_read_disk_set_standard_lookup.3 MLINKS+= archive_read_disk.3 archive_read_disk_set_symlink_hybrid.3 MLINKS+= archive_read_disk.3 archive_read_disk_set_symlink_logical.3 MLINKS+= archive_read_disk.3 archive_read_disk_set_symlink_physical.3 MLINKS+= archive_read_disk.3 archive_read_disk_set_uname_lookup.3 MLINKS+= archive_read_disk.3 archive_read_disk_uname.3 MLINKS+= archive_read_set_options.3 archive_read_set_filter_option.3 MLINKS+= archive_read_set_options.3 archive_read_set_format_option.3 MLINKS+= archive_read_set_options.3 archive_read_set_option.3 MLINKS+= archive_util.3 archive_clear_error.3 MLINKS+= archive_util.3 archive_compression.3 MLINKS+= archive_util.3 archive_compression_name.3 MLINKS+= archive_util.3 archive_copy_error.3 MLINKS+= archive_util.3 archive_errno.3 MLINKS+= archive_util.3 archive_error_string.3 MLINKS+= archive_util.3 archive_file_count.3 MLINKS+= archive_util.3 archive_filter_code.3 MLINKS+= archive_util.3 archive_filter_count.3 MLINKS+= archive_util.3 archive_filter_name.3 MLINKS+= archive_util.3 archive_format.3 MLINKS+= archive_util.3 archive_format_name.3 MLINKS+= archive_util.3 archive_position.3 MLINKS+= archive_util.3 archive_set_error.3 MLINKS+= archive_write_blocksize.3 archive_write_get_bytes_in_last_block.3 MLINKS+= archive_write_blocksize.3 archive_write_get_bytes_per_block.3 MLINKS+= archive_write_blocksize.3 archive_write_set_bytes_in_last_block.3 MLINKS+= archive_write_blocksize.3 archive_write_set_bytes_per_block.3 MLINKS+= archive_write_disk.3 archive_write_data_block.3 MLINKS+= archive_write_disk.3 archive_write_disk_new.3 MLINKS+= archive_write_disk.3 archive_write_disk_set_group_lookup.3 MLINKS+= archive_write_disk.3 archive_write_disk_set_options.3 MLINKS+= archive_write_disk.3 archive_write_disk_set_skip_file.3 MLINKS+= archive_write_disk.3 archive_write_disk_set_standard_lookup.3 MLINKS+= archive_write_disk.3 archive_write_disk_set_user_lookup.3 MLINKS+= archive_write_filter.3 archive_write_add_filter_bzip2.3 MLINKS+= archive_write_filter.3 archive_write_add_filter_compress.3 MLINKS+= archive_write_filter.3 archive_write_add_filter_gzip.3 MLINKS+= archive_write_filter.3 archive_write_add_filter_lzip.3 MLINKS+= archive_write_filter.3 archive_write_add_filter_lzma.3 MLINKS+= archive_write_filter.3 archive_write_add_filter_none.3 MLINKS+= archive_write_filter.3 archive_write_add_filter_program.3 MLINKS+= archive_write_filter.3 archive_write_add_filter_xz.3 MLINKS+= archive_write_format.3 archive_write_set_format_cpio.3 MLINKS+= archive_write_format.3 archive_write_set_format_pax.3 MLINKS+= archive_write_format.3 archive_write_set_format_pax_restricted.3 MLINKS+= archive_write_format.3 archive_write_set_format_shar.3 MLINKS+= archive_write_format.3 archive_write_set_format_shar_dump.3 MLINKS+= archive_write_format.3 archive_write_set_format_ustar.3 MLINKS+= archive_write_free.3 archive_write_close.3 MLINKS+= archive_write_free.3 archive_write_fail.3 MLINKS+= archive_write_free.3 archive_write_finish.3 MLINKS+= archive_write_open.3 archive_write_open_FILE.3 MLINKS+= archive_write_open.3 archive_write_open_fd.3 MLINKS+= archive_write_open.3 archive_write_open_file.3 MLINKS+= archive_write_open.3 archive_write_open_filename.3 MLINKS+= archive_write_open.3 archive_write_open_memory.3 MLINKS+= archive_write_set_options.3 archive_write_set_filter_option.3 MLINKS+= archive_write_set_options.3 archive_write_set_format_option.3 MLINKS+= archive_write_set_options.3 archive_write_set_option.3 MLINKS+= libarchive.3 archive.3 .if ${MK_TESTS} != "no" SUBDIR+= tests .endif .include Index: stable/10/lib/libarchive/tests/Makefile =================================================================== --- stable/10/lib/libarchive/tests/Makefile (revision 324417) +++ stable/10/lib/libarchive/tests/Makefile (revision 324418) @@ -1,571 +1,574 @@ # $FreeBSD$ SRCTOP= ${.CURDIR:H:H:H} TESTSDIR= ${TESTSBASE}/lib/libarchive LIBARCHIVEDIR= ${SRCTOP}/contrib/libarchive ATF_TESTS_SH+= functional_test BINDIR= ${TESTSDIR} PROGS+= libarchive_test CFLAGS+= -I${.CURDIR} -I${.CURDIR:H} -I${.OBJDIR} CFLAGS+= -I${LIBARCHIVEDIR}/libarchive -I${LIBARCHIVEDIR}/libarchive/test CFLAGS+= -I${LIBARCHIVEDIR}/test_utils CFLAGS+= -DHAVE_LIBLZMA=1 -DHAVE_LZMA_H=1 # Uncomment to link against dmalloc #LDADD+= -L/usr/local/lib -ldmalloc #CFLAGS+= -I/usr/local/include -DUSE_DMALLOC .PATH: ${LIBARCHIVEDIR}/libarchive/test TESTS_SRCS= \ test_acl_nfs4.c \ test_acl_pax.c \ test_acl_platform_nfs4.c \ test_acl_platform_posix1e.c \ test_acl_posix1e.c \ test_acl_text.c \ test_archive_api_feature.c \ test_archive_clear_error.c \ test_archive_cmdline.c \ test_archive_digest.c \ test_archive_getdate.c \ test_archive_match_time.c \ test_archive_match_owner.c \ test_archive_match_path.c \ test_archive_pathmatch.c \ test_archive_read_add_passphrase.c \ test_archive_read_close_twice.c \ test_archive_read_close_twice_open_fd.c \ test_archive_read_close_twice_open_filename.c \ test_archive_read_multiple_data_objects.c \ test_archive_read_next_header_empty.c \ test_archive_read_next_header_raw.c \ test_archive_read_open2.c \ test_archive_read_set_filter_option.c \ test_archive_read_set_format_option.c \ test_archive_read_set_option.c \ test_archive_read_set_options.c \ test_archive_read_support.c \ test_archive_set_error.c \ test_archive_string.c \ test_archive_string_conversion.c \ test_archive_write_add_filter_by_name.c \ test_archive_write_set_filter_option.c \ test_archive_write_set_format_by_name.c \ test_archive_write_set_format_filter_by_ext.c \ test_archive_write_set_format_option.c \ test_archive_write_set_option.c \ test_archive_write_set_options.c \ test_archive_write_set_passphrase.c \ test_bad_fd.c \ test_compat_bzip2.c \ test_compat_cpio.c \ test_compat_gtar.c \ test_compat_gzip.c \ test_compat_lz4.c \ test_compat_lzip.c \ test_compat_lzma.c \ test_compat_lzop.c \ test_compat_mac.c \ test_compat_pax_libarchive_2x.c \ test_compat_perl_archive_tar.c \ test_compat_plexus_archiver_tar.c \ test_compat_solaris_tar_acl.c \ test_compat_solaris_pax_sparse.c \ test_compat_star_acl.c \ test_compat_tar_hardlink.c \ test_compat_uudecode.c \ test_compat_uudecode_large.c \ test_compat_xz.c \ test_compat_zip.c \ + test_compat_zstd.c \ test_empty_write.c \ test_entry.c \ test_entry_strmode.c \ test_extattr_freebsd.c \ test_filter_count.c \ test_fuzz.c \ test_gnutar_filename_encoding.c \ test_link_resolver.c \ test_open_fd.c \ test_open_failure.c \ test_open_file.c \ test_open_filename.c \ test_pax_filename_encoding.c \ test_read_data_large.c \ test_read_disk.c \ test_read_disk_directory_traversals.c \ test_read_disk_entry_from_file.c \ test_read_extract.c \ test_read_file_nonexistent.c \ test_read_filter_compress.c \ test_read_filter_grzip.c \ test_read_filter_lrzip.c \ test_read_filter_lzop.c \ test_read_filter_lzop_multiple_parts.c \ test_read_filter_program.c \ test_read_filter_program_signature.c \ test_read_filter_uudecode.c \ test_read_format_7zip.c \ test_read_format_7zip_encryption_data.c \ test_read_format_7zip_encryption_header.c \ test_read_format_7zip_encryption_partially.c \ test_read_format_7zip_malformed.c \ test_read_format_ar.c \ test_read_format_cab.c \ test_read_format_cab_filename.c \ test_read_format_cpio_afio.c \ test_read_format_cpio_bin.c \ test_read_format_cpio_bin_Z.c \ test_read_format_cpio_bin_be.c \ test_read_format_cpio_bin_bz2.c \ test_read_format_cpio_bin_gz.c \ test_read_format_cpio_bin_le.c \ test_read_format_cpio_bin_lzip.c \ test_read_format_cpio_bin_lzma.c \ test_read_format_cpio_bin_xz.c \ test_read_format_cpio_filename.c \ test_read_format_cpio_odc.c \ test_read_format_cpio_svr4_gzip.c \ test_read_format_cpio_svr4c_Z.c \ test_read_format_cpio_svr4_bzip2_rpm.c \ test_read_format_cpio_svr4_gzip_rpm.c \ test_read_format_empty.c \ test_read_format_gtar_filename.c \ test_read_format_gtar_gz.c \ test_read_format_gtar_lzma.c \ test_read_format_gtar_sparse.c \ test_read_format_gtar_sparse_skip_entry.c \ test_read_format_iso_Z.c \ test_read_format_iso_multi_extent.c \ test_read_format_iso_xorriso.c \ test_read_format_isorr_rr_moved.c \ test_read_format_isojoliet_bz2.c \ test_read_format_isojoliet_long.c \ test_read_format_isojoliet_rr.c \ test_read_format_isojoliet_versioned.c \ test_read_format_isorr_bz2.c \ test_read_format_isorr_ce.c \ test_read_format_isorr_new_bz2.c \ test_read_format_isozisofs_bz2.c \ test_read_format_lha.c \ test_read_format_lha_bugfix_0.c \ test_read_format_lha_filename.c \ test_read_format_mtree.c \ test_read_format_mtree_crash747.c \ test_read_format_pax_bz2.c \ test_read_format_rar.c \ test_read_format_rar_encryption_data.c \ test_read_format_rar_encryption_header.c \ test_read_format_rar_encryption_partially.c \ test_read_format_rar_invalid1.c \ test_read_format_raw.c \ test_read_format_tar.c \ test_read_format_tar_concatenated.c \ test_read_format_tar_empty_filename.c \ test_read_format_tar_empty_pax.c \ test_read_format_tar_filename.c \ test_read_format_tbz.c \ test_read_format_tgz.c \ test_read_format_tlz.c \ test_read_format_txz.c \ test_read_format_tz.c \ test_read_format_ustar_filename.c \ test_read_format_warc.c \ test_read_format_xar.c \ test_read_format_zip.c \ test_read_format_zip_comment_stored.c \ test_read_format_zip_encryption_data.c \ test_read_format_zip_encryption_header.c \ test_read_format_zip_encryption_partially.c \ test_read_format_zip_filename.c \ test_read_format_zip_high_compression.c \ test_read_format_zip_jar.c \ test_read_format_zip_mac_metadata.c \ test_read_format_zip_malformed.c \ test_read_format_zip_msdos.c \ test_read_format_zip_nested.c \ test_read_format_zip_nofiletype.c \ test_read_format_zip_padded.c \ test_read_format_zip_sfx.c \ test_read_format_zip_traditional_encryption_data.c \ test_read_format_zip_winzip_aes.c \ test_read_format_zip_winzip_aes_large.c \ test_read_format_zip_with_invalid_traditional_eocd.c \ test_read_format_zip_zip64.c \ test_read_large.c \ test_read_pax_truncated.c \ test_read_position.c \ test_read_set_format.c \ test_read_too_many_filters.c \ test_read_truncated.c \ test_read_truncated_filter.c \ test_sparse_basic.c \ test_tar_filenames.c \ test_tar_large.c \ test_warn_missing_hardlink_target.c \ test_ustar_filenames.c \ test_ustar_filename_encoding.c \ test_write_disk.c \ test_write_disk_appledouble.c \ test_write_disk_failures.c \ test_write_disk_hardlink.c \ test_write_disk_hfs_compression.c \ test_write_disk_lookup.c \ test_write_disk_mac_metadata.c \ test_write_disk_no_hfs_compression.c \ test_write_disk_perms.c \ test_write_disk_secure.c \ test_write_disk_secure744.c \ test_write_disk_secure745.c \ test_write_disk_secure746.c \ test_write_disk_sparse.c \ test_write_disk_symlink.c \ test_write_disk_times.c \ test_write_filter_b64encode.c \ test_write_filter_bzip2.c \ test_write_filter_compress.c \ test_write_filter_gzip.c \ test_write_filter_gzip_timestamp.c \ test_write_filter_lrzip.c \ test_write_filter_lz4.c \ test_write_filter_lzip.c \ test_write_filter_lzma.c \ test_write_filter_lzop.c \ test_write_filter_program.c \ test_write_filter_uuencode.c \ test_write_filter_xz.c \ + test_write_filter_zstd.c \ test_write_format_7zip.c \ test_write_format_7zip_empty.c \ test_write_format_7zip_large.c \ test_write_format_ar.c \ test_write_format_cpio.c \ test_write_format_cpio_empty.c \ test_write_format_cpio_newc.c \ test_write_format_cpio_odc.c \ test_write_format_gnutar.c \ test_write_format_gnutar_filenames.c \ test_write_format_iso9660.c \ test_write_format_iso9660_boot.c \ test_write_format_iso9660_empty.c \ test_write_format_iso9660_filename.c \ test_write_format_iso9660_zisofs.c \ test_write_format_mtree.c \ test_write_format_mtree_absolute_path.c \ test_write_format_mtree_classic.c \ test_write_format_mtree_classic_indent.c \ test_write_format_mtree_fflags.c \ test_write_format_mtree_no_separator.c \ test_write_format_mtree_quoted_filename.c \ test_write_format_pax.c \ test_write_format_raw.c \ test_write_format_raw_b64.c \ test_write_format_shar_empty.c \ test_write_format_tar.c \ test_write_format_tar_empty.c \ test_write_format_tar_sparse.c \ test_write_format_tar_ustar.c \ test_write_format_tar_v7tar.c \ test_write_format_warc.c \ test_write_format_warc_empty.c \ test_write_format_xar.c \ test_write_format_xar_empty.c \ test_write_format_zip.c \ test_write_format_zip_compression_store.c \ test_write_format_zip_empty.c \ test_write_format_zip_empty_zip64.c \ test_write_format_zip_file.c \ test_write_format_zip_file_zip64.c \ test_write_format_zip_large.c \ test_write_format_zip_zip64.c \ test_write_open_memory.c \ test_write_read_format_zip.c \ test_xattr_platform.c \ test_zip_filename_encoding.c # Deterministic failures: # Crashes with SIGBUS BROKEN_TESTS+= test_archive_rmd160 # Fails with `libarchive/test/test_archive_crypto.c:121: md != actualmd` BROKEN_TESTS+= test_archive_sha384 # Fails with `test_compat_pax_libarchive_2x.c:122: ARCHIVE_WARN != archive_read_next_header(a, &ae)` BROKEN_TESTS+= test_compat_pax_libarchive_2x # Fails with `test_read_disk_directory_traversals.c:1094: File at has atime 886622, 1443306049 seconds ago` BROKEN_TESTS+= test_read_disk_directory_traversals # Non-deterministic failures: # (Times out?) [and] crashes BROKEN_TESTS+= test_fuzz_rar # Build the test program. SRCS.libarchive_test= \ ${TESTS_SRCS} \ read_open_memory.c \ list.h DPADD.libarchive_test= ${LIBARCHIVE} LDADD.libarchive_test= -larchive .PATH: ${LIBARCHIVEDIR}/test_utils SRCS.libarchive_test+= test_main.c \ test_utils.c # list.h is just a list of all tests, as indicated by DEFINE_TEST macro lines list.h: ${TESTS_SRCS} Makefile @(cd ${LIBARCHIVEDIR}/libarchive/test && \ grep -E -h ^DEFINE_TEST ${.ALLSRC:N*Makefile} | \ egrep -v '${BROKEN_TESTS:tW:C/ /|/g}') > ${.TARGET}.tmp @mv ${.TARGET}.tmp ${.TARGET} CLEANFILES+= list.h list.h.tmp FILES+= README FILES+= test_acl_pax_posix1e.tar.uu FILES+= test_acl_pax_nfs4.tar.uu FILES+= test_archive_string_conversion.txt.Z.uu FILES+= test_compat_bzip2_1.tbz.uu FILES+= test_compat_bzip2_2.tbz.uu FILES+= test_compat_cpio_1.cpio.uu FILES+= test_compat_gtar_1.tar.uu FILES+= test_compat_gtar_2.tar.uu FILES+= test_compat_gzip_1.tgz.uu FILES+= test_compat_gzip_2.tgz.uu FILES+= test_compat_lz4_1.tar.lz4.uu FILES+= test_compat_lz4_2.tar.lz4.uu FILES+= test_compat_lz4_3.tar.lz4.uu FILES+= test_compat_lz4_B4.tar.lz4.uu FILES+= test_compat_lz4_B4BD.tar.lz4.uu FILES+= test_compat_lz4_B4BDBX.tar.lz4.uu FILES+= test_compat_lz4_B5.tar.lz4.uu FILES+= test_compat_lz4_B5BD.tar.lz4.uu FILES+= test_compat_lz4_B6.tar.lz4.uu FILES+= test_compat_lz4_B6BD.tar.lz4.uu FILES+= test_compat_lz4_B7.tar.lz4.uu FILES+= test_compat_lz4_B7BD.tar.lz4.uu FILES+= test_compat_lzip_1.tlz.uu FILES+= test_compat_lzip_2.tlz.uu FILES+= test_compat_lzma_1.tlz.uu FILES+= test_compat_lzma_2.tlz.uu FILES+= test_compat_lzma_3.tlz.uu FILES+= test_compat_lzop_1.tar.lzo.uu FILES+= test_compat_lzop_2.tar.lzo.uu FILES+= test_compat_lzop_3.tar.lzo.uu FILES+= test_compat_mac-1.tar.Z.uu FILES+= test_compat_mac-2.tar.Z.uu FILES+= test_compat_pax_libarchive_2x.tar.Z.uu FILES+= test_compat_perl_archive_tar.tar.uu FILES+= test_compat_plexus_archiver_tar.tar.uu FILES+= test_compat_solaris_pax_sparse_1.pax.Z.uu FILES+= test_compat_solaris_pax_sparse_2.pax.Z.uu FILES+= test_compat_solaris_tar_acl.tar.uu FILES+= test_compat_star_acl_nfs4.tar.uu FILES+= test_compat_star_acl_posix1e.tar.uu FILES+= test_compat_tar_hardlink_1.tar.uu FILES+= test_compat_uudecode_large.tar.Z.uu FILES+= test_compat_xz_1.txz.uu FILES+= test_compat_zip_1.zip.uu FILES+= test_compat_zip_2.zip.uu FILES+= test_compat_zip_3.zip.uu FILES+= test_compat_zip_4.zip.uu FILES+= test_compat_zip_5.zip.uu FILES+= test_compat_zip_6.zip.uu FILES+= test_compat_zip_7.xps.uu +FILES+= test_compat_zstd_1.tar.zst.uu FILES+= test_fuzz.cab.uu FILES+= test_fuzz.lzh.uu FILES+= test_fuzz_1.iso.Z.uu FILES+= test_pax_filename_encoding.tar.uu FILES+= test_rar_multivolume_multiple_files.part1.rar.uu FILES+= test_rar_multivolume_multiple_files.part2.rar.uu FILES+= test_rar_multivolume_multiple_files.part3.rar.uu FILES+= test_rar_multivolume_multiple_files.part4.rar.uu FILES+= test_rar_multivolume_multiple_files.part5.rar.uu FILES+= test_rar_multivolume_multiple_files.part6.rar.uu FILES+= test_rar_multivolume_single_file.part1.rar.uu FILES+= test_rar_multivolume_single_file.part2.rar.uu FILES+= test_rar_multivolume_single_file.part3.rar.uu FILES+= test_rar_multivolume_uncompressed_files.part01.rar.uu FILES+= test_rar_multivolume_uncompressed_files.part02.rar.uu FILES+= test_rar_multivolume_uncompressed_files.part03.rar.uu FILES+= test_rar_multivolume_uncompressed_files.part04.rar.uu FILES+= test_rar_multivolume_uncompressed_files.part05.rar.uu FILES+= test_rar_multivolume_uncompressed_files.part06.rar.uu FILES+= test_rar_multivolume_uncompressed_files.part07.rar.uu FILES+= test_rar_multivolume_uncompressed_files.part08.rar.uu FILES+= test_rar_multivolume_uncompressed_files.part09.rar.uu FILES+= test_rar_multivolume_uncompressed_files.part10.rar.uu FILES+= test_read_filter_grzip.tar.grz.uu FILES+= test_read_filter_lrzip.tar.lrz.uu FILES+= test_read_filter_lzop.tar.lzo.uu FILES+= test_read_filter_lzop_multiple_parts.tar.lzo.uu FILES+= test_read_format_7zip_bcj2_bzip2.7z.uu FILES+= test_read_format_7zip_bcj2_copy_1.7z.uu FILES+= test_read_format_7zip_bcj2_copy_2.7z.uu FILES+= test_read_format_7zip_bcj2_copy_lzma.7z.uu FILES+= test_read_format_7zip_bcj2_deflate.7z.uu FILES+= test_read_format_7zip_bcj2_lzma1_1.7z.uu FILES+= test_read_format_7zip_bcj2_lzma1_2.7z.uu FILES+= test_read_format_7zip_bcj2_lzma2_1.7z.uu FILES+= test_read_format_7zip_bcj2_lzma2_2.7z.uu FILES+= test_read_format_7zip_bcj_bzip2.7z.uu FILES+= test_read_format_7zip_bcj_copy.7z.uu FILES+= test_read_format_7zip_bcj_deflate.7z.uu FILES+= test_read_format_7zip_bcj_lzma1.7z.uu FILES+= test_read_format_7zip_bcj_lzma2.7z.uu FILES+= test_read_format_7zip_bzip2.7z.uu FILES+= test_read_format_7zip_copy.7z.uu FILES+= test_read_format_7zip_copy_2.7z.uu FILES+= test_read_format_7zip_deflate.7z.uu FILES+= test_read_format_7zip_delta_lzma1.7z.uu FILES+= test_read_format_7zip_delta_lzma2.7z.uu FILES+= test_read_format_7zip_empty_archive.7z.uu FILES+= test_read_format_7zip_empty_file.7z.uu FILES+= test_read_format_7zip_encryption.7z.uu FILES+= test_read_format_7zip_encryption_header.7z.uu FILES+= test_read_format_7zip_encryption_partially.7z.uu FILES+= test_read_format_7zip_lzma1.7z.uu FILES+= test_read_format_7zip_lzma1_2.7z.uu FILES+= test_read_format_7zip_lzma1_lzma2.7z.uu FILES+= test_read_format_7zip_lzma2.7z.uu FILES+= test_read_format_7zip_malformed.7z.uu FILES+= test_read_format_7zip_malformed2.7z.uu FILES+= test_read_format_7zip_ppmd.7z.uu FILES+= test_read_format_7zip_symbolic_name.7z.uu FILES+= test_read_format_ar.ar.uu FILES+= test_read_format_cab_1.cab.uu FILES+= test_read_format_cab_2.cab.uu FILES+= test_read_format_cab_3.cab.uu FILES+= test_read_format_cab_filename_cp932.cab.uu FILES+= test_read_format_cpio_bin_be.cpio.uu FILES+= test_read_format_cpio_bin_le.cpio.uu FILES+= test_read_format_cpio_filename_cp866.cpio.uu FILES+= test_read_format_cpio_filename_eucjp.cpio.uu FILES+= test_read_format_cpio_filename_koi8r.cpio.uu FILES+= test_read_format_cpio_filename_utf8_jp.cpio.uu FILES+= test_read_format_cpio_filename_utf8_ru.cpio.uu FILES+= test_read_format_cpio_svr4_bzip2_rpm.rpm.uu FILES+= test_read_format_cpio_svr4_gzip_rpm.rpm.uu FILES+= test_read_format_gtar_filename_cp866.tar.Z.uu FILES+= test_read_format_gtar_filename_eucjp.tar.Z.uu FILES+= test_read_format_gtar_filename_koi8r.tar.Z.uu FILES+= test_read_format_gtar_sparse_1_13.tar.uu FILES+= test_read_format_gtar_sparse_1_17.tar.uu FILES+= test_read_format_gtar_sparse_1_17_posix00.tar.uu FILES+= test_read_format_gtar_sparse_1_17_posix01.tar.uu FILES+= test_read_format_gtar_sparse_1_17_posix10.tar.uu FILES+= test_read_format_gtar_sparse_1_17_posix10_modified.tar.uu FILES+= test_read_format_gtar_sparse_skip_entry.tar.Z.uu FILES+= test_read_format_iso.iso.Z.uu FILES+= test_read_format_iso_2.iso.Z.uu FILES+= test_read_format_iso_joliet.iso.Z.uu FILES+= test_read_format_iso_joliet_by_nero.iso.Z.uu FILES+= test_read_format_iso_joliet_long.iso.Z.uu FILES+= test_read_format_iso_joliet_rockridge.iso.Z.uu FILES+= test_read_format_iso_multi_extent.iso.Z.uu FILES+= test_read_format_iso_rockridge.iso.Z.uu FILES+= test_read_format_iso_rockridge_ce.iso.Z.uu FILES+= test_read_format_iso_rockridge_new.iso.Z.uu FILES+= test_read_format_iso_rockridge_rr_moved.iso.Z.uu FILES+= test_read_format_iso_xorriso.iso.Z.uu FILES+= test_read_format_iso_zisofs.iso.Z.uu FILES+= test_read_format_lha_bugfix_0.lzh.uu FILES+= test_read_format_lha_filename_cp932.lzh.uu FILES+= test_read_format_lha_header0.lzh.uu FILES+= test_read_format_lha_header1.lzh.uu FILES+= test_read_format_lha_header2.lzh.uu FILES+= test_read_format_lha_header3.lzh.uu FILES+= test_read_format_lha_lh0.lzh.uu FILES+= test_read_format_lha_lh6.lzh.uu FILES+= test_read_format_lha_lh7.lzh.uu FILES+= test_read_format_lha_withjunk.lzh.uu FILES+= test_read_format_mtree.mtree.uu FILES+= test_read_format_mtree_crash747.mtree.bz2.uu FILES+= test_read_format_mtree_nomagic.mtree.uu FILES+= test_read_format_mtree_nomagic2.mtree.uu FILES+= test_read_format_mtree_nomagic3.mtree.uu FILES+= test_read_format_rar.rar.uu FILES+= test_read_format_rar_binary_data.rar.uu FILES+= test_read_format_rar_compress_best.rar.uu FILES+= test_read_format_rar_compress_normal.rar.uu FILES+= test_read_format_rar_encryption_data.rar.uu FILES+= test_read_format_rar_encryption_header.rar.uu FILES+= test_read_format_rar_encryption_partially.rar.uu FILES+= test_read_format_rar_invalid1.rar.uu FILES+= test_read_format_rar_multi_lzss_blocks.rar.uu FILES+= test_read_format_rar_multivolume.part0001.rar.uu FILES+= test_read_format_rar_multivolume.part0002.rar.uu FILES+= test_read_format_rar_multivolume.part0003.rar.uu FILES+= test_read_format_rar_multivolume.part0004.rar.uu FILES+= test_read_format_rar_noeof.rar.uu FILES+= test_read_format_rar_ppmd_lzss_conversion.rar.uu FILES+= test_read_format_rar_sfx.exe.uu FILES+= test_read_format_rar_subblock.rar.uu FILES+= test_read_format_rar_unicode.rar.uu FILES+= test_read_format_rar_windows.rar.uu FILES+= test_read_format_raw.bufr.uu FILES+= test_read_format_raw.data.Z.uu FILES+= test_read_format_raw.data.uu FILES+= test_read_format_tar_concatenated.tar.uu FILES+= test_read_format_tar_empty_filename.tar.uu FILES+= test_read_format_tar_empty_pax.tar.Z.uu FILES+= test_read_format_tar_filename_koi8r.tar.Z.uu FILES+= test_read_format_ustar_filename_cp866.tar.Z.uu FILES+= test_read_format_ustar_filename_eucjp.tar.Z.uu FILES+= test_read_format_ustar_filename_koi8r.tar.Z.uu FILES+= test_read_format_warc.warc.uu FILES+= test_read_format_zip.zip.uu FILES+= test_read_format_zip_comment_stored_1.zip.uu FILES+= test_read_format_zip_comment_stored_2.zip.uu FILES+= test_read_format_zip_encryption_data.zip.uu FILES+= test_read_format_zip_encryption_header.zip.uu FILES+= test_read_format_zip_encryption_partially.zip.uu FILES+= test_read_format_zip_filename_cp866.zip.uu FILES+= test_read_format_zip_filename_cp932.zip.uu FILES+= test_read_format_zip_filename_koi8r.zip.uu FILES+= test_read_format_zip_filename_utf8_jp.zip.uu FILES+= test_read_format_zip_filename_utf8_ru.zip.uu FILES+= test_read_format_zip_filename_utf8_ru2.zip.uu FILES+= test_read_format_zip_high_compression.zip.uu FILES+= test_read_format_zip_jar.jar.uu FILES+= test_read_format_zip_length_at_end.zip.uu FILES+= test_read_format_zip_mac_metadata.zip.uu FILES+= test_read_format_zip_malformed1.zip.uu FILES+= test_read_format_zip_msdos.zip.uu FILES+= test_read_format_zip_nested.zip.uu FILES+= test_read_format_zip_nofiletype.zip.uu FILES+= test_read_format_zip_padded1.zip.uu FILES+= test_read_format_zip_padded2.zip.uu FILES+= test_read_format_zip_padded3.zip.uu FILES+= test_read_format_zip_sfx.uu FILES+= test_read_format_zip_symlink.zip.uu FILES+= test_read_format_zip_traditional_encryption_data.zip.uu FILES+= test_read_format_zip_ux.zip.uu FILES+= test_read_format_zip_with_invalid_traditional_eocd.zip.uu FILES+= test_read_format_zip_winzip_aes128.zip.uu FILES+= test_read_format_zip_winzip_aes256.zip.uu FILES+= test_read_format_zip_winzip_aes256_large.zip.uu FILES+= test_read_format_zip_winzip_aes256_stored.zip.uu FILES+= test_read_format_zip_zip64a.zip.uu FILES+= test_read_format_zip_zip64b.zip.uu FILES+= test_read_large_splitted_rar_aa.uu FILES+= test_read_large_splitted_rar_ab.uu FILES+= test_read_large_splitted_rar_ac.uu FILES+= test_read_large_splitted_rar_ad.uu FILES+= test_read_large_splitted_rar_ae.uu FILES+= test_read_pax_schily_xattr.tar.uu FILES+= test_read_splitted_rar_aa.uu FILES+= test_read_splitted_rar_ab.uu FILES+= test_read_splitted_rar_ac.uu FILES+= test_read_splitted_rar_ad.uu FILES+= test_read_too_many_filters.gz.uu FILES+= test_splitted_rar_seek_support_aa.uu FILES+= test_splitted_rar_seek_support_ab.uu FILES+= test_splitted_rar_seek_support_ac.uu FILES+= test_write_disk_appledouble.cpio.gz.uu FILES+= test_write_disk_hfs_compression.tgz.uu FILES+= test_write_disk_mac_metadata.tar.gz.uu FILES+= test_write_disk_no_hfs_compression.tgz.uu .include Index: stable/10/usr.bin/bsdcat/tests/Makefile =================================================================== --- stable/10/usr.bin/bsdcat/tests/Makefile (revision 324417) +++ stable/10/usr.bin/bsdcat/tests/Makefile (revision 324418) @@ -1,69 +1,72 @@ # $FreeBSD$ PACKAGE= tests _LIBARCHIVEDIR= ${SRCTOP}/contrib/libarchive ATF_TESTS_SH+= functional_test BINDIR= ${TESTSDIR} PROGS+= bsdcat_test CFLAGS+= -DPLATFORM_CONFIG_H=\"${SRCTOP}/lib/libarchive/config_freebsd.h\" CFLAGS+= -I${SRCTOP}/lib/libarchive -I${.OBJDIR} CFLAGS+= -I${.OBJDIR} CFLAGS+= -I${_LIBARCHIVEDIR}/cat -I${_LIBARCHIVEDIR}/cat/test CFLAGS+= -I${_LIBARCHIVEDIR}/libarchive CFLAGS+= -I${_LIBARCHIVEDIR}/libarchive_fe -I${_LIBARCHIVEDIR}/test_utils # Uncomment to link against dmalloc #LDADD+= -L/usr/local/lib -ldmalloc #CFLAGS+= -I/usr/local/include -DUSE_DMALLOC .PATH: ${_LIBARCHIVEDIR}/cat/test TESTS_SRCS= \ test_0.c \ test_empty_gz.c \ test_empty_lz4.c \ test_empty_xz.c \ + test_empty_zstd.c \ test_error.c \ test_error_mixed.c \ test_expand_Z.c \ test_expand_bz2.c \ test_expand_gz.c \ test_expand_lz4.c \ test_expand_mixed.c \ test_expand_plain.c \ test_expand_xz.c \ + test_expand_zstd.c \ test_help.c \ test_version.c SRCS.bsdcat_test= list.h \ ${TESTS_SRCS} .PATH: ${_LIBARCHIVEDIR}/test_utils SRCS.bsdcat_test+= test_main.c \ test_utils.c LIBADD.bsdcat_test= archive list.h: ${TESTS_SRCS} Makefile @(cd ${_LIBARCHIVEDIR}/tar/test && \ grep -h DEFINE_TEST ${.ALLSRC:N*Makefile}) > ${.TARGET}.tmp @mv ${.TARGET}.tmp ${.TARGET} CLEANFILES+= list.h list.h.tmp FILES+= test_empty.gz.uu FILES+= test_empty.lz4.uu FILES+= test_empty.xz.uu FILES+= test_expand.Z.uu FILES+= test_expand.bz2.uu FILES+= test_expand.gz.uu FILES+= test_expand.lz4.uu FILES+= test_expand.plain.uu FILES+= test_expand.xz.uu +FILES+= test_expand.zst.uu .include Index: stable/10/usr.bin/cpio/tests/Makefile =================================================================== --- stable/10/usr.bin/cpio/tests/Makefile (revision 324417) +++ stable/10/usr.bin/cpio/tests/Makefile (revision 324418) @@ -1,124 +1,127 @@ # $FreeBSD$ SRCTOP= ${.CURDIR:H:H:H} TESTSDIR= ${TESTSBASE}/usr.bin/cpio LIBARCHIVEDIR= ${SRCTOP}/contrib/libarchive ATF_TESTS_SH+= functional_test BINDIR= ${TESTSDIR} PROGS+= bsdcpio_test CFLAGS+= -DPLATFORM_CONFIG_H=\"${SRCTOP}/lib/libarchive/config_freebsd.h\" CFLAGS+= -I${SRCTOP}/lib/libarchive -I${.OBJDIR} CFLAGS+= -I${.OBJDIR} CFLAGS+= -I${LIBARCHIVEDIR}/cpio -I${LIBARCHIVEDIR}/cpio/test CFLAGS+= -I${LIBARCHIVEDIR}/libarchive CFLAGS+= -I${LIBARCHIVEDIR}/libarchive_fe -I${LIBARCHIVEDIR}/test_utils # Uncomment to link against dmalloc #LDADD+= -L/usr/local/lib -ldmalloc #CFLAGS+= -I/usr/local/include -DUSE_DMALLOC .PATH: ${LIBARCHIVEDIR}/cpio CPIO_SRCS= cmdline.c .PATH: ${LIBARCHIVEDIR}/libarchive_fe CPIO_SRCS+= err.c .PATH: ${LIBARCHIVEDIR}/cpio/test TESTS_SRCS= \ test_0.c \ test_basic.c \ test_cmdline.c \ test_extract_cpio_Z.c \ test_extract_cpio_bz2.c \ test_extract_cpio_grz.c \ test_extract_cpio_gz.c \ test_extract_cpio_lrz.c \ test_extract_cpio_lz.c \ test_extract_cpio_lz4.c \ test_extract_cpio_lzma.c \ test_extract_cpio_lzo.c \ test_extract_cpio_xz.c \ + test_extract_cpio_zstd.c \ test_format_newc.c \ test_gcpio_compat.c \ test_missing_file.c \ test_option_0.c \ test_option_B_upper.c \ test_option_C_upper.c \ test_option_J_upper.c \ test_option_L_upper.c \ test_option_Z_upper.c \ test_option_a.c \ test_option_b64encode.c \ test_option_c.c \ test_option_d.c \ test_option_f.c \ test_option_grzip.c \ test_option_help.c \ test_option_l.c \ test_option_lrzip.c \ test_option_lz4.c \ test_option_lzma.c \ test_option_lzop.c \ test_option_m.c \ test_option_passphrase.c \ test_option_t.c \ test_option_u.c \ test_option_uuencode.c \ test_option_version.c \ test_option_xz.c \ test_option_y.c \ test_option_z.c \ + test_option_zstd.c \ test_owner_parse.c \ test_passthrough_dotdot.c \ test_passthrough_reverse.c SRCS.bsdcpio_test= list.h \ ${CPIO_SRCS} \ ${TESTS_SRCS} .PATH: ${LIBARCHIVEDIR}/test_utils SRCS.bsdcpio_test+= test_main.c \ test_utils.c DPADD.bsdcpio_test= ${LIBARCHIVE} LDADD.bsdcpio_test= -larchive list.h: ${TESTS_SRCS} Makefile @(cd ${LIBARCHIVEDIR}/tar/test && \ grep -h DEFINE_TEST ${.ALLSRC:N*Makefile}) > ${.TARGET}.tmp @mv ${.TARGET}.tmp ${.TARGET} CLEANFILES+= list.h list.h.tmp FILES+= test_extract.cpio.Z.uu FILES+= test_extract.cpio.bz2.uu FILES+= test_extract.cpio.grz.uu FILES+= test_extract.cpio.gz.uu FILES+= test_extract.cpio.lrz.uu FILES+= test_extract.cpio.lz.uu FILES+= test_extract.cpio.lz4.uu FILES+= test_extract.cpio.lzma.uu FILES+= test_extract.cpio.lzo.uu FILES+= test_extract.cpio.xz.uu +FILES+= test_extract.cpio.zst.uu FILES+= test_gcpio_compat_ref.bin.uu FILES+= test_gcpio_compat_ref.crc.uu FILES+= test_gcpio_compat_ref.newc.uu FILES+= test_gcpio_compat_ref.ustar.uu FILES+= test_gcpio_compat_ref_nosym.bin.uu FILES+= test_gcpio_compat_ref_nosym.crc.uu FILES+= test_gcpio_compat_ref_nosym.newc.uu FILES+= test_gcpio_compat_ref_nosym.ustar.uu FILES+= test_option_f.cpio.uu FILES+= test_option_m.cpio.uu FILES+= test_option_passphrase.zip.uu FILES+= test_option_t.cpio.uu FILES+= test_option_t.stdout.uu FILES+= test_option_tv.stdout.uu .include Index: stable/10/usr.bin/tar/tests/Makefile =================================================================== --- stable/10/usr.bin/tar/tests/Makefile (revision 324417) +++ stable/10/usr.bin/tar/tests/Makefile (revision 324418) @@ -1,122 +1,125 @@ # $FreeBSD$ SRCTOP= ${.CURDIR:H:H:H} TESTSDIR= ${TESTSBASE}/usr.bin/tar LIBARCHIVEDIR= ${SRCTOP}/contrib/libarchive ATF_TESTS_SH+= functional_test BINDIR= ${TESTSDIR} CFLAGS+= -DPLATFORM_CONFIG_H=\"${SRCTOP}/lib/libarchive/config_freebsd.h\" CFLAGS+= -I${SRCTOP}/lib/libarchive -I${.OBJDIR} CFLAGS+= -I${LIBARCHIVEDIR}/tar -I${LIBARCHIVEDIR}/tar/test CFLAGS+= -I${LIBARCHIVEDIR}/libarchive CFLAGS+= -I${LIBARCHIVEDIR}/test_utils # Uncomment to link against dmalloc #LDADD+= -L/usr/local/lib -ldmalloc #CFLAGS+= -I/usr/local/include -DUSE_DMALLOC PROGS+= bsdtar_test .PATH: ${LIBARCHIVEDIR}/tar/test TESTS_SRCS= \ test_0.c \ test_basic.c \ test_copy.c \ test_empty_mtree.c \ test_extract_tar_bz2.c \ test_extract_tar_grz.c \ test_extract_tar_gz.c \ test_extract_tar_lrz.c \ test_extract_tar_lz.c \ test_extract_tar_lz4.c \ test_extract_tar_lzma.c \ test_extract_tar_lzo.c \ test_extract_tar_xz.c \ + test_extract_tar_zstd.c \ test_format_newc.c \ test_help.c \ test_leading_slash.c \ test_missing_file.c \ test_option_C_upper.c \ test_option_H_upper.c \ test_option_L_upper.c \ test_option_O_upper.c \ test_option_T_upper.c \ test_option_U_upper.c \ test_option_X_upper.c \ test_option_a.c \ test_option_acls.c \ test_option_b.c \ test_option_b64encode.c \ test_option_exclude.c \ test_option_fflags.c \ test_option_gid_gname.c \ test_option_grzip.c \ test_option_j.c \ test_option_k.c \ test_option_keep_newer_files.c \ test_option_lrzip.c \ test_option_lz4.c \ test_option_lzma.c \ test_option_lzop.c \ test_option_n.c \ test_option_newer_than.c \ test_option_nodump.c \ test_option_older_than.c \ test_option_passphrase.c \ test_option_q.c \ test_option_r.c \ test_option_s.c \ test_option_uid_uname.c \ test_option_uuencode.c \ test_option_xattrs.c \ test_option_xz.c \ test_option_z.c \ + test_option_zstd.c \ test_patterns.c \ test_print_longpath.c \ test_stdio.c \ test_strip_components.c \ test_symlink_dir.c \ test_version.c SRCS.bsdtar_test= \ ${TESTS_SRCS} \ list.h .PATH: ${LIBARCHIVEDIR}/test_utils SRCS.bsdtar_test+= test_main.c \ test_utils.c DPADD.bsdtar_test= ${LIBARCHIVE} LDADD.bsdtar_test= -larchive list.h: ${TESTS_SRCS} Makefile @(cd ${LIBARCHIVEDIR}/tar/test && \ grep -h DEFINE_TEST ${.ALLSRC:N*Makefile}) > ${.TARGET}.tmp @mv ${.TARGET}.tmp ${.TARGET} CLEANFILES+= list.h list.h.tmp FILES+= test_extract.tar.Z.uu FILES+= test_extract.tar.bz2.uu FILES+= test_extract.tar.grz.uu FILES+= test_extract.tar.gz.uu FILES+= test_extract.tar.lrz.uu FILES+= test_extract.tar.lz.uu FILES+= test_extract.tar.lz4.uu FILES+= test_extract.tar.lzma.uu FILES+= test_extract.tar.lzo.uu FILES+= test_extract.tar.xz.uu +FILES+= test_extract.tar.zst.uu FILES+= test_leading_slash.tar.uu FILES+= test_option_keep_newer_files.tar.Z.uu FILES+= test_option_passphrase.zip.uu FILES+= test_option_s.tar.Z.uu FILES+= test_patterns_2.tar.uu FILES+= test_patterns_3.tar.uu FILES+= test_patterns_4.tar.uu FILES+= test_print_longpath.tar.Z.uu .include Index: stable/10 =================================================================== --- stable/10 (revision 324417) +++ stable/10 (revision 324418) Property changes on: stable/10 ___________________________________________________________________ Modified: svn:mergeinfo ## -0,0 +0,1 ## Merged /head:r324148