diff --git a/contrib/gnu-sort/ABOUT-NLS b/contrib/gnu-sort/ABOUT-NLS index f0833956378d..57e4b6760cfb 100644 --- a/contrib/gnu-sort/ABOUT-NLS +++ b/contrib/gnu-sort/ABOUT-NLS @@ -1,393 +1,432 @@ Notes on the Free Translation Project ************************************* Free software is going international! The Free Translation Project is a way to get maintainers of free software, translators, and users all together, so that will gradually become able to speak many languages. A few packages already provide translations for their messages. If you found this `ABOUT-NLS' file inside a distribution, you may assume that the distributed package does use GNU `gettext' internally, itself available at your nearest GNU archive site. But you do _not_ need to install GNU `gettext' prior to configuring, installing or using this package with messages translated. Installers will find here some useful hints. These notes also explain how users should proceed for getting the programs to use the available translations. They tell how people wanting to contribute and work at translations should contact the appropriate team. When reporting bugs in the `intl/' directory or bugs which may be related to internationalization, you should tell about the version of `gettext' which is used. The information can be found in the `intl/VERSION' file, in internationalized packages. Quick configuration advice ========================== If you want to exploit the full power of internationalization, you should configure it using ./configure --with-included-gettext to force usage of internationalizing routines provided within this package, despite the existence of internationalizing capabilities in the operating system where this package is being installed. So far, only the `gettext' implementation in the GNU C library version 2 provides as many features (such as locale alias, message inheritance, automatic charset conversion or plural form handling) as the implementation here. It is also not possible to offer this additional functionality on top of a `catgets' implementation. Future versions of GNU `gettext' will very likely convey even more functionality. So it might be a good idea to change to GNU `gettext' as soon as possible. So you need _not_ provide this option if you are using GNU libc 2 or you have installed a recent copy of the GNU gettext package with the included `libintl'. INSTALL Matters =============== Some packages are "localizable" when properly installed; the programs they contain can be made to speak your own native language. Most such packages use GNU `gettext'. Other packages have their own ways to internationalization, predating GNU `gettext'. By default, this package will be installed to allow translation of messages. It will automatically detect whether the system already provides the GNU `gettext' functions. If not, the GNU `gettext' own library will be used. This library is wholly contained within this package, usually in the `intl/' subdirectory, so prior installation of the GNU `gettext' package is _not_ required. Installers may use special options at configuration time for changing the default behaviour. The commands: ./configure --with-included-gettext ./configure --disable-nls will respectively bypass any pre-existing `gettext' to use the internationalizing routines provided within this package, or else, _totally_ disable translation of messages. When you already have GNU `gettext' installed on your system and run configure without an option for your new package, `configure' will probably detect the previously built and installed `libintl.a' file and will decide to use this. This might be not what is desirable. You should use the more recent version of the GNU `gettext' library. I.e. if the file `intl/VERSION' shows that the library which comes with this package is more recent, you should use ./configure --with-included-gettext to prevent auto-detection. The configuration process will not test for the `catgets' function and therefore it will not be used. The reason is that even an emulation of `gettext' on top of `catgets' could not provide all the extensions of the GNU `gettext' library. Internationalized packages have usually many `po/LL.po' files, where LL gives an ISO 639 two-letter code identifying the language. Unless translations have been forbidden at `configure' time by using the `--disable-nls' switch, all available translations are installed together with the package. However, the environment variable `LINGUAS' may be set, prior to configuration, to limit the installed set. `LINGUAS' should then contain a space separated list of two-letter codes, stating which languages are allowed. Using This Package ================== As a user, if your language has been installed for this package, you only have to set the `LANG' environment variable to the appropriate `LL_CC' combination. Here `LL' is an ISO 639 two-letter language code, and `CC' is an ISO 3166 two-letter country code. For example, let's suppose that you speak German and live in Germany. At the shell prompt, merely execute `setenv LANG de_DE' (in `csh'), `export LANG; LANG=de_DE' (in `sh') or `export LANG=de_DE' (in `bash'). This can be done from your `.login' or `.profile' file, once and for all. You might think that the country code specification is redundant. But in fact, some languages have dialects in different countries. For example, `de_AT' is used for Austria, and `pt_BR' for Brazil. The country code serves to distinguish the dialects. The locale naming convention of `LL_CC', with `LL' denoting the language and `CC' denoting the country, is the one use on systems based on GNU libc. On other systems, some variations of this scheme are used, such as `LL' or `LL_CC.ENCODING'. You can get the list of locales supported by your system for your country by running the command `locale -a | grep '^LL''. Not all programs have translations for all languages. By default, an English message is shown in place of a nonexistent translation. If you understand other languages, you can set up a priority list of languages. This is done through a different environment variable, called `LANGUAGE'. GNU `gettext' gives preference to `LANGUAGE' over `LANG' for the purpose of message handling, but you still need to have `LANG' set to the primary language; this is required by other parts of the system libraries. For example, some Swedish users who would rather read translations in German than English for when Swedish is not available, set `LANGUAGE' to `sv:de' while leaving `LANG' to `sv_SE'. In the `LANGUAGE' environment variable, but not in the `LANG' environment variable, `LL_CC' combinations can be abbreviated as `LL' to denote the language's main dialect. For example, `de' is equivalent to `de_DE' (German as spoken in Germany), and `pt' to `pt_PT' (Portuguese as spoken in Portugal) in this context. Translating Teams ================= For the Free Translation Project to be a success, we need interested people who like their own language and write it well, and who are also able to synergize with other translators speaking the same language. Each translation team has its own mailing list. The up-to-date list of teams can be found at the Free Translation Project's homepage, `http://www.iro.umontreal.ca/contrib/po/HTML/', in the "National teams" area. If you'd like to volunteer to _work_ at translating messages, you should become a member of the translating team for your own language. The subscribing address is _not_ the same as the list itself, it has `-request' appended. For example, speakers of Swedish can send a message to `sv-request@li.org', having this message body: subscribe Keep in mind that team members are expected to participate _actively_ in translations, or at solving translational difficulties, rather than merely lurking around. If your team does not exist yet and you want to start one, or if you are unsure about what to do or how to get started, please write to `translation@iro.umontreal.ca' to reach the coordinator for all translator teams. The English team is special. It works at improving and uniformizing the terminology in use. Proven linguistic skill are praised more than programming skill, here. Available Packages ================== Languages are not equally supported in all packages. The following -matrix shows the current state of internationalization, as of January +matrix shows the current state of internationalization, as of July 2002. The matrix shows, in regard of each package, for which languages PO files have been submitted to translation coordination, with a translation percentage of at least 50%. - Ready PO files bg ca cs da de el en eo es et fi fr - +-------------------------------------+ - a2ps | [] [] [] [] | - bash | [] [] [] [] | - bfd | [] [] | - binutils | [] [] | - bison | [] [] [] | - clisp | [] [] [] [] | - cpio | [] [] [] [] | - diffutils | [] [] [] [] [] [] | - enscript | [] [] | - error | [] [] | - fetchmail | () [] [] [] () | - fileutils | [] [] [] [] [] | - findutils | [] [] [] [] [] | - flex | [] [] [] | - gas | [] | - gawk | [] [] | - gcal | [] [] | - gcc | [] [] | - gettext | [] [] [] [] [] | - gnupg | [] [] [] [] [] [] | - gprof | [] [] | - grep | [] [] [] [] [] [] | - hello | [] [] [] [] [] [] [] [] | - id-utils | [] [] [] | - indent | [] [] [] [] | - jpilot | () [] [] [] | - jwhois | [] [] | - kbd | [] | - ld | [] [] | - libc | [] [] [] [] [] [] [] | - lilypond | [] [] | - lynx | [] [] [] [] | - m4 | [] [] [] [] [] | - make | [] [] [] [] | - mysecretdiary | [] [] | - nano | [] () [] [] [] [] | - nano_1_0 | [] () [] [] [] [] | - opcodes | [] [] [] | - parted | [] [] [] [] | - ptx | [] [] [] [] [] | - python | | - recode | [] [] [] [] [] [] | - sed | [] [] [] [] [] [] [] [] | - sh-utils | [] [] [] [] [] [] [] [] | - sharutils | [] [] [] [] [] [] | - sketch | () [] () | - soundtracker | [] [] [] | - sp | | - tar | [] [] [] [] [] [] | - texinfo | [] [] [] [] [] | - textutils | [] [] [] [] | - util-linux | [] [] [] [] | - wdiff | [] [] [] [] [] | - wget | [] [] [] [] [] [] [] [] | - +-------------------------------------+ - bg ca cs da de el en eo es et fi fr - 0 8 12 31 36 9 1 9 37 15 1 49 + Ready PO files be bg ca cs da de el en eo es et fi fr + +----------------------------------------+ + a2ps | [] [] [] [] | + bash | [] [] [] [] | + bfd | [] [] | + binutils | [] [] | + bison | [] [] [] [] | + clisp | [] [] [] [] | + clisp | | + clisplow | | + cpio | [] [] [] [] | + darkstat | () | + diffutils | [] [] [] [] [] [] | + enscript | [] [] | + error | [] [] [] | + fetchmail | [] () [] [] [] () | + fileutils | [] [] [] [] [] | + findutils | [] [] [] [] [] [] | + flex | [] [] [] [] [] | + gas | [] [] | + gawk | [] [] [] | + gcal | [] [] | + gcc | [] [] | + gettext | [] [] [] [] [] | + gnupg | [] [] [] [] [] [] [] | + gprof | [] [] | + gpsdrive | () () () () () | + grep | [] [] [] [] [] [] [] | + gretl | [] | + gthumb | () () () | + hello | [] [] [] [] [] [] [] [] [] | + id-utils | [] [] [] | + indent | [] [] [] [] [] | + jpilot | () [] [] [] | + jwhois | [] [] | + kbd | [] [] [] | + ld | [] [] | + libc | [] [] [] [] [] [] [] | + libiconv | [] [] [] [] | + lifelines | () () | + lilypond | [] [] [] | + lingoteach | [] [] | + lingoteach_lessons| () () | + lynx | [] [] [] [] [] | + m4 | [] [] [] [] [] | + make | [] [] [] [] | + man-db | [] () () [] () () | + mysecretdiary | [] [] [] | + nano | [] () [] [] [] [] | + nano_1_0 | [] () [] [] [] [] | + opcodes | [] [] [] | + parted | [] [] [] [] [] | + ptx | [] [] [] [] [] [] [] | + python | | + recode | [] [] [] [] [] [] | + sed | [] [] [] [] [] [] [] [] [] | + sh-utils | [] [] [] [] | + sharutils | [] [] [] [] [] [] | + sketch | () [] () | + soundtracker | [] [] [] | + sp | [] | + tar | [] [] [] [] [] [] | + texinfo | [] [] [] [] [] | + textutils | [] [] [] [] [] | + util-linux | [] [] [] [] [] [] | + vorbis-tools | [] | + wastesedge | | + wdiff | [] [] [] [] [] [] | + wget | [] [] [] [] [] [] [] [] [] [] | + +----------------------------------------+ + be bg ca cs da de el en eo es et fi fr + 0 2 19 10 30 44 9 1 12 45 16 3 53 gl he hr hu id it ja ko lv nb nl nn +-------------------------------------+ a2ps | () () [] | - bash | | + bash | [] | bfd | [] | binutils | [] | - bison | [] | + bison | [] [] [] [] | clisp | [] | - cpio | [] [] [] | - diffutils | [] [] | - enscript | [] | + clisp | | + clisplow | | + cpio | [] [] [] [] | + darkstat | () | + diffutils | [] [] [] [] [] | + enscript | [] [] | error | [] | - fetchmail | | - fileutils | [] [] | - findutils | [] [] [] [] [] [] | + fetchmail | [] | + fileutils | [] [] [] | + findutils | [] [] [] [] [] [] [] [] | flex | [] | gas | | gawk | [] | gcal | | gcc | [] | - gettext | [] | - gnupg | [] [] [] | + gettext | [] [] | + gnupg | [] [] [] [] | gprof | | - grep | [] [] | - hello | [] [] [] [] [] [] [] [] [] | - id-utils | [] | - indent | [] [] [] | + gpsdrive | [] () () | + grep | [] [] [] [] [] [] [] | + gretl | | + gthumb | () () | + hello | [] [] [] [] [] [] [] [] [] [] [] [] | + id-utils | [] [] | + indent | [] [] [] [] | jpilot | () () | - jwhois | | + jwhois | [] [] | kbd | | ld | | libc | [] [] [] [] | + libiconv | [] [] [] | + lifelines | | lilypond | [] [] | - lynx | [] [] | + lingoteach | [] | + lingoteach_lessons| | + lynx | [] [] [] [] | m4 | [] [] [] [] | - make | [] [] [] [] | - mysecretdiary | | - nano | [] [] [] () () [] | - nano_1_0 | [] [] [] () () [] | - opcodes | | + make | [] [] [] [] [] [] | + man-db | () () | + mysecretdiary | [] | + nano | [] [] [] [] [] [] | + nano_1_0 | [] [] [] [] [] | + opcodes | [] [] | parted | [] [] [] | - ptx | [] [] [] [] | + ptx | [] [] [] [] [] | python | | recode | [] [] [] | - sed | [] [] [] [] [] [] [] | - sh-utils | [] [] [] [] [] | + sed | [] [] [] [] [] [] [] [] | + sh-utils | [] [] [] | sharutils | [] [] [] | sketch | () | - soundtracker | [] | + soundtracker | [] [] | sp | | - tar | [] [] [] | + tar | [] [] [] [] [] | texinfo | [] [] [] | - textutils | [] [] | + textutils | [] [] [] | util-linux | () [] | - wdiff | | - wget | [] [] [] [] [] | + vorbis-tools | | + wastesedge | | + wdiff | [] [] [] | + wget | [] [] [] [] [] [] | +-------------------------------------+ gl he hr hu id it ja ko lv nb nl nn - 20 6 1 3 6 11 22 9 1 6 17 4 + 23 9 12 18 14 13 26 9 1 8 19 4 - no pl pt pt_BR ru sk sl sv tr uk zh - +-------------------------------------+ - a2ps | () () () [] [] [] () | 8 - bash | | 4 - bfd | [] [] | 5 - binutils | [] | 4 - bison | [] [] [] | 7 - clisp | | 5 - cpio | [] [] [] [] | 11 - diffutils | [] [] [] | 11 - enscript | [] [] [] | 6 - error | [] [] | 5 - fetchmail | () () | 3 - fileutils | [] [] [] [] | 11 - findutils | [] [] [] [] [] [] | 17 - flex | [] [] | 6 - gas | [] | 2 - gawk | [] [] | 5 - gcal | [] | 3 - gcc | [] | 4 - gettext | [] [] [] [] | 10 - gnupg | [] [] [] | 12 - gprof | [] [] | 4 - grep | [] [] [] [] [] | 13 - hello | [] [] [] [] [] [] [] | 24 - id-utils | [] [] | 6 - indent | [] [] [] [] | 11 - jpilot | () () | 3 - jwhois | () () | 2 - kbd | [] [] | 3 - ld | [] [] | 4 - libc | [] [] [] [] [] [] | 17 - lilypond | [] | 5 - lynx | [] [] [] | 9 - m4 | [] [] [] | 12 - make | [] [] [] [] | 12 - mysecretdiary | [] | 3 - nano | () [] [] [] | 12 - nano_1_0 | () [] [] [] | 12 - opcodes | [] [] | 5 - parted | [] [] [] | 10 - ptx | [] [] [] [] [] [] | 15 - python | | 0 - recode | [] [] [] [] | 13 - sed | [] [] [] [] [] [] | 21 - sh-utils | [] [] [] [] [] [] [] [] [] | 22 - sharutils | [] [] | 11 - sketch | () | 1 - soundtracker | | 4 - sp | | 0 - tar | [] [] [] [] [] [] [] | 16 - texinfo | [] [] | 10 - textutils | [] [] | 8 - util-linux | [] [] [] | 8 - wdiff | [] [] [] [] | 9 - wget | [] [] [] [] [] [] | 19 - +-------------------------------------+ - 35 teams no pl pt pt_BR ru sk sl sv tr uk zh - 54 domains 5 12 2 11 25 10 11 39 29 4 1 463 + no pl pt pt_BR ru sk sl sv tr uk zh_TW + +----------------------------------------+ + a2ps | () () () [] [] [] [] [] | 10 + bash | [] | 6 + bfd | [] [] | 5 + binutils | [] [] | 5 + bison | [] [] [] | 11 + clisp | | 5 + clisp | | 0 + clisplow | | 0 + cpio | [] [] [] [] | 12 + darkstat | [] () | 1 + diffutils | [] [] [] [] [] [] | 17 + enscript | [] [] [] [] | 8 + error | [] [] [] | 7 + fetchmail | () () [] | 6 + fileutils | [] [] [] [] [] | 13 + findutils | [] [] [] [] [] [] [] | 21 + flex | [] [] [] | 9 + gas | [] | 3 + gawk | [] [] | 6 + gcal | [] [] | 4 + gcc | [] | 4 + gettext | [] [] [] [] [] [] | 13 + gnupg | [] [] [] | 14 + gprof | [] [] | 4 + gpsdrive | [] [] () | 3 + grep | [] [] [] [] | 18 + gretl | | 1 + gthumb | () () [] | 1 + hello | [] [] [] [] [] [] [] | 28 + id-utils | [] [] [] [] | 9 + indent | [] [] [] [] [] | 14 + jpilot | () () [] | 4 + jwhois | [] () () [] [] | 7 + kbd | [] [] | 5 + ld | [] [] | 4 + libc | [] [] [] [] [] [] | 17 + libiconv | [] [] [] [] | 11 + lifelines | [] | 1 + lilypond | [] | 6 + lingoteach | [] [] | 5 + lingoteach_lessons| | 0 + lynx | [] [] [] [] | 13 + m4 | [] [] [] | 12 + make | [] [] [] [] | 14 + man-db | | 3 + mysecretdiary | [] [] [] | 7 + nano | [] [] [] [] | 15 + nano_1_0 | [] [] [] [] | 14 + opcodes | [] [] | 7 + parted | [] [] [] | 11 + ptx | [] [] [] [] [] [] [] | 19 + python | | 0 + recode | [] [] [] [] [] [] | 15 + sed | [] [] [] [] [] [] | 23 + sh-utils | [] | 8 + sharutils | [] [] [] [] | 13 + sketch | [] () [] | 4 + soundtracker | [] | 6 + sp | | 1 + tar | [] [] [] [] [] [] [] | 18 + texinfo | [] [] | 10 + textutils | [] [] [] [] [] | 13 + util-linux | [] [] [] | 10 + vorbis-tools | [] | 2 + wastesedge | | 0 + wdiff | [] [] [] [] [] | 14 + wget | [] [] [] [] [] [] [] [] | 24 + +----------------------------------------+ + 36 teams no pl pt pt_BR ru sk sl sv tr uk zh_TW + 67 domains 4 15 2 24 26 12 10 47 42 4 8 594 Some counters in the preceding matrix are higher than the number of visible blocks let us expect. This is because a few extra PO files are used for implementing regional variants of languages, or language dialects. For a PO file in the matrix above to be effective, the package to which it applies should also have been internationalized and distributed as such by its maintainer. There might be an observable lag between the mere existence a PO file and its wide availability in a distribution. - If January 2002 seems to be old, you may fetch a more recent copy of + If July 2002 seems to be old, you may fetch a more recent copy of this `ABOUT-NLS' file on most GNU archive sites. The most up-to-date matrix with full percentage details can be found at `http://www.iro.umontreal.ca/contrib/po/HTML/matrix.html'. Using `gettext' in new packages =============================== If you are writing a freely available program and want to internationalize it you are welcome to use GNU `gettext' in your package. Of course you have to respect the GNU Library General Public License which covers the use of the GNU `gettext' library. This means in particular that even non-free programs can use `libintl' as a shared library, whereas only free software can use `libintl' as a static library or use modified versions of `libintl'. Once the sources are changed appropriately and the setup can handle to use of `gettext' the only thing missing are the translations. The Free Translation Project is also available for packages which are not developed inside the GNU project. Therefore the information given above applies also for every other Free Software Project. Contact `translation@iro.umontreal.ca' to make the `.pot' files available to the translation teams. diff --git a/contrib/gnu-sort/ChangeLog b/contrib/gnu-sort/ChangeLog index 87e22a2d1bf8..bc43c75876f1 100644 --- a/contrib/gnu-sort/ChangeLog +++ b/contrib/gnu-sort/ChangeLog @@ -1,8574 +1,8743 @@ +2002-07-29 Jim Meyering + + * Version 2.1. + + * src/tail.c: Adjust command examples in comments to use + POSIX 1003.1-2001 option syntax. + * src/sort.c: Likewise. + From Paul Eggert. + +2002-07-28 Jim Meyering + + * configure.ac: Require automake-1.6b. + +2002-07-20 Jim Meyering + + * Version 2.0.22. + + * Upgrade to cvs automake -- required to work with 2.53b. + * configure.ac: Require automake-1.6a. + + * Upgrade to gettext-0.11.3. + * Upgrade to autoconf-2.53b. + +2002-07-02 Jim Meyering + + * src/cat.c, src/cksum.c, src/comm.c, src/csplit.c, src/cut.c (usage): + Use the PACKAGE_BUGREPORT e-mail address, rather than hard-coding it. + * src/expand.c, src/fmt.c, src/fold.c, src/head.c, src/join.c: Likewise. + * src/md5sum.c, src/nl.c, src/od.c, src/paste.c, src/pr.c: Likewise. + * src/sort.c, src/split.c, src/sum.c, src/tac.c, src/tail.c: Likewise. + * src/tr.c, src/tsort.c, src/unexpand.c, src/uniq.c, src/wc.c: Likewise. + +2002-06-22 Jim Meyering + + * configure.ac: Invoke AC_GNU_SOURCE very early. + +2002-06-15 Jim Meyering + + * src/cksum.c: Include before system.h. + +2002-05-22 Jim Meyering + + * src/od.c [struct dummy]: Renamed. Don't rely on portability of + zero-length arrays. Instead, use sizes of -1 or +1. + Reported by Eric Backus. + +2002-05-14 Jim Meyering + + `od -t f8' didn't work. + This bug was introduced with my change of 2000-10-22 (textutils-2.0.8). + * src/od.c (ulonglong_t): Move declaration to precede new use. + [enum size_spec] (N_SIZE_SPECS): New member. + (width_bytes): Add initializer corresponding to ulonglong_t type. + (struct assert_width_bytes_matches_size_spec_decl): Declare. + Based on a patch from Tony Kocurko. + +2002-04-29 Paul Eggert + + Use ENABLE_NLS only for gettext-related stuff, not also for + setlocale-related stuff. From a patch suggested by Bruno + Haible. + + Since setlocale is now almost universal, this patch also + removes some of the optimizations that clutter up the code and + which don't help all that much even on hosts that lack + setlocale. + + * lib/hard-locale.c: Upgrade to version used in GNU Diffutils 2.8.1. + * m4/prereq.m4 (jm_PREREQ_HARD_LOCALE): Check for stdlib.h. + Do not check for alloca.h (no longer used) or stdbool.h (was never + used?). Add AM_C_PROTOTYPES since hard-locale.h uses it. + + * src/sys2.h (HAVE_SETLOCALE): Define to 0 if not defined. + * src/comm.c (compare_files): Replace #ifdef ENABLE_NLS + with if (HAVE_SETLOCALE). + * src/join.c (keycmp): Likewise. + * src/sort.c (keycompare, compare): Likewise. + + * src/comm.c (hard_LC_COLLATE): Define even if ! ENABLE_NLS. + * src/join.c (hard_LC_COLLATE): Likewise. + * src/sort.c (hard_LC_COLLATE): Likewise. + + * src/comm.c (main): Always initialize hard_LC_COLLATE. + Put initialization next to other locale-related stuff. + * src/join.c (main): Likewise. + * src/sort.c (main): Likewise. + + * src/sort.c: Include even if ! ENABLE_NLS. + (decimal_point, th_sep): Depend on HAVE_SETLOCALE, not ENABLE_NLS. + (main): Likewise. + (MONTHTAB_CONST): Remove; all uses removed. + (struct_month_cmp, inittables): Do not depend on ENABLE_NLS. + (main): hard_LC_TIME locale does not depend on ENABLE_NLS. + +2002-04-29 Jim Meyering + + * src/tail.c (parse_obsolescent_option): Do not limit the maximum + line/byte count (when specified via obsolete option syntax) to be + less than ULONG_MAX on systems where OFF_T_MAX is larger. + +2002-04-28 Paul Eggert + + * src/tail.c (tail_forever): Do not treat + max_n_unchanged_stats_between_opens specially (as if it were + infinite) if it has its maximal value. Similarly for + max_n_consecutive_size_changes_between_opens. + +2002-04-28 Jim Meyering + + Avoid new warnings from gcc. + * src/pr.c [struct COLUMN] : Declare member to be const. + (init_fps): Declare local variable `firstname' to be const. + +2002-04-25 Jim Meyering + + * src/wc.c (usage): Adjust ordering to match that of default output. + Suggestion from Gaute Hvoslef Kvalnes. + + * configure.ac (AM_GNU_GETTEXT): Add external arg. + (AC_CONFIG_FILES): Remove intl/Makefile. + * Makefile.am (SUBDIRS): Remove intl. + * src/Makefile.am (INCLUDES): Remove -I../intl. + +2002-04-15 Jim Meyering + + * src/tr.c (usage): Give a verbose description of --squeeze-repeats, + including the mention of SET1. Reported by Dan Jacobson. + +2002-04-12 Jim Meyering + + * src/Makefile.am (LDADD): List ../lib/libfetish.a both before + and after @LIBINTL@. Thanks to Paul Eggert for the fix and to + Bruno Haible for diagnosing the problem. + +2002-03-16 Jim Meyering + + * src/Makefile.am (datadir): Don't override $(datadir) + which might be set by --datadir and different from $(prefix)/share. + Patch from Albert Chin-A-Young. + +2002-03-10 Jim Meyering + + * configure.ac (AM_INIT_AUTOMAKE): Specify the required version + of automake (1.6), and options (gnits dist-bzip2), rather than... + * Makefile.am (AUTOMAKE_OPTIONS): ...here. Remove definition. + +2002-03-09 Jim Meyering + + * Makefile.am (EXTRA_DIST): Remove config/config.rpath, + now that automake knows about it. + +2002-03-05 Jim Meyering + + * Makefile.am (EXTRA_DIST): Add config/config.rpath. + +2002-02-23 Paul Eggert + + * src/cat.c (simple_cat, cat, main): Use char * rather than + unsigned char *, to avoid some technical violations of the C + standard, and to avoid the need for casts. I guess the + unsigned char * was probably there for improved performance + with ancient compilers, but the code wasn't using the + unsignedness any more. + Reported by Nelson H. F. Beebe. + + (cat): int -> size_t for full_write arg. This doesn't + fix any bugs since the buffer size is small, but it makes + the code a bit clearer. + 2002-02-18 Jim Meyering * Version 2.0.21. * tests/tail/Test.pm: For tests matching /^(obs|err-[134])/, put _POSIX2_VERSION=199209 in the environment, so the tests succeed even when the environment would have contained _POSIX2_VERSION=200112. * tests/uniq/Test.pm (test_vector): Rename tests of obsolete usage to have prefix of `obs'. Adjust environment for those tests as above. * tests/mk-script: When getting input via a pipe, apply the environment settings to the command in question, not to `cat'. No longer allow more than one element in the $Test::env or $Test::env_default arrays. No caller used that feature. 2002-02-17 Jim Meyering * tests/head/Test.pm: Accept pre-POSIX 1003.1-2001 options. * tests/pr/Test.pm: Likewise. * tests/sort/Test.pm: Likewise. * Makefile.cfg (wget_files): Reflect new location of these files. 2002-02-16 Jim Meyering * ABOUT-NLS, intl/*: Update to Gettext 0.11. * INSTALL: Update to Autoconf 2.52h. * src/Makefile.am (datadir): Use `share' in place of @DATADIRNAME@, per Gettext-0.11. (LDADD): Use @LIBINTL@ instead of @INTLLIBS@, per Gettext-0.11. * configure.ac (AC_CONFIG_AUX_DIR): New macro invocation. (ALL_LINGUAS): Remove: now in po/LINGUAS as per Gettext 0.11. * config: New subdirectory, containing the following files from .: config.guess, config.sub, depcomp, missing, install-sh, mkinstalldirs. Move the following files here from doc: texinfo.tex, mdate-sh. 2002-02-15 Paul Eggert Add support for _POSIX2_VERSION, which lets you pick which POSIX version you want the utilities to conform to. Remove warnings about failure to conform to a future POSIX version. * NEWS, doc/coreutils.texi: Document _POSIX2_VERSION. * doc/coreutils.texi (Standards Conformance): New section. * src/expand.c: Include posixver.h. (usage): Document only the intersection of the old and new behaviors, to encourage portability. * src/fold.c: Likewise. * src/head.c: Likewise. * src/od.c: Likewise. * src/pr.c: Likewise. * src/sort.c: Likewise. * src/split.c: Likewise. * src/uniq.c: Likewise. * src/tail.c: Likewise. * src/unexpand.c: Likewise. * src/expand.c (shortopts): Remove; no longer needed. * src/od.c (short_options): Likewise. * src/pr.c (short_options): Likewise. * src/sort.c (short_options): Likewise. * src/split.c (shortopts): Likewise. * src/uniq.c (shortopts): Likewise. * src/unexpand.c (shortopts): Likewise. * src/od.c (COMMON_SHORT_OPTIONS): New macro. * src/pr.c (COMMON_SHORT_OPTIONS): Likewise. * src/sort.c (COMMON_SHORT_OPTIONS): Likewise. * src/expand.c (main): Revert to previous behavior, except report an error and exit if the obsolete syntax is used and if conforming to the new standard. * src/fold.c (main): Likewise. * src/head.c (main): Likewise. * src/split.c (main): Likewise. * src/uniq.c (main): Likewise. * src/tail.c (main, parse_obsolescent_option): Likewise. * src/unexpand.c (main): Likewise. * src/od.c (STRINGS_OPTION, WIDTH_OPTION): Remove; no longer needed. All uses changed back to the corresponding short options. * src/pr.c (SEP_STRING_OPTION): Likewise. * src/od.c (main): Parse options using POSIX 1003.1-2001 rules if conforming to that standard. Do not warn of obsolete options. * src/pr.c (main): Likewise. * src/sort.c (main): Likewise. * src/system.h (_POSIX2_VERSION, POSIX2_VERSION): Remove definitions. (OPTARG_POSIX, OBSOLETE_OPTION_WARNINGS): Likewise. 2002-02-16 Jim Meyering * tests/misc/split-a: New test for change of 2002-02-05. * tests/misc/Makefile.am (TESTS): Add split-a. 2002-02-12 Jim Meyering * src/split.c (DEFAULT_SUFFIX_LENGTH): Define constant. (suffix_length): Use it here. (usage): Use it here. 2002-02-05 Paul Eggert Change 'split' to conform to POSIX. It now accepts an -a or --suffix-length option, and refuses to generate more files than there are suffixes. * NEWS, doc/coreutils.texi: Document this. * src/split.c: Include "dirname.h". (outbase): Now a global var. (outfile_end): Remove. (suffix_length): New var. (shortopts, longopts, main): Add -a or --suffix-length. (next_file_name): Implement -a. Do not extend the suffix length. Check for file names that are too long. (main): Move outfile initialization to next_file_name. * src/split.c (shortopts): Remove -v (a typo). 2002-02-11 Jim Meyering * tests/misc/head-pos: New file/test for the change of 2002-02-08. * tests/misc/Makefile.am (TESTS): Add head-pos. * src/tr.c (unquote): Make comment a little clearer. 2002-02-08 Paul Eggert * src/head.c (head_lines): If we have read too much data, try to seek back to the position we would have gotten to had we been reading one byte at a time. POSIX currently doesn't require this, but it's easy to do and some software relies on it. 2002-02-02 Jim Meyering * src/uniq.c (main): Prepend `warning: ' to the diagnostic, so it's consistent with all of the other `... is obsolete...' diagnostics. 2002-01-31 Paul Eggert Add more support for POSIX 1003.1-2001, which requires removal for support of obsolete "-N" option syntax in expand, head, fold, split, tail, unexpand, uniq, and which prohibits options with optional arguments in od and pr. * NEWS: Document this. * doc/coreutils.texi: Likewise. * src/expand.c (usage): Likewise. * src/fold.c (usage): Likewise. * src/head.c (usage): Likewise. * src/od.c (usage): Likewise. * src/pr.c (usage): Likewise. * src/split.c (usage): Likewise. * src/unexpand.c (usage): Likewise. * src/uniq.c (usage): Likewise. * NEWS: Improve doc for previous fix in this area. * src/expand.c (shortopts): New constant. * src/od.c (short_options): Likewise. * src/pr.c (short_options): Likewise. * src/sort.c (short_options): Likewise. * src/split.c (shortopts): Likewise. * src/unexpand.c (shortopts): Likewise. * src/uniq.c (shortopts): Likewise. * src/expand.c (main): Check for obsolete options. * src/fold.c (main): Likewise. * src/head.c (main): Likewise. * src/od.c (main): Likewise. * src/pr.c (main): Likewise. * src/sort.c (main): Likewise. * src/split.c (main): Likewise. * src/tail.c (parse_obsolescent_option, main): Likewise. * src/unexpand.c (main): Likewise. * src/uniq.c (main): Likewise. * src/head.c (header_mode_option): New constant. * src/od.c: New enum for long options. * src/pr.c (SEP_STRING_OPTION): New enum value. (long_options): Use it. * src/split.c (main): Use -1, not EOF, for getopt_long. * src/system.h (OPTARG_POSIX, OBSOLETE_OPTION_WARNINGS): New macros. * src/fold.c (main): Fix bug: -- -N was treated as -N --. 2002-02-01 Jim Meyering * src/tail.c (start_lines): Handle the case in which bytes_read is zero. 2002-01-28 Jim Meyering * Makefile.maint (ac-check): Remove, now that we no longer have to duplicate that information. (local-check): Remove ac-check. * Makefile.am (AUTOMAKE_OPTIONS): Require automake-1.5d. * configure.ac (AM_INIT_AUTOMAKE): Use zero-arg form of this macro. (ALL_LINGUAS): Add the bokmål dialect of Norwegian (nb). 2002-01-27 Jim Meyering tail -n +2 would perform an extra read after encountering EOF * src/tail.c (start_lines): Detect EOF, inform caller. (tail_lines): Upon EOF in start_lines, return immediately. 2002-01-22 Jim Meyering * Version 2.0.20. 2002-01-22 Paul Eggert * src/od.c (skip): Remove redundant check for overflow. fseeko's seek argument is already known to be in range, since it is less than the file size. 2002-01-22 Jim Meyering * po/POTFILES.in: Add lib/xmemcoll.c. 2002-01-22 Paul Eggert * src/tac.c (save_stdin): Report proper errno value after fwrite failures. Do not bother to rewind the temp file, as it'll be read backwards anyway. 2002-01-22 Paul Eggert * src/ptx.c (swallow_file_in_memory): Work even if `open' returns 0. Check for `close' error. 2002-01-22 Paul Eggert Port to glibc 2.2.5, whose mmap stdio positions the underlying file descriptor at a nonzero offset after an fopen. * m4/jm-macros.m4 (jm_MACROS): Require AC_FUNC_FSEEKO. * src/od.c (skip): Use fseeko rather than lseek. (Also, check for overflow when converting uintmax_t to off_t.) * src/sys2.h (EOVERFLOW, fseeko): New macros. 2001-01-21 Paul Eggert * src/comm.c: Include xmemcoll.h, not memcoll.h. Use xmemcoll instead of memcoll. * src/join.c: Likewise. * src/sort.c: Likewise. * src/sort.c (die): Do not invoke cleanup; atexit does this now. (main): Use atexit to invoke cleanup on exit. Set xmemcoll_exit_failure to SORT_FAILURE. 2002-01-20 Paul Eggert * src/comm.c (compare_files): Don't assume that the difference between two size_t values can be stored in an int; this doesn't work, for example, on 64-bit Solaris. * src/sort.c (CMP_WITH_IGNORE): Likewise. * src/join.c (keycmp): Likewise. Also, fix the type of two locals to be size_t instead of int. 2002-01-20 Paul Eggert * src/csplit.c: Don't include "xalloc.h", as system.h already does that via sys2.h. * src/sort.c: Likewise. 2002-01-19 Jim Meyering * configure.ac: Remove code that set LIBOBJS in support of ansi2knr. The latest version of autoconf objected to it, and besides, I don't care too much about K&R support these days. 2002-01-18 Jim Meyering Fix tr so it no longer gets a failed assertion for [::] or [==]. * src/tr.c (xmemdup): Rename from `substr' and rewrite to take only pointer/length parameters. (build_spec_list): Update sole caller. Properly diagnose the invalid specs [::] and [==]. Pawel Prokop reported that `tr '[::]' x' elicits a failed assertion. * tests/tr/Test.pm: Add tests for the above fix. 2002-01-14 Paul Eggert * src/sort.c (main): Issue a warning for obsolete usage, unless POSIXLY_CORRECT. * src/tail.c (parse_obsolescent_option): Likewise. * src/uniq.c (main): Likewise. 2002-01-14 Paul Eggert Add support for POSIX 1003.1-2001, which requires removal for support of obsolete "+" option syntax in sort, tail, and uniq. * NEWS: Document this. * src/sort.c (usage, main): Implement this. * src/tail.c (usage, parse_obsolescent_option): Likewise. * src/uniq.c (usage, main): Likewise. * src/system.h (_POSIX2_VERSION, POSIX2_VERSION): New macros. * src/sys2.h (ISDIGIT, S_ISVTX, S_IXUGO): Comment fix. 2002-01-14 Jim Meyering * tests/misc/sort: New test, for just-fixed bug. * tests/misc/Makefile.am (TESTS): Add sort. (TESTS_ENVIRONMENT): Add PROG=$$tst. This is a gross kludge that depends on the name of an automake-generated local variable. 2002-01-13 Paul Eggert * src/sort.c (mergefps): Fix bug when eliminating empty files. Bug reported by James Hutt. 2002-01-13 Jim Meyering * tests/help-version: Tweak to work with the programs in diffutils. 2002-01-09 Jim Meyering * configure.in (ALL_LINGUAS): Add Turkish (tr). 2001-12-18 Paul Eggert * src/sort.c (usage, specify_sort_size): Accept 'K' (which is now preferred, as it connotes 1024) as well as 'k'. * NEWS: Document this change. 2001-12-17 Jim Meyering * src/uniq.c (usage): Remove stray `)' in --help output. Reported by Padraig Brady. * Makefile.maint (real_dir): Remove unused variable. (url_dir_list): Don't set define it here, but... * Makefile.cfg (url_dir_list): ...define it here instead. (hosts, a_host, b_host): Remove now-unused variables. (alpha_subdir, a_url_dir, b_url_dir): Remove now-unused variables. 2001-12-12 Bruno Haible Paul Eggert * src/wc.c (wc): Use ISSPACE and iswspace in addition to hard-coding the ASCII space character. 2001-12-08 Jim Meyering * Version 2.0.19. * Makefile.cfg (wget_files): Remove ansi2knr.c, since it's no longer available at that location. 2001-12-03 Jim Meyering * src/od.c, src/tail.c, src/tr.c (usage): I will not split translatable strings in the middle of a sentence. I will not split translatable strings in the middle of a sentence. I will not split translatable strings in the middle of a sentence. ... * src/join.c (usage): Merge lone "\n" into adjacent string. * src/md5sum.c (usage): Likewise. * src/paste.c (usage): Likewise. Reported by Karl Eichwalder. 2001-12-02 Jim Meyering * Makefile.maint (emit-upload-commands): Fix typo: s/distdir/PACKAGE/. 2001-12-01 Jim Meyering * Version 2.0.18. * Makefile.maint (po-check): Check for uses of _() not just in .c files, but also in .h files. Ideally, there would be just one translation for all --help and --version description strings. Before this change, they differed only in the number of blanks between the option name and the description. Someday gettext may be smart enough to merge such strings and to reconstruct the proper spacing at run time. In the mean time, now there is one string for each, and those descriptions in the --help output are no longer aligned with the others. * src/sys2.h (HELP_OPTION_DESCRIPTION): New macro. (VERSION_OPTION_DESCRIPTION): New macro. * src/cat.c, src/cksum.c, src/comm.c, src/csplit.c, src/cut.c: * src/expand.c, src/fmt.c, src/fold.c, src/head.c, src/join.c: * src/md5sum.c, src/nl.c, src/od.c, src/paste.c, src/pr.c, src/ptx.c: * src/sort.c, src/split.c, src/sum.c, src/sys2.h, src/tac.c: * src/tail.c, src/tr.c, src/tsort.c, src/unexpand.c, src/uniq.c: * src/wc.c (usage): Use new macros, HELP_OPTION_DESCRIPTION and VERSION_OPTION_DESCRIPTION instead of hard-coding --help and --version descriptions. * po/POTFILES.in: Add src/sys2.h. 2001-11-27 Jim Meyering * src/system.h (SET_MODE) [O_BINARY]: Call setmode, not set_mode. From Matthew Smith. 2001-11-26 Jim Meyering * src/cat.c (cat): Don't test whether the full_write return value (of type size_t) is less than 0. Reported by Nelson H. F. Beebe, as a warning from Irix 6.5's C compiler. 2001-11-25 Jim Meyering * src/cksum.c (usage): Indent --help and --version strings to start in the 7th column. * src/pr.c (usage): Likewise. 2001-11-23 Jim Meyering Factor out some common strings to make translation easier. * cat.c, cksum.c, comm.c, csplit.c, cut.c, expand.c, fmt.c, fold.c: * head.c, join.c, md5sum.c, nl.c, od.c, paste.c, pr.c, ptx.c: * sort.c, split.c, sum.c, tac.c, tail.c, tr.c, tsort.c, unexpand.c: * uniq.c, wc.c: Split usage strings so that --help and --version descriptions are alone in their own string. Likewise for the one that says ``Mandatory arguments to long options are mandatory for short options too.'' Suggestion from Karl Eichwalder. * src/ptx.c (main): Don't split copyright string in the middle of a sentence. 2001-11-22 Jim Meyering * Version 2.0.17. csplit could get a failed assertion: printf 'a\n\n'|csplit - '/^$/' 2 * src/csplit.c: No longer include assert.h. (process_line_count): Remove invalid assertion. The test that caused the failure has been in the code since before 1992-11-08, but since 1996 it's been in an assertion. That assertion was disabled by default until textutils-1.22g (1999-01-10). Reported by Eric Pemente. * tests/misc/csplit: New file. Test for the above fix and others. * tests/misc/Makefile.am (TESTS): Add csplit. * tests/Makefile.am (EXTRA_DIST): Add lang-default. 2001-11-19 Jim Meyering * src/tail.c (xlseek): New function. Call it instead of lseek, in most cases, so any failure is reported. 2001-11-18 Jim Meyering * config.sub: Update from master repository. * config.guess: Likewise. 2001-11-17 Jim Meyering * Makefile.cfg ($(srcdir)/m4/jm-glibc-io.m4): Use $(move_if_change), not the literal `move-if-change'. * src/tail.c (tail_forever): Add a cast. Avoid compiler warnings. * src/fmt.c: Remove `unsigned' attribute from type of global `prefix'. (copy_rest): Cast to (unsigned char *) before dereferencing. (get_prefix): Likewise. * configure.ac: Replace use of the one-arg form of AC_INIT with a use of the 3-arg form and a use of AC_CONFIG_SRCDIR. 2001-11-14 Paul Eggert * src/tail.c (file_lines): Remove unnecessary cast. (tail_lines): Could have called file_lines even though the first lseek failed. Fix that. 2001-11-14 Jim Meyering * src/tail.c (file_lines): Add a parameter, start_pos. Work properly even when the read pointer is not at beginning of file. (tail_lines): Call file_lines for any regular file, as long as lseek can be used to seek to its end, not just when the initial read pointer is at beginning of file. 2001-11-13 Jim Meyering * src/tail.c (tail_lines): Move declaration of local `length' into scope where it's used. (tail_file): Likewise for local `stats'. * tests/tail-2/Makefile.am (TESTS): Add proc-ksyms and start-middle. * tests/tail-2/start-middle: New test, for the bug fixed on 1995-07-24. `tail /proc/ksyms' would segfault on Linux. * src/tail.c (tail_lines): Use status of lseek (...SEEK_END) call in deciding whether to call file_lines or pipe_lines. From Herbert Xu. * tests/tail-2/proc-ksyms: New test, for the above fix. 2001-11-11 Jim Meyering * src/od.c (struct tspec): Declare function pointer with prototype. * src/sum.c (main): Declare function pointer with prototype. * src/tsort.c (count_items): Mark parameter as unused. * src/sort.c (struct_month_cmp): Guard definition with the same cpp condition that guards the use. * src/tail.c (xwrite): Remove assertion that size_t N >= 0. * src/pr.c (struct COLUMN) [print_func]: Declare with a protype. (struct COLUMN) [char_func]: Declare with a protype. * src/od.c (parse_old_offset): Declare to be static. * src/join.c (make_blank): Declare to be static. (prfield): Declare local to be of type size_t, not int. Some help strings were very long. Split them so that they're no longer than the magic length 509 that ISO C89 compilers are required to support. Sorry, translators :-( * src/md5sum.c (usage): Split --help output into smaller pieces. * src/ptx.c (main, usage): Likewise. * src/fmt.c (usage): Likewise. * src/wc.c (usage): Split --help output into smaller pieces. Use fputs, not printf. * src/cat.c (usage): Likewise. * src/cut.c (usage): Likewise. * src/csplit.c (usage): Likewise. * src/expand.c (usage): Likewise. * src/head.c (usage): Likewise. * src/join.c (usage): Likewise. * src/nl.c (usage): Likewise. * src/od.c (usage): Likewise. * src/pr.c (usage): Likewise. * src/sort.c (usage): Likewise. * src/split.c (usage): Likewise. * src/tac.c (usage): Likewise. * src/tr.c (usage): Likewise. * src/unexpand.c (usage): Likewise. * src/uniq.c (usage): Likewise. * src/tail.c (usage): Likewise. * src/sys2.h (alloca) [__GNUC__]: Define only if not already defined. * src/Makefile.am (sort_LDADD): Reflect spelling change: s/POW_LIBM/POW_LIB/. 2001-11-10 Jim Meyering * src/nl.c (proc_text): Use `puts' to output a string of spaces, not printf. This avoids a warning from gcc's -Wformat-security. 2001-11-09 Jim Meyering * configure.ac: Use AC_CONFIG_FILES(...) and call AC_OUTPUT with no arguments. 2001-11-05 Jim Meyering * src/sort.c (usage): Don't recommend setting LC_COLLATE=C. That can cause problems (now documented in coreutils.texi). 2001-11-04 Jim Meyering * uniq.c, unexpand.c, tail.c, tac.c, split.c, sort.c, pr.c, paste.c: * od.c, nl.c, head.c, fold.c, expand.c, cut.c, csplit.c (usage): Say that ``Mandatory arguments to long options are mandatory for short options too.'' * src/ptx.c (usage): Add one-line description. 2001-10-28 Jim Meyering * tests/sum/sysv: New test for the fix below. * tests/sum/Makefile.am (TESTS): Add sysv. 2001-10-27 Paul Eggert * src/sum.c: Include human.h. (ROTATE_RIGHT): Remove; it was slow and the comment was misleading anyway. (bsd_sum_file): Avoid unsigned and long when they're not needed. Compute total input bytes as uintmax_t, not long, so that the code works even with files whose size does not fit in 'long'. (sysv_sum_file): Likewise. (sysv_sum_file): Do not reduce checksum until the end; this is what System V does. Reported by Nick Lawes. 2001-10-27 Jim Meyering Give an accurate diagnostic when `head --bytes=30M' fails. * src/head.c (string_to_integer): Check explicitly for overflow, and lump everything else together as `invalid'. 2001-10-04 Jim Meyering Rename --all-repeated argument `precede' to `prepend'. * src/uniq.c (enum delimit_method): s/DM_PRECEDE/DM_PREPEND/ and change all uses. (delimit_method_string): s/precede/prepend/ * tests/uniq/Test.pm: Adjust tests accordingly. Patches by Padraig Brady. * src/uniq.c (usage): Correct typo in description of --all-repeated. Patch by Padraig Brady. 2001-09-30 Jim Meyering * Version 2.0.16. tail -F would segfault * src/tail.c (parse_options): Add missing `break;' from last change. 2001-09-28 Jim Meyering * configure.ac: Tell automake to use the file name `config.hin' rather than the two-`.' config.h.in. 2001-09-26 Jim Meyering * man/Makefile.am (common_dep): Define it, so we depend on configure.ac. * configure.ac: Renamed from configure.in. 2001-09-25 Jim Meyering * src/Makefile.am (sort_DEPENDENCIES): Remove definition altogether. It adds nothing to the existing: $(PROGRAMS): ../lib/libfetish.a. 2001-09-24 Paul Eggert * src/Makefile.am (sort_DEPENDENCIES): Depend only on ../lib/libfetish.a, not on $(LDADD), since the latter may contain `-lintl'. This resembles yesterday's patch for the `$(PROGRAMS)' rule. 2001-09-23 Jim Meyering * Version 2.0.15. * src/Makefile.am ($(PROGRAMS)): Depend only on ../lib/libfetish.a, not on $(LDADD), since the former may contain `-lintl'. * Makefile.am (SUBDIRS): When using newer gettext (in which intl/libintl.h is created by rules intl/Makefile) `intl' must precede `lib'. 2001-09-22 Jim Meyering * ABOUT-NLS: Upgrade to gettext 0.10.40 * intl: Upgrade entire directory to gettext 0.10.40 * po/ChangeLog: New file. * Makefile.am (DISTCLEANFILES): Remove definition, since intl/libintl.h is no longer created via configure. * tests/od/od-N: New test for the 2001-09-19 fix by Ian Bruce. * tests/od/Makefile.am: New file. * tests/od: New directory. * tests/Makefile.am (SUBDIRS): Add od. * configure.in (AC_OUTPUT): Add tests/od/Makefile. 2001-09-19 Paul Eggert * src/od.c (MAX_ADDRESS_LENGTH, pseudo_offset, format_address, n_bytes_to_skip, skip, format_address_none, format_address_std, format_address_paren, format_address_label, write_block, parse_old_offset, dump, dump_strings, main): Use uintmax_t, not off_t, for file addresses, so that we can handle multiple large files even if the sum of their sizes exceeds off_t limits. (print_s_char, print_char, print_s_short, print_short, print_int, print_long, print_long_long, print_float, print_double, print_long_double, dump_hexl_mode_trailer, print_named_ascii, print_ascii, decode_one_format): Use size_t, not off_t, for in-memory byte counts. (end_offset): New var. (dump, dump_strings): Use it. (main): Set it, but check for overflow while doing so. (skip): Report an error if an in-range lseek fails on a regular file, as something's seriously wrong. Check for negative regular file sizes (possible with some broken NFS implementations). (parse_old_offset): Now all offsets are valid, so return a success boolean and take a pointer to an offset as an argument. All callers changed. (dump_strings): Check for overflow when computing end_offset - string_min. (main): Remove OFF_T_MAX checks that are no longer needed. Don't bother assigning through temporary when there's no size limit to check. 2001-09-19 Jim Meyering * src/od.c (open_next_file): Use SETVBUF, not setbuf. (skip): Revert part of last change: use lseek, not fseek. When --read-bytes=N (-N N) is used, disable input buffering in the standard I/O library. Otherwise, od would read more input than requested. This could have caused problems when reading from pipes, character devices, or open file descriptors inherited from a parent process. * src/od.c (open_next_file): New function, factored out of... (skip): Adapt to use open_next_file. (read_char): Likewise. (read_block): Likewise. (main): Likewise. (dump): Fix an off-by-one error that could have made od fail to report a read error when reading from a named file (not stdin). (check_and_close): Account for the fact that in_stream may now be NULL. (usage): Correct descriptions of -j and -N options. Patch by Ian Bruce. 2001-09-16 Jim Meyering * Makefile.am (AUTOMAKE_OPTIONS): Require automake-1.5. * man/Makefile.am: Revamp this file, as for fileutils. * src/md5sum.c (digest_check): On systems for which setmode actually does something, arrange to read the file containing checksum strings in text mode. Based on a patch from Chris Faylor. * src/system.h (SET_MODE): Define. 2001-09-12 Jim Meyering The command `echo a|./fmt -2147483647' would cause fmt to segfault. * src/fmt.c (fmt_paragraph): Test for sentinal directly, rather than doing arithmetic with it's potentially large (INT_MAX) length. * tests/fmt/basic: Add tests for the above. tail: accept new option: -F, equivalent to `--follow=name --retry', for compatibility with FreeBSD and NetBSD versions of tail. * src/tail.c (usage): Describe new option. (parse_options): Accept it. Patch by Christian Kurz, via Herbert Xu. The command `echo foo|fmt -w N' for N>=32K would cause fmt to segfault. * src/fmt.c [struct Word]: Declare members length, space, and line_length to be of type `int', not short. (MAXCOST): Define using TYPE_MAXIMUM. Reported by Herbert Xu. 2001-08-31 Jim Meyering * src/fmt.c (bool): Remove typedef, now that it's in system.h. (TRUE): Define to `true', not 1. (FALSE): Define to `false', not 0. 2001-08-30 Paul Eggert * src/cat.c: Include full-write.h. (full_write): Remove decl; not needed. * src/split.c: Likewise. * src/cat.c (simple_cat, cat): Don't assume read and write size fits in int. 2001-08-28 Jim Meyering * src/fmt.c (main): Correct the diagnostic for invalid argument to -w. (main): Diagnose an invalid, old-style width option. 2001-08-27 Jim Meyering * Makefile.maint: Undo last change. Instead, simply remove `N_(', since the `Copyright ...' line should not be translated. 2001-08-26 Jim Meyering * Makefile.maint (copyright-check): Update to reflect fact that the copyright year is now defined separately. 2001-08-08 Charles Randall * src/sort.c (mergefps): Do not allocate at least sort_size bytes for each merge buffer. Instead, allocate at least sort_size bytes total. 2001-08-25 Jim Meyering * tests/uniq/Test.pm: Re-enable some commented-out tests. 2001-08-25 Paul Eggert * src/uniq.c (main): Fix a typo in the previous patch: missing a bounds check for examples like `uniq a b c'. * uniq.c (main): Fix some more incompatibilities with POSIX.2, (e.g. `uniq +3 --' did not work) by invoking getopt_long with leading '-', resembling what was done to 'sort' on 2001-03-20. Recognize an +N option only if it is an integer in range, and (if POSIXLY_CORRECT) only if a file name argument has not been seen; otherwise silently treat it as a file name. If the user specifies too many operands, output the first one in the error message, as a diagnostic aid. 2001-08-24 Paul Eggert * src/uniq.c (main): Don't report an error when given the valid options `+3 -d'. 2001-08-24 Paul Eggert * src/uniq.c (different): Don't assume that lengths can fit into size_t. Tune code for the common case where the line lengths differ: we avoid comparing them entirely in that case. 2001-08-25 Jim Meyering * tests/uniq/Test.pm: Add tests for -s 0 and -w 0. 2001-08-24 Paul Eggert Remove arbitrary restrictions on sizes, fixing a bug reported by Geoff Whale. * src/uniq.c (skip_fields, skip_chars, check_chars): Now size_t, not int. (size_opt): New function. Do not arbitrarily reject size zero. Change the wording of the error message slightly, for convenience. (find_field): Use size_t, not int, to compute sizes. (different, main): check_chars==0 is no longer a special case, as it defaults to SIZE_MAX. (main): Check for overflow with args like -234234234234234. Use 'size_opt' to convert optional arguments to sizes. 2001-08-25 Jim Meyering * tests/uniq/Test.pm: Add new test, #54, for the just-fixed bug. 2001-08-24 Paul Eggert * src/uniq.c (find_field): Don't count trailing newline to be part of the field. This disagrees with POSIX.2, but it's gotta be a bug in the standard. An interpretations request has been submitted to PASC. Reported by Geoff Whale. 2001-08-18 Jim Meyering * Makefile.maint: Merge in changes from autoconf. * Makefile.maint ($(v_etc_file)): Define. (copyright-check): Use the new variable. (my-distcheck): Depend on local-check so those checks are once again run by `make distcheck'. * src/head.c (head_bytes): Read no more than the specified number of bytes. Patch by Ian Bruce. * tests/misc/head-c: New test for the above. * tests/misc/Makefile.am: New file. * tests/misc: New directory. * tests/Makefile.am (SUBDIRS): Add misc. * configure.in (AC_OUTPUT): Add tests/misc/Makefile. 2001-08-14 Jim Meyering * src/Makefile.am (INCLUDES): Add -I../lib so sys2.h can include the new, generated file, unlocked-io.h. 2001-08-13 Jim Meyering * src/cat.c (AUTHORS): Mark string for translation, since it contains the English word `and'. * src/comm.c: Likewise * src/csplit.c: Likewise * src/cut.c: Likewise * src/md5sum.c: Likewise * src/nl.c: Likewise * src/paste.c: Likewise * src/pr.c: Likewise * src/sort.c: Likewise * src/split.c: Likewise * src/sum.c: Likewise * src/tac.c: Likewise * src/uniq.c: Likewise * src/wc.c: Likewise 2001-08-12 Jim Meyering * tests/mk-script ($program_name): Define. (main): Use $program_name, rather than long-inaccurate `build-script'. (main): Iterate over the sorted keys (rather than `while...each...'), so this script produces the same output regardless of which version of perl (e.g., perl5.6) is used. 2001-08-04 Jim Meyering * Makefile.am (EXTRA_DIST): Add Makefile.cfg. * man/Makefile.am (EXTRA_DIST): Remove duplicate definition. 2001-07-08 Jim Meyering * src/tail.c (usage): Reformat so help2man generates a properly indented man page. Patch by Herbert Xu. 2001-07-04 Jim Meyering * Makefile.cfg: New file with package-specific definitions. * Makefile.am (EXTRA_DIST): Add Makefile.cfg. 2001-07-01 Jim Meyering * src/sort.c (usage): Suggest setting LC_COLLATE=C, rather than LC_ALL=C. 2001-06-10 Jim Meyering * tests/sort/Test.pm: Add a test based on a report from Herbert Xu. 2001-05-20 Jim Meyering * src/uniq.c: Rename new option values: s/all/precede/ and s/minimum/separate/. * tests/uniq/Test.pm: Reflect the above renamings. 2001-05-12 Bruno Haible * src/tail.c (parse_obsolescent_option): Accept a b suffix. 2001-05-19 Jim Meyering msgmerge-0.10.37 complains about some `invalid multibyte sequences.' * po/el.po: Remove the offending entries. * po/ko.po: Likewise. * po/zh.po: Likewise. 2001-05-19 Jim Meyering Support new modes for uniq's --all-repeated option. The default behavior is unchanged. * src/uniq.c: Include argmatch.h. (usage): Update. (check_file): Implement it. (main): Handle new, optional arguments. * tests/uniq/Test.pm: New tests for the above. Patch by Padraig Brady. 2001-05-12 Bruno Haible * src/tail.c (parse_obsolescent_option): Use t_count_lines, not count_lines, in error message. 2001-05-05 Jim Meyering * src/comm.c (usage): Correct description of -3. Reported by Soeren Sonnenburg. 2001-04-22 Jim Meyering * Version 2.0.14. * configure: Regenerate using a patched copy of autoconf-2.49e to work around a bug in its test for a working memcmp function. 2001-03-28 Paul Eggert * src/cksum.c: Fix bugs when computing length of large files. Add overflow and write error checks. Use uint_fast32_t, not unsigned long, to do checksum calculation, as C99 says uint_fast32_t should be no slower and might be faster. Include and "system.h" even if CRCTAB is defined, so that the code will compile if CRCTAB is defined. Include "human.h" if CRCTAB is not defined. (uint_fast32_t): Define if it appears that stdint.h didn't. (BIT, remainder, main, crctab): Use uint_fast32_t, not unsigned long, for checksums. (fill_r, remainder, main): Use ANSI prototypes. (fill_r, remainder, main): Omit duplicate code. (main): Use uintmax_t, not unsigned long, for file lengths. Use size_t, not long, for result of fread. Check for overflow when computing file lengths. Check for write error immediately after printing a line, so that we don't write to stdout indefinitely after an error. 2001-04-14 Paul Eggert * src/sort.c (default_sort_size): Leave a 1/16 margin for RSS. Suggestion from Solar Designer. 2001-04-14 Jim Meyering * src/wc.c (usage): Tweak --help output: s/line,/newline,/ 2001-04-13 Jim Meyering * src/sort.c (main): Add a comment justifying the use of `multi-character' rather than `multibyte' in a diagnostic. From Paul Eggert. 2001-04-12 Paul Eggert * src/sort.c (RLIMIT_AS): Do not define; just use conditional code, since RLIMIT_RSS is similar (and is not standardized). (default_sort_size): Don't allocate more than the RSS limit, if this host has such a limit. 2001-03-20 Paul Eggert * NEWS, doc/textutils.texi: sort now accepts long options and checks POSIX option syntax. * doc/textutils.texi: Document --, -, sort long options, and sort -o after files. * src/sort.c: Include . (usage, main): Add support for long options, and check option syntax as POSIX requires, though (as usual for GNU apps) options can follow file names unless POSIXLY_CORRECT is set. Many diagnostic revamped. (long_options): New constant. (badfieldspec): New arg MSGID. Mark as noreturn. (parse_field_count): New arg MSGID; if null, just return null on error. (new_key): Renamed from key_init. All callers changed. Now allocates the new key. 2001-03-18 Jim Meyering * Version 2.0.13. * tests/Fetish.pm: If the $DJDIR envvar is defined, set SHELL to $DJDIR/bin/bash.exe. Patch from Richard Dawe, based on a suggestion from Eli Zaretskii. 2001-03-17 Paul Eggert * src/sort.c, src/tail.c, src/uniq.c (usage): Warn that the +N form will be withdrawn. * doc/textutils.texi: Likewise. 2001-03-08 Paul Eggert * NEWS, doc/textutils.texi: Document pr changes. * src/pr.c: Include mbswidth.h. (standard_header, header, test_suite): Remove. (date_format, date_text, file_text, header_width_available): New vars. (long_options, main, init_header, usage): Add new -D or --date-format option. (CHARS_FOR_DATE_AND_PAGE, T_BUF_FMT, T_BUF_SIZE, NO_DATE): Remove. (init_header): Allow arbitrary width for date format. Change "Page %5d" to "Page %d", since the code no longer assumes fixed width. Do not assume that localtime succeeds. (init_header, print_header, usage): Do not truncate headers. (init_header, print_header): Defer width calculations until page is printed, since "Page 100000" is wider than "Page 1". Count columns, not bytes, in page headers. Custom headers take up only the center, not the whole header. (print_header): Use printf rather than fprintf(stdout). * po/Makefile.in.in (install-data-yes): Install LC_TIME locale for all packages, not just for fileutils. (uninstall): Remove LC_TIME locale too. * src/sys2.h (INT_STRLEN_BOUND): New macro. * tests/pr/0F, tests/pr/0FF, tests/pr/2-S_f-t_notab, tests/pr/2-Sf-t_notab, tests/pr/2f-t_notab, tests/pr/2s_f-t_notab, tests/pr/2s_w60f-t_nota, tests/pr/2sf-t_notab, tests/pr/2sw60f-t_notab, tests/pr/2w60f-t_notab, tests/pr/3-0F, tests/pr/3-5l17f-t, tests/pr/3-FF, tests/pr/3a2l17-FF, tests/pr/3a3f-0F, tests/pr/3a3l15-t, tests/pr/3a3l8f-t, tests/pr/3b2l17-FF, tests/pr/3b3f-0F, tests/pr/3b3f-0FF, tests/pr/3b3f-FF, tests/pr/3b3l15-t, tests/pr/3b3l8f-t, tests/pr/3f-0F, tests/pr/3f-FF, tests/pr/3l17f-t, tests/pr/3l24-t, tests/pr/3ml17f-t, tests/pr/3ml24-FF, tests/pr/3ml24-t, tests/pr/3ml24-t-FF, tests/pr/4-7l24-FF, tests/pr/4l24-FF, tests/pr/a2l15-FF, tests/pr/a2l17-FF, tests/pr/a3-0F, tests/pr/a3f-0F, tests/pr/a3f-0FF, tests/pr/a3f-FF, tests/pr/a3l15-t, tests/pr/a3l17f-lm, tests/pr/a3l8f-t, tests/pr/b2l15-FF, tests/pr/b2l17-FF, tests/pr/b3-0F, tests/pr/b3f-0F, tests/pr/b3f-0FF, tests/pr/b3f-FF, tests/pr/b3l15-t, tests/pr/b3l17f-lm, tests/pr/b3l8f-t, tests/pr/FF, tests/pr/Ja3l17f-lm, tests/pr/Jb3l17f-lm, tests/pr/Jml17f-lm-lo, tests/pr/l17f-t, tests/pr/l24-FF, tests/pr/l24-t, tests/pr/ml17f-0F, tests/pr/ml17f-lm-lo, tests/pr/ml17f-t, tests/pr/ml17f-t-0F, tests/pr/ml20-FF-t, tests/pr/ml24-FF, tests/pr/ml24-t, tests/pr/ml24-t-FF, tests/pr/n+2-5l17f-0FF, tests/pr/n+2l17f-0FF, tests/pr/n+2l17f-bl, tests/pr/n+3-7l24-FF, tests/pr/n+3l17f-0FF, tests/pr/n+3l17f-bl, tests/pr/n+3ml13f-bl-FF, tests/pr/n+3ml17f-bl-tn, tests/pr/n+3ml17f-tn-bl, tests/pr/n+4-8a2l17-FF, tests/pr/n+4b2l10f-0FF, tests/pr/n+5-8b3l10f-FF, tests/pr/n+5a3l6f-0FF, tests/pr/n+6a2l17-FF, tests/pr/n+6b3l6f-FF, tests/pr/n+7l24-FF, tests/pr/n+8l20-FF, tests/pr/nJml17f-lmlmlo, tests/pr/nJml17f-lmlolm, tests/pr/nl17f-bl, tests/pr/nN1+3l17f-bl, tests/pr/nN15l17f-bl, tests/pr/nSml13-bl-FF, tests/pr/nSml13-t-t-FF, tests/pr/nSml13-t-tFFFF, tests/pr/nSml17-bl-FF, tests/pr/nSml17-t-t-FF, tests/pr/nSml17-t-tFFFF, tests/pr/o3a3l17f-tn, tests/pr/o3a3Sl17f-tn, tests/pr/o3a3Snl17f-tn, tests/pr/o3b3l17f-tn, tests/pr/o3b3Sl17f-tn, tests/pr/o3b3Snl17f-tn, tests/pr/o3Jml17f-lm-lo, tests/pr/o3ml17f-bl-tn, tests/pr/o3mSl17f-bl-tn, tests/pr/o3mSnl17fbltn, tests/pr/Test.pm, tests/pr/W-72l17f-ll, tests/pr/W20l17f-ll, tests/pr/W26l17f-ll, tests/pr/W27l17f-ll, tests/pr/W28l17f-ll, tests/pr/W35a3l17f-lm, tests/pr/W35b3l17f-lm, tests/pr/W35Ja3l17f-lm, tests/pr/W35Jb3l17f-lm, tests/pr/W35Jml17f-lmlo, tests/pr/W35ml17f-lm-lo, tests/pr/W72Jl17f-ll, tests/pr/w72l17f-ll: Adjust to minor spacing changes in pr headers. 2001-03-11 Paul Eggert * src/sort.c (die): New message arg, to describe failures better. All callers changed. 2001-03-11 Paul Eggert * src/sort.c: Decrease buffer size when only merging or checking. (sort_size): Now the user-specified sort size. (MIN_MERGE_BUFFER_SIZE): New macro. (MIN_SORT_SIZE): Use it. (merge_buffer_size): New variable. (fillbuf): Increase merge_buffer_size if a longer line is encountered. (checkfp, mergefps): Do not allocate a buffer smaller than merge_buffer_size. (sort): Use the default_sort_size if sort_size is zero. (main): Do not set sort_size to default_sort_size. 2001-03-06 Paul Eggert * src/sort.c (sort): If all the input files are empty, create an empty output file. * tests/sort/Test.pm: Test for this. 2001-03-13 Jim Meyering * src/sort.c (merge): Move declarations of local variables into the inner scope where they're used. (sort): Likewise. 2001-03-06 Paul Eggert Avoid the need for a copy of the input file when the input and output overlap, e.g. 'sort F -o F'. With -m, though, a copy is still needed sometimes. * doc/textutils.texi: 'sort F -o F' no longer needs to copy F. * src/sort.c: Do not include closeout.h. (xfopen): Use stdout if *how != 'r'. (mergefps): Remove FPS arg. Open all input files, and close all files when done. If OFP is null, open the output file (but after opening input files). All callers changed. (first_same_file): New function. (sort, merge): Remove arg OFP; we now open the output file as needed. All callers changed. (merge): New arg MAX_MERGE. All callers changed. (sort): For "sort F -o F", close the input before opening the output. (main): Do not use close_stdout; 'sort' and 'merge' now close stdout. (This also fixes a close-stdout-twice bug.) Remove test for overlapping input and output files, as 'sort' no longer needs to worry about overlap, and 'merge' checks for overlap itself. Use first_same_file to inform 'merge' about how much to merge at the top level, to avoid overlap. 2001-03-12 Jim Meyering `fmt --prefix=S' would not work properly for any string S containing a byte with the high bit set. * src/fmt.c (prefix): Declare to be of type unsigned char, not `char'. (get_prefix): Likewise for local, `p'. Reported by François Pinard. Add a test for the above-fixed problem. * tests/fmt: New directory/files. * configure.in (AC_OUTPUT): Add tests/fmt/Makefile. * tests/Makefile.am (SUBDIRS): Add fmt. 2001-03-05 Andreas Schwab * src/tac.c (save_stdin): Use mkstemp to create temporary file. 2001-03-09 Jim Meyering * Makefile.maint (my_distdir): Define new variable, and use this in place of most old uses of $(distdir). * tests/help-version: Ensure that /dev/full is a character device (using test -c) as well as being writable, before trying to write to it. Otherwise, the test could mistakenly append a newline to an existing, regular, writable, /dev/full file. Suggested by Ulrich Drepper. 2001-03-08 Jim Meyering * tests/sha1sum/sample-vec: Insert the `--text' argument for each test. Reported by Matthew Smith. 2001-03-04 Jim Meyering * Makefile.maint (my_distdir): Define. Use it in place of $(distdir) almost everywhere. * Version 2.0.12. * Makefile.maint (alpha): Use $(PACKAGE)-$(VERSION), not $(distdir), since the latter now has a `$(top_distdir)/' prefix. 2001-03-01 Paul Eggert * src/sort.c (die): New function. (create_temp_file, xfopen, xfclose, write_bytes, sort_buffer_size, fillbuf, main): Use it to regularize error messages. The only change in behavior is that write_bytes and the final close used to say "write error" but now give just the output file name, which should be enough. 2001-03-01 Paul Eggert * src/sort.c (xfclose): Add FILE arg, and report the file name on error. All callers changed. 2001-03-01 Paul Eggert * src/sort.c (main): When fclose (stdin) fails, do not mention the output file in the error message; mention "-" instead. 2001-02-28 Paul Eggert * src/sort.c (xfopen): Set have_read_stdin to 1 only if file is "-". Use fopen_safer, not fopen, to avoid subtle bugs when fopen returns stdin, stdout, or stderr. (xfclose): stdout is no longer a special case. (main): Close output file, don't just flush it; there might be an error on the close. 2001-02-27 Paul Eggert * src/sort.c (initbuf): If the desired size cannot be allocated, repeatedly halve it until allocation succeeds. Reported by Solar Designer. 2001-02-26 Paul Eggert * src/sort.c (parse_field_count): Comment fix. 2001-02-25 Paul Eggert * src/sort.c: Tune allocation and comparison of nodes representing temp files. This improved CPU performance of 'sort -S 1 *.[ch]' by 17% on my host. (struct tempnode): name member now uses struct hack. (temphead): Now a pointer, not a structure. All uses changed. (create_temp_file): Allocate node using struct hack. (zaptemp): Free node using struct hack. Use pointer comparison, not string comparison. 2001-02-25 Paul Eggert * src/sort.c (eolchar, trim_trailing_blanks): Now static. 2001-02-25 Paul Eggert 'sort' race condition fixes. Defend against a DoS attack where someone else creates a temporary file with the same name as ours. Use mkstemp to do this, supplying our own mkstemp if the system doesn't have one. Also, fix a race condition during cleanup on hosts without sigaction. * src/sort.c (NAME_MAX_IN_DIR): Remove. (sigprocmask, sigset_t): New macros, defined only on older hosts. (caught_signals): New var. (xtmpfopen, tempname): Removed. (create_temp_file): New function, combining the functions of the old xtmpfopen and tempname. All callers changed. Use mkstemp to create the file. (sighandler): On hosts without sigaction, ignore signals while cleaning up, instead of letting them interrupt cleanup. (main): Initialize caught_signals. On hosts with sigaction, block all caught signals while handling one. Remove duplicate code. 2001-02-22 Jim Meyering * src/sort.c (keycompare): Move declarations of locals, lena and lenb, into the inner scope where they are used. 2001-02-19 Jim Meyering * src/sort.c (AUTHORS): Add Paul Eggert. * src/wc.c (wc): Rename innermost `buf' to avoid shadowing warning. (wc): Rename local `wc' to avoid shadowing function name. 2001-02-18 Paul Eggert Fix a race condition: freed storage accessed during a signal handler. * src/sort.c (struct tempnode.next): Now volatile. (zaptemp): Free the file name after removing it from the temp list, not before, because a signal can arrive between the two actions and cleanup () traverses the list. 2001-02-18 Paul Eggert Check for input size, and do not overallocate memory. Also check for memory quotas. Revamp storage management so that line tables and character data are taken from the same buffer. Line tables are now in reverse order, since they grow down while the character data grow up. * src/sort.c: (): Include if HAVE_SYS_RESOURCE_H. (struct rlimit, getrlimit): Define a replacement if RLIMIT_DATA is not defined. (RLIMIT_AS): Define to RLIMIT_DATA if not defined. (struct lines): Remove. (struct buffer): New members nlines, line_bytes, eof. Remove member newline_free; no longer needed, since the code no longer runs out of line table space. (SORTALLOC_MIN, SORTALLOC_DEFAULT_MIN): Remove. (sort_size): Renamed from sortalloc; now applies to the sum of the character data and the line table, not just the character data. (MIN_SORT_SIZE, INPUT_FILE_SIZE_GUESS): New macros. (linelength): remove. (specify_sort_size): Don't worry about the distinction between the character data and the line table; that is now the caller's responsibility. (default_sort_size): Return the value, instead of being executed for side effect. Return half of available memory, or 1/16 of total memory, whichever is greater; except do not exceed 1/2 of quota. (sort_buffer_size): New function. (initbuf): New arg LINE_BYTES. Ensure that the line array is properly aligned. Initialize the new set of struct buffer members. (buffer_linelim): New function. (fillbuf): Return int, not size_t, since the callers merely care whether the result is nonzero. New arg FILE so that error messages can report the file name. Keep track of eof. Initialize the line table too, taking its memory from the input buffer's memory; this subsumes the old findlines function and removes the need for worrying about running out of line table entries. (checkfp, mergefps, sortlines, merge, sort): Adjust to the new storage management regime, in particular the fact that line tables are now filled in by fillbuf and are in reverse order. (checkfp): Now takes char *, not const char *, since subroutines require that now. Rewrite to avoid lint and duplicate code. If line length alloc calculation overflows, simply allocate enough memory to hold the line. (mergefps): New arg FILES, used for buffer size calculation and error messages. Rewrite to avoid lint. Do not loop if savealloc*2 overflows. (mergefps, merge): Zap temporary files eagerly rather than lazily; this is needed because we now pass FILES to mergefps. (sortlines): Args now point at end of arrays, not at beginnings. (sort): Do not allocate temporary line array for sortlines; instead, take the space from the same buffer. (main): Adjust to sort_size and default_sort_size changes. 2001-02-18 Jim Meyering Rename test input files to avoid conflicts on case-insensitive file systems. * tests/pr/2-Sf-t_notab: Rename from 2Sf-t_notab. * tests/pr/2-S_f-t_notab: Rename from 2S_f-t_notab. * tests/pr/W-72l17f-ll: Rename from W72l17f-ll. * tests/pr/Test.pm: Update file names to reflect renamings. Reported by Matthew Smith. 2001-01-20 Jim Meyering * configure.in: Remove jm_CHECK_ALL_TYPES. Now it's invoked by jm_MACROS. 2001-01-17 Jim Meyering * src/cksum.c (main): Use PACKAGE, not GNU_PACKAGE. * src/tsort.c (main): Likewise. * src/sort.c (main): Likewise. (usage): Convert each TAB in --help output to a sequence of 8 spaces. 2001-01-07 Jim Meyering * src/tail.c (usage): Split a string that was longer than 2048 bytes. 2001-01-03 Jim Meyering * src/sort.c (main): Remove embedded \n from diagnostic. 2001-01-02 Jim Meyering * src/od.c (ulonglong_t): Define place-holder type to avoid some #if directives. (LONGEST_INTEGRAL_TYPE): Remove definition. (MAX_INTEGRAL_TYPE_SIZE): Use ulonglong_t instead of LONGEST_INTEGRAL_TYPE. (print_long_long): Compile this function even on systems without long long support. (decode_one_format): Remove #if directive. * src/od.c (decode_one_format): Guard use of print_long_long with `#if HAVE_UNSIGNED_LONG_LONG'. From Darren Salt. Change all `#ifdef HAVE_UNSIGNED_LONG_LONG' to use `#if' instead. 2000-12-23 Jim Meyering * src/sys2.h [HAVE_INTTYPES_H]: Include . 2000-12-19 Jim Meyering * Version 2.0.11. 2000-12-18 Paul Eggert * NEWS, doc/textutils.texi: New "sort" option -S SIZE. * src/sys2.h (UINTMAX_MAX): New macro, taken from C99. * src/sort.c: Include physmem.h. (SORTALLOC, mergealloc, LINEALLOC): Remove. (sortalloc): Default to zero at program startup. (SORTALLOC_MIN, SORTALLOC_DEFAULT_MIN): New macros. (usage, main): Add support for new -S SIZE option. (specify_sort_size, default_sort_size): New functions. (initlines): Do not let alloc exceed limit. (findlines): Likewise. (checkfp, mergefps, sort): Use sortalloc to size everything else, instead of relying on precomputed sizes. 2000-12-17 Jim Meyering * doc/texinfo.tex: Update from master repository. * config.sub: Likewise. * config.guess: Likewise. 2000-12-11 Jim Meyering * Version 2.0.10. 2000-12-07 Jim Meyering * src/od.c (address_base): Declare to be static. 2000-12-06 Paul Eggert * src/od.c (address_base, address_pad_len): New var. (output_address_fmt_string, address_fmt_buffer, address_pad): Remove. (flag_pseudo_start): Now int, not long int. (pseudo_offset): Now off_t, not long int. (n_specs, n_specs_allocated): Now size_t, not unsigned int. (format_address, format_address_none, format_address_std, format_address_label): Now accepts an extra char argument (an extra char to print if nonzero), and prints instead of returning a string. All callers changed. (bytes_per_block): Now size_t, not int. (format_address_none): Do not even print the extra char argument. This simplifies the callers. (format_address_std, format_address_label): Print off_t ourself instead of trying to use autoconfigured format. This is faster and more portable. (format_address_paren): New function. (dump): Remove unnecessary cast. (expand_address_fmt): Remove. (main): Use size_t, off_t, etc. instead of builtin types where this is advisable. Adjust to above changes. Remove unnecessary cast. 2000-12-03 Jim Meyering * src/tail.c (tail_file): Initialize ignore, dev, and ino members, when tailing forever and the open failed. Otherwise, we could get uninitialized memory references of those fields in recheck. * tests/tail-2/Makefile.am (TESTS): Add assert-2. * tests/tail-2/assert-2: New file. * Version 2.0.9. Make od print valid addresses for offsets of 2^32 and larger, and allow byte offset (-j) and byte count (-N) to be 2^32 and larger. * src/od.c (MAX_ADDRESS_LENGTH): Don't hard-code as a literal. Rather, define in terms of the type, off_t. (string_min): Declare to be of type size_t. (flag_dump_strings): Declare to be of type int. (print_s_char): Declare the n_bytes parameter and the local, `i', to be of type off_t. (print_char): Likewise. (print_s_short): Likewise. (print_short): Likewise. (print_int): Likewise. (print_long): Likewise. (print_long_long): Likewise. (print_float): Likewise. (print_double): Likewise. (print_long_double): Likewise. (dump_hexl_mode_trailer): Likewise. (print_named_ascii): Likewise. (print_ascii): Likewise. (write_block): Likewise. (print_ascii): Declare local, `print_function' with a prototype. Change a few `>' comparisons to the equivalent `<' form. (parse_options): Declare `tmp' to be of type uintmax_t. Use xstrtoumax, not xstrtoul. Fail if the specified offset if larger than OFF_T_MAX. (dump_strings): Declare local `i' to be of type size_t. Remove the now-unnecessary cast-to-off_t. (main) [IF_LINT]: Initialize desired_width to avoid a warning. Declare `tmp' to be of type uintmax_t. Use xstrtoumax, not xstrtoul. Fail if minimum string length is larger than SIZE_MAX. Fail if specified width is larger than ULONG_MAX. * src/od.c (format_address): Use off_t, not long unsigned_int as the parameter type. (format_address_none): Likewise. Mark parameter as unused. (format_address_std): Likewise. (format_address_label): Likewise. (print_ascii): Mark format string parameter as unused. (write_block): Use off_t, not long unsigned_int as offset type. (expand_address_fmt): New function. (main): Use it to expand each address format string template. Reported by Mark Nudelman, via Andreas Jaeger. * src/sys2.h (OFF_T_MIN): Define here instead. (OFF_T_MAX): Likewise. (CHAR_BIT): Define. * src/tail.c (parse_options): Use xstrtoumax to parse the byte and line offset. Give a better diagnostic when the requested offset is still representable but larger than OFF_T_MAX. (OFF_T_MIN): Remove definition. (OFF_T_MAX): Likewise. 2000-12-02 Jim Meyering * src/sort.c (checkfp): Rename local `buf' to avoid shadowing previous declaration. * src/sort.c (NONZERO): Define and use it to make the code a tiny bit more readable. * doc/textutils.texi (sort invocation): Clarify how -t works when a sort key specifies a range of fields. From Karl O. Pinc. 2000-11-26 Paul Eggert * src/od.c (skip): Use lseek instead of worrying about fseeko or fseek. This should be portable, as we seek before doing any I/O. (fseeko): Remove; no longer used. 2000-11-30 Jim Meyering * src/sort.c: s/SIZE_T_MAX/SIZE_MAX/. 2000-11-30 Paul Eggert * src/sys2.h: Include if HAVE_STDINT_H. (SIZE_MAX): Renamed from SIZE_T_MAX, as C99 uses SIZE_MAX. All uses changed. 2000-11-30 Jim Meyering * src/sort.c: SIZE_MAX is not defined, so s/SIZE_MAX/SIZE_T_MAX/, and... * src/sys2.h (SIZE_T_MAX): ... define. 2000-11-29 Paul Eggert Port GNU "sort" to hosts where sizes don't fit in "int", e.g. 64-bit Solaris (sparc). * src/sort.c ("human.h", "xstrtol.h"): Include. (struct line): length member is now size_t, not int. (struct lines): Likewise for used, alloc, limit members. (struct buffer): Likewise for used, alloc, left, newline_free members. (struct keyfield): Likewise for sword, schar, eword, echar members. (sortalloc, mergealloc, linelength): Now size_t, not int. (initbuf, fillbuf, initlines, begfield, limfield, findlines, numcompare, getmonth, keycompare, compare, checkfp, mergefps, sortlines, sort): Accept, return, and use size_t for sizes, not int. (fillbuf, initlines, findlines, checkfp, sort): Check for overflow when computing buffer sizes. (begfield, limfield): Do not index past end of array. (checkfp): Return a boolean, not a line number, as the line number may not fit in int. All callers changed. Use uintmax_t for line numbers, not int. (sort): Don't allocate tmp until we need it (and know the right size). (parse_field_count): New function. (main): Use it to check for overflow in field counts. "outfile" is now a pointer to const. 2000-11-27 Jim Meyering * src/checksum.h: Don't include system.h here. * src/md5.c: Include config.h, stdio.h, sys/types.h. and system.h here instead. * src/sha1sum.c: Likewise. 2000-11-18 Paul Eggert * src/pr.c (main): Do not assume EOF == -1. Handle the case correctly when digits options immediately precede a non-option. 2000-11-18 Jim Meyering * doc/textutils.texi: Change GNU to @sc{gnu} in many places. Update to use the Free Documentation Licence. * configure, config.h.in, Makefile.in, etc.: Regenerate using the very latest version (in CVS) of autoconf. 2000-11-17 Jim Meyering * GNUmakefile (.NOTPARALLEL): New target. Prevent unwanted parallelism. Suggestion from Ulrich Drepper. 2000-11-16 John David Anglin * tsort.c: Include sys/types.h before system.h. 2000-11-16 Jim Meyering * src/join.c: Declare global `tab' to be of type *unsigned* char, so join works with 8-bit delimiter characters. Reported by Antonio Rendas. * tests/join/Test.pm (8-bit-t): New test for the above fix. 2000-11-13 Paul Eggert * src/sys2.h (ST_TIME_CMP_NS): Fix typo: ST_MTIM_NSEC should be tested with #ifdef, not with #if. 2000-11-11 Jim Meyering * doc/textutils.texi: More minor rewording and grammar correction. From Brian Youmans. 2000-11-06 Paul Eggert * src/od.c (skip): st_size is not defined for directories. 2000-11-03 Jim Meyering * tests/sha1sum/sample-vec: New file, with tests derived from http://csrc.nist.gov/cryptval/shs/sha1-vectors.zip * tests/sha1sum/Makefile.am (TESTS): Add sample-vec. 2000-10-30 Jim Meyering * configure, config.h.in, Makefile.in, etc.: Regenerate using the very latest version (in CVS) of autoconf. * src/wc.c: Include "system.h" after wctype.h (not before) to avoid a compilation error on Solaris 5.5.1 systems. Reported by Vin Shelton. 2000-10-23 Jim Meyering * Makefile.maint: Clean up version-related variables. * Version 2.0.8. * Makefile.am (EXTRA_DIST): Add .prev-version. 2000-10-22 Jim Meyering * tests/Makefile.am (SUBDIRS): Add sha1sum. * tests/sha1sum/basic-1: New file. * configure.in (AC_OUTPUT): Add tests/sha1sum/Makefile. * src/Makefile.am (bin_PROGRAMS): Add sha1sum. (md5sum_SOURCES): Define. (sha1sum_SOURCES): Define. (noinst_HEADERS): Add checksum.h. * src/md5sum.c: Factor out the differences between MD5 and SHA1, and parameterize so this code may be used by both md5sum and the new program, sha1sum. Loosely based on a patch from Scott Miller. * src/checksum.h: New file. * src/md5.c: New file that simply defines `algorithm'. * src/sha1sum.c: Likewise. * man/Makefile.am (man_MANS): Add sha1sum.1. * man/sha1sum.x: New file. Support 8-byte integers, assuming they're printable with e.g., %lld. * src/od.c: Add support for printing data as unsigned long long integers. 2000-10-21 Jim Meyering The command, `yes ''|./cat -n' would stop printing after INT_MAX lines. * src/cat.c (cat): Never let `newlines' exceed 3. 2000-10-21 Jim Meyering Prevent a counter buffer overrun when numbering lines and when processing 100 billion lines (or more) of input. * src/cat.c (LINE_COUNTER_BUF_LEN): Define to allow numbering as many as 10^18 - 1 lines (the old limit was 10^11 - 1, and could be exceeded without too much trouble). Use this symbol rather than hard-coding the constant everywhere. (next_line_num): Rather than overrunning for input with more lines, mark the line number by putting a `>' in the leftmost slot. Patch by Jan Nieuwenhuizen. * src/sort.c (SORT_OUT_OF_ORDER): Define. (main): Use it instead of hard-coding the `1'. 2000-10-10 Jim Meyering * src/sort.c (main): Use EXIT_SUCCESS rather than 0. Fail when checking (-c) with more than one file argument, rather than simply ignoring the extra arguments. 2000-09-07 Jim Meyering * src/cut.c: Remove obsolete comment block. (usage): Note that when using -f, lines with no delimiters are also printed. * doc/textutils.texi (cut invocation): Likewise. 2000-08-23 Jim Meyering * src/md5sum.c: Include unconditionally, to be consistent with all the other programs in this directory. * src/tsort.c: Likewise. 2000-08-19 Jim Meyering * src/comm.c (writeline): Correct comments. From Bruno Haible. 2000-08-11 Paul Eggert * doc/textutils.texi: Portable shell scripts should specify global options before key fields. Move global LC_CTYPE remark to each sort option that depends on LC_CTYPE. sort -g depends on LC_NUMERIC. Add @vindex where it's missing. "radix character" -> "decimal-point character", to match Standard C terminology, which is easier for most people to follow. "comm" does not consider trailing newlines to be significant. 2000-08-10 Paul Eggert * doc/textutils.texi: Recommend against the System V syntax for tr ranges, and don't use it in examples. Use POSIX classes rather than ranges, for portability. * src/tr.c (usage): Don't describe System V syntax, as it doesn't always work. * src/sort.c (usage): Describe -d and -i in a locale-independent way. * doc/Makefile.am (constants.texi): Use the C locale so that [A-Z] works as expected. 2000-08-07 Paul Eggert * src/cut.c (cut_fields): Invoke xalloc_die instead of printing our own message. (cut_fields): Check for I/O error as well as end-of-file 2000-08-06 Bruno Haible * src/wc.c: Add support for multi-byte locales. (iswprint, mbrtowc, wcwidth): Provide default definitions. (total_bytes): New variable. (print_bytes): New variable. (longopts): Change abbreviation for --chars from 'c' to 'm'. (usage): Update. (write_counts): Add `bytes' argument. (wc): New variables `bytes', `count_bytes', `count_chars', `count_complicated'. The old code determines `bytes', not `chars'. New case for MB_CUR_MAX > 1. A non-printable non-space character does not increment the line position or start a word. Update `total_bytes'. (main): Initialize `print_bytes' and `total_bytes' to 0. Accept 'm' option. Pass `total_bytes' to write_counts. * doc/textutils.texi (wc invocation): Update accordingly. 2000-08-06 Bruno Haible * src/head.c (head): Call write_header here. (head_file): ... not here. * src/cat.c (closeout_func): Remove variable. (close_stdout_wrapper): Remove unused function. (main): Remove assignment to closeout_func. * src/fold.c (fold_file): Remove ferror(stdout) check, already done in close_stdout. 2000-08-06 Paul Eggert * src/sort.c (usage): Warn more succinctly about the effects of the locale on sort order. 2000-08-06 Jim Meyering * Version 2.0g. * src/tail.c (pipe_lines): Declare local `cp' to be const. 2000-08-04 Greg McGary * src/tail.c (pipe_lines): Add variable `nbytes' so we can free `tmp' immediately after read loop. Don't process an empty file. [This fixes a buffer-underrun error -- detected thanks to bounded pointers -jmm] 2000-08-05 Jim Meyering Fix buffer-overrun error. * src/pr.c (init_header): Allocated buffer was sometimes too small. Reported by Greg McGary (who found this bug using his bounded-pointers-enabled gcc). (init_header): Move declarations of several variables into the inner scope where they are used. * tests/pr/Test.pm (narrow-1): New test, for the above. * src/csplit.c (MAX): Remove definition (It's in sys2.h). 2000-08-04 Jim Meyering * src/sort.c (main): Rename local `t' to `tmp_dir' to avoid shadowing a previous local by that name. (usage): Warn that GNU sort is now locale-aware, and suggest people put LC_ALL=POSIX in their environment. * src/pr.c (store_columns): Remove conjunct that would dereference an out-of-bounds pointer. Reported by Greg McGary (who found this bug using his bounded-pointers-enabled gcc). 2000-08-03 Jim Meyering * tests/Makefile.am.in ($(srcdir)/$x-tests): Use -I$(srcdir), not `-I.'. (Makefile.am): Likewise. Reported by Greg McGary. 2000-08-01 Jim Meyering * doc/textutils.texi (tr invocation): Note that ranges are not portable. Update examples not to use ranges. 2000-07-31 Jim Meyering * src/cut.c (cut_fields): Use `virtual memory exhausted', not `Memory exhausted'. 2000-07-30 Jim Meyering * src/help-version: Remove file. Move it to ... * tests/help-version: ...here * tests/Makefile.am (TESTS): Set to help-version. (TESTS_ENVIRONMENT): Define. (EXTRA_DIST): Add $(TESTS). * src/Makefile.am (EXTRA_DIST): Remove help-version. (TESTS): Remove definition. (TESTS_ENVIRONMENT): Remove definition. 2000-07-28 Paul Eggert sort's -T option can now appear multiple times. Thanks to Charles Randall for suggesting this idea. * doc/textutils.texi, NEWS: Document this. * src/sort.c (temp_dir): Remove. (temp_dirs, temp_dir_count, temp_dir_alloc): New vars. (process_id): New var. (usage): Describe new use of -T. (add_temp_dir): New function. (tempname): Use new temp_dirs array. Do not discard information from the process-id or sequence number, unless we have short file names. (sighandle): Use process_id instead of getpid. (main): Initialize process_id. Add support for the new use of -T. 2000-07-23 Jim Meyering * tests/head/Test.pm: Rename tests (and hence files) to have a short enough prefix so that they're not truncated on 8+3 file systems. Reported by Eli Zaretskii. 2000-07-09 Jim Meyering * config.guess: Update from FSF. * config.sub: Likewise. 2000-07-02 Jim Meyering * Version 2.0f. 2000-07-01 Jim Meyering * src/cut.c: Change type of global, field_1_bufsize, from int to size_t. 2000-06-29 Jim Meyering * src/tr.c (find_bracketed_repeat): Add a cast to suppress a warning. Reported by Gerhard Poul. * src/tail.c (enum): Remove comma from end of enumerator list. From Gerhard Poul. 2000-06-25 Jim Meyering * configure.in (ALL_LINGUAS): Add Danish (da). 2000-06-24 Jim Meyering * doc/textutils.texi: Lots of minor rewording and grammar correction. From Brian Youmans. * tests/sum/basic-1: New file. * tests/sum/Makefile.am: New file. * configure.in (AC_OUTPUT): Add tests/md5sum/Makefile. * tests/Makefile.am (SUBDIRS): Add sum. 2000-06-22 Bruno Haible * src/sum.c (sysv_sum_file): Avoid overflowing 32-bit accumulator on files whose bytes sum to 2^32 or larger. The smallest such file contains 16,843,010 bytes, nearly all of which have the value 0xff. 2000-06-22 Bruno Haible * src/cat.c (main): Correct a comment. 2000-06-21 Jim Meyering * src/cut.c (getstr): Remove private (and out of date) version of this function. (cut_fields): Adjust caller to use the just-extended one in ../lib. Reported by Paul Sauer. 2000-06-20 Jim Meyering * src/uniq.c (main): Fix off-by-argc test, so +N-style options are recognized once again. Reported by Geoff Keunning. Fix typo in diagnostic: s/compare/skip/. 2000-05-27 Jim Meyering * src/tail.c: Arrange to call close_stdout upon exit. Don't close stdout explicitly. (usage): Add missing backslash at end of line. (write_header): Remove now-unused COMMENT parameter. Update all callers. 2000-05-24 Jim Meyering * src/head.c (head_file): Use STDIN_FILENO in place of `0'. (main): Likewise. 2000-05-20 Jim Meyering * src/cat.c: Arrange to call close_stdout upon exit. Don't close stdout explicitly. Replace uses of global constant, output_desc, with uses of STDOUT_FILENO. Arrange to call close_stdout upon exit. Don't close stdout explicitly. * src/cksum.c: Likewise. * src/comm.c: Likewise. * src/csplit.c: Likewise. * src/cut.c: Likewise. * src/expand.c: Likewise. * src/fmt.c: Likewise. * src/fold.c: Likewise. * src/head.c: Likewise. * src/join.c: Likewise. * src/md5sum.c: Likewise. * src/nl.c: Likewise. * src/od.c: Likewise. * src/paste.c: Likewise. * src/pr.c: Likewise. * src/ptx.c: Likewise. * src/sort.c: Likewise. (but set exit status and file name, too) * src/split.c: Likewise. * src/sum.c: Likewise. * src/tac.c: Likewise. * src/tr.c: Likewise. * src/tsort.c: Likewise. * src/unexpand.c: Likewise. * src/uniq.c: Likewise. * src/wc.c:: Likewise. 2000-05-13 Jim Meyering * src/help-version: Run each program successfully at least once. 2000-05-12 Jim Meyering * src/head.c (string_to_integer): Restrict base to 10. Reported by Joseph S. Myers. * tests/head/Test.pm: Add a few tests for this. New option: --first-only * src/unexpand.c (anonymous enum) [CONVERT_FIRST_ONLY_OPTION]: Define. (long_options): Add `first-only'. (main): Handle new option. 2000-05-11 Jim Meyering * tests/unexpand/basic-1: New tests. 2000-04-17 Bruno Haible * src/system.h [__BEOS__]: Ignore O_BINARY and O_TEXT. 2000-04-16 Jim Meyering * src/tail.c: Prepare to remove option: --max-consecutive-size-changes. Doesn't seem useful. (usage): Remove description. * doc/textutils.texi (tail invocation): Remove description. 2000-04-12 Jim Meyering * src/tail.c (usage): Tweak --help output. Suggestions from Karl Berry. Move some macros into m4/. * configure.in (_GNU_SOURCE): Don't define here. (AC_SYS_LARGEFILE): Don't use here. (AM_C_PROTOTYPES): Don't use here. * acconfig.h: Remove now-unused file. 2000-04-09 Jim Meyering * doc/textutils.texi: Include constants.texi. (tail invocation): Use `@value's instead of referring to the output of `tail --help' for defaults. * doc/Makefile.am (constants.texi): New rule. (EXTRA_DIST): Add constants.texi. (MAINTAINERCLEANFILES): Add it here, too. (textutils.dvi, textutils.info): Depend on constants.texi. * src/tail.c (usage): Clarify descriptions of the --max-* options. From Karl Berry. 2000-04-03 Jim Meyering * configure.in (AC_CHECK_HEADERS): Remove langinfo.h. Now it's in m4/. (AC_CHECK_FUNCS): Remove nl_langinfo. Now it's in m4/. 2000-03-17 Jim Meyering * configure.in (ALL_LINGUAS): Add Slovenian (sl). 2000-03-12 Jim Meyering Merge from fileutils. * src/system.h "pathmax.h": Include. : Include if appropriate. : Likewise. : Likewise. : Likewise. : Likewise. (major, minor, makedev): Define if needed. (struct utimbuf): Declare if missing. (ST_BLKSIZE, ST_NBLOCKS, ST_NBLOCKSIZE): Define. * src/cksum.c (usage): Correct --help output. Prompted by a report from Gregory Leblanc. 2000-03-05 Paul Eggert * src/sort.c (struct buffer.newline_free): New member. (initbuf, findlines): Set it. (fillbuf): Do not double the size of a full buffer to append a newline unless the buffer is known to be newline free. 2000-03-05 Jim Meyering * Version 2.0e. 2000-03-04 Jim Meyering Don't get failed assertion for `tail -f directory'. * src/tail.c: Revert most of 1999-10-20 change. Instead, ... (IS_TAILABLE_FILE_TYPE): Define to produce similar result. (recheck): Use it here. (tail_file): Use it here. Reported by François Pinard. Once we encounter a file that is not of IS_TAILABLE_FILE_TYPE, marke it as such and ignore it forever after. * src/tail.c (struct File_spec): New member. (recheck): Initialize new member. (tail_file): Likewise. (tail_forever): Skip the file if it's marked as ignorable. 2000-03-03 Jim Meyering * src/sort.c (fillbuf): Move declaration of local, cc, into scope of `while' loop where it's used. 2000-03-02 Paul Eggert Big performance improvement when sorting many small files, building on a suggestion by Charles Randall. * src/sort.c (fillbuf): Skip memmove if it would be a no-op, as many memmove implementations are slow in that case. Don't examine leftover bytes for eolchar, since they may be left over from a previous file, and we want to read from this file. (sort): At end of file, if there is more input and buffer room, concatenate the next input file. 2000-02-29 Jim Meyering * src/join.c (make_blank): Add an explicit cast to `unsigned char *' to placate HPUX's C compiler. Reported by Bob Proulx. 2000-01-30 Jim Meyering * Version 2.0d. * Makefile.maint (my-distcheck): Depend on po-check, so I'll be warned about any required changes at `make distcheck' time rather than at `make alpha' time. (writable-files): New target. (my-distcheck): Depend on it. (alpha): Likewise. * po/POTFILES.in: Add these: lib/makepath.c, lib/rpmatch.c, and lib/same.c 2000-01-29 Jim Meyering * configure.in: Remove lib-check for cposix now that we use m4/isc-posix.m4. (jm_LIB_CHECK): Use this even though the library checks aren't required for this patchage (so all three lib/Makefile.in files will be the same). 2000-01-24 Mark Kettenis * src/tsort.c (exit_status): New variable. (loop): New varibale. (count_items, scan_zeroes): Change return type to int. (detect_loop): Complete rewrite to correctly implement detection of loops. Also change return type to int. (recurse_tree): Stop if ACTION returns non-zero. This involves changing the return type of this function and ACTION to int. (walk_tree): Change return type of ACTION to int. (tsort): Continue sort after a loop has been detected (and broken). Set exit_status to 1 if a loop was detected. (main): Use exit_status to determine exit code. * tests/tsort/basic-1: Change expected output for cycle-1 and cycle-2 tests. 2000-01-23 Jim Meyering * configure.in (AM_FUNC_STRTOD): Move to m4/. (AC_SUBST(POW_LIBM)): Likewise. 2000-01-22 Jim Meyering * src/sort.c (keycompare): Use global, hard_LC_COLLATE in place of local that is sometimes undeclared. * configure.in (ALL_LINGUAS): Add Brazilian Portuguese (pt_BR). (AC_CHECK_HEADERS): Move these checks into m4/. * src/tr.c ("xstrtol.h"): Include. (Filter): Rename from PFL. (non_neg_strtol): Remove function. (find_bracketed_repeat): Use xstrtoul instead of non_neg_strtol. (squeeze_filter, set_initialize, main): Use size_t and ssize_t in place of long and int in several decls. (read_and_delete, read_and_xlate): Likewise, and remove assertion. * tests/tr/Test.pm (o-rep-1, o-rep-2): New tests. * src/cut.c: Correct copying notice to use GPL, per author's request. * src/paste.c: Likewise. (AUTHORS): Add djm. 1999-09-19 Bruno Haible * src/tr.c (PFI): Return `long', not `int'. (PFL): Rename from PFI now that it returns long. (squeeze_filter): Declare as `long': i, nr, out_len. 2000-01-17 Paul Eggert Tweak sort performance. * src/sort.c (hard_LC_CTYPE): Remove. (keylist): Renamed from keyhead. Now a pointer, not a mostly-unused struct. All uses changed. (findlines, keycompare, CMP_WITH_IGNORE, compare, checkfp, mergefps, sort): Tune and use a more consistent style for reallocation. (keycompare, main): Don't worry about LC_CTYPE; it's buggy with multibyte chars anyway. (compare): Invoke alloca (0) after each call to keycompare, not just the ones that return nonzero. This avoids a memory leak on architectures without builtin alloca that occurs sometimes when a file contains all duplicate lines. 2000-01-17 Paul Eggert * src/csplit.c (interrupt_handler, main): Don't use SA_INTERRUPT to decide whether to call sigaction, as POSIX.1 doesn't require SA_INTERRUPT and some systems (e.g. Solaris 7) don't define it. Use SA_NOCLDSTOP instead; it's been part of POSIX.1 since day 1 (in 1988). * src/sort.c (sighandler, main): Likewise. 2000-01-16 Jim Meyering * lib/Makefile.am: Merge with fileutils/lib/Makefile.am. * configure.in: Remove AM_FUNC_ERROR_AT_LINE, jm_FUNC_STRFTIME, AC_FUNC_VPRINTF, and AC_FUNC_ALLOCA. Now they're in m4/. 2000-01-15 Jim Meyering * configure.in (AC_REPLACE_FUNCS): Remove these: memcpy memset stpcpy strpbrk strtol strtoul. Now that's in m4/. unexpand could infloop * src/unexpand.c (TAB_STOP_SENTINEL): Define. (unexpand): Use it instead of INT_MAX. Declare column and pending to be `unsigned'. Increment pending and column counters only if column is smaller than TAB_STOP_SENTINEL. * tests/Makefile.am (SUBDIRS): Add unexpand. * tests/unexpand/basic-1: New tests for the above-fixed bug. * configure.in (AC_OUTPUT): Add tests/unexpand/Makefile. Reported by John Kodis. 2000-01-12 Paul Eggert * src/sort.c (fillbuf): Avoid quadratic behavior with long lines. Also, stop worrying about ancient memchr bug (misbehavior when size is zero), since other code doesn't worry either. 2000-01-11 Jim Meyering Sync lib/ directories of sh-utils, and textutils. * lib/Makefile.am (libfetish_a_SOURCES): Add getdate.y, posixtm.c, basename.c, canon-host.c, readutmp.c, stripslash.c, xgetcwd.c, xgethostname.c. (noinst_HEADERS): Add getdate.h, group-member.h, posixtm.h, pathmax.h, readutmp.h * lib/Makefile.am: s/tu/fetish/ * src/Makefile.am: s/libtu/libfetish/ 2000-01-10 Jim Meyering * Version 2.0c. * Makefile.maint (announcement): Include URLs for xdelta files. (alpha): Use scp, not ncftp. * lib/Makefile.am (noinst_HEADERS): Add nanosleep.h. 2000-01-08 Jim Meyering * Version 2.0b. More nits. * src/cut.c (OUTPUT_DELIMITER_OPTION): Define this and use it instead of `CHAR_MAX + n'. * src/pr.c (PAGES_OPTION, COLUMNS_OPTION): Likewise. 2000-01-07 Jim Meyering * tests/tsort/Makefile.am (TESTS_ENVIRONMENT): Add `pwd`/ prefix to exported PATH value (though not strictly necessary, here). * tests/md5sum/Makefile.am: Likewise. Nits. * lib/memcasecmp.c: Use `#if' instead of `#ifdef' for `HAVE_CONFIG_H'. Capitalize all macro parameters. (memcasecmp): Ansideclify. Don't cast away `const'ness of parameters. * lib/strpbrk.c (strpbrk): Ansideclify. Use `#if' instead of `#ifdef' for `HAVE_CONFIG_H'. Suggestions from François Pinard. 2000-01-06 Jim Meyering * tests/tail-2/assert: Tell the user just before sleeping for 7 seconds. * tests/tail-2/Makefile.am (TESTS): Remove fflush test. It didn't test the losing behavior, and left a stray tail process to boot. 2000-01-01 Jim Meyering * Version 2.0a. * lib/Makefile.am (lstat.c): Adapt rule to handle new parts of xstat.in. 1999-12-12 Jim Meyering Move 120+ lines of stat.h-related macros from system.h (not shared) to sys2.h, which is shared between fileutils, sh-utils, textutils. * src/system.h: Move them from here... * src/sys2.h: ... to here. 1999-12-09 Jim Meyering * configure.in (ALL_LINGUAS): Add Galician (gl). 1999-12-06 Jim Meyering * tests/tail-2/fflush: New test for latest change. * tests/tail-2/Makefile.am (TESTS): Add fflush. 1999-10-26 Marc Boucher * src/tail.c (main): Flush stdout before switching to unbuffered mode and calling tail_forever. Required only on Solaris2.7 -- on other systems, using setvbuf to switch to unbufferd mode does the flush. 1999-11-22 Jim Meyering * lib/Makefile.am (DISTCLEANFILES): Put $(BUILT_SOURCES) here rather than in $(MAINTAINERCLEANFILES). * src/tail.c (recheck): Handle a race condition (including reuse) that would lead to a failed assertion. Reported by Ken Pizzini. (tail_forever): Record errno before using it in call to `error' which might change it. (tail_file): Likewise. New test for the above. * tests/tail-2/assert: New file. * tests/tail-2/Makefile.am: New file. * tests/tail-2: New directory. * tests/Makefile.am (SUBDIRS): Add tail-2. * configure.in (AC_OUTPUT): Add tests/tail-2/Makefile. 1999-11-05 Jim Meyering * src/system.h: Use HAVE_STRUCT_STAT_ST_BLKSIZE not HAVE_ST_BLKSIZE. * configure.in: Move some type/header/member tests into m4/jm-macros.m4 (jm_CHECK_ALL_TYPES) so they are shared by all of fileutils, textutils, and sh-utils. 1999-11-04 Paul Eggert * src/sort.c (SORTALLOC): New macro. (sortalloc, mergealloc, LINEALLOC): Use it. (sortalloc, mergealloc, linelength): Now const. (sortalloc): Increase from 0.5 to 8 MB. (mergealloc): Increase from 16 to 256 kB. (LINEALLOC): Increase from 0.25 to 4 MB. 1999-11-03 Paul Eggert * NEWS, doc/textutils.texi: Do not consider newline to be part of a line when comparing lines in `sort' and `comm'. POSIX.2 requires that we consider newline, but this is a bug in the spec and the bug will likely be fixed. * src/comm.c (compare_files): Likewise. * src/sort.c (begfield, limfield, findlines, keycompare, compare): Likewise. * tests/sort/Test.pm: Update tests `use-nl' and `11d' to reflect this change. * lib/linebuffer.c (readline): Do not leave room for an extra byte after the newline; it's no longer needed. * src/sort.c (sortalloc, mergealloc, fillbuf, checkfp, mergefps): Likewise. * lib/memcoll.c (memcoll): The two arguments cannot be adjacent any more, so remove the alloca/copy workaround for that case. 1999-11-03 Jim Meyering Fix so that `tail -fn 2 file' works again. * src/tail.c (anonymous enum) [LONG_FOLLOW_OPTION]: Define. (long_options): Use LONG_FOLLOW_OPTION here, instead of 'f'. (main): Remove the `::' after the `f' in getopt_long string. Add `case LONG_FOLLOW_OPTION' after `case 'f':'. Based on a patch from Tim Waugh. * src/tail.c (anonymous enum) [RETRY_OPTION]: Define. [MAX_UNCHANGED_STATS_OPTION]: Likewise. [MAX_CONSECUTIVE_SIZE_CHANGES_OPTION]: Likewise. [PID_OPTION]: Likewise. (long_options): Use *_OPTION instead of CHAR_MAX + N. (main): Likewise. 1999-11-02 Jim Meyering * man/help2man: Import version 1.018. * man/Makefile.maint ($(man_MANS)): Remove use of --name=... option. * man/*.x: Include one-line summary in [NAME] section. * man/Makefile.summ (cat-summary): Remove now-unused file. * man/Makefile.am (EXTRA_DIST): Remove Makefile.summ. * man/Makefile.maint: Include Makefile.summ with leading `-'. This file is shared by fileutils and sh-utils, both of which still have the file (albeit nearly empty now). Suggestion for clean-up from Akim Demaille. 1999-11-01 Jim Meyering * lib/Makefile.am (MAINTAINERCLEANFILES): Set to $(BUILT_SOURCES). 1999-10-20 Jim Meyering * src/tail.c (recheck): Don't refuse to tail a non-regular, non-pipe. (tail_file): Likewise. 1999-10-07 Jim Meyering * configure.in (ALL_LINGUAS): Add Japanese (ja). 1999-10-04 Jim Meyering * depcomp: New file, for automake's new dependency support. * missing: New version, from automake's user-dep-gen-branch. 1999-09-28 Jim Meyering * src/system.h: Update from the system.h from fileutils. * lib/error.c (error_at_line): Use strerror portably here, too, just as was done in error. 1999-09-02 Jim Meyering * src/cut.c: Remove xstrdup declaration. * src/sort.c: Likewise. * src/tsort.c: Likewise. 1999-08-22 Jim Meyering * src/tsort.c (detect_loop): There's no loop if k->top is NULL. Niklas Edmundsson showed how to make tsort segfault. * tests/tsort/Makefile.am: Rewrite to use Fetish.pm-based framework. * tests/tsort/basic-1: New file. Add a few more tests. * tests/tsort/Test.pm: Remove. * man/help2man: Escape backslashes so tr's manual comes out right. Reported by Andreas Schwab. * src/tsort.c (usage): Correct address for bug reports. (tsort): Remove trailing `\n' in error format. * src/md5sum.c (md5_check): Remove trailing `\n' in error format. * src/sys2.h (IF_LINT): Define new macro. * src/od.c (dump): Use IF_LINT macro instead of #ifdef lint... (main): Likewise. * src/paste.c (paste_parallel): Likewise. * src/pr.c (read_line): Likewise. * src/sort.c (checkfp): Likewise. (mergefps): Likewise. * src/cksum.c (cksum): Constify a char*. * src/comm.c (writeline): Likewise. * src/uniq.c (check_file): Move declarations of local variables into the scopes where they're used. (min): Remove macro definition. (different): Use MIN, not min. (SWAP_LINES): New macro. (check_file): Use it here. * src/uniq.c (check_file): Generate each line of output earlier, when possible. It is possible when using none of these options: --count, -repeated, --all-repeated, --unique. Based on a patch from Ian Turner. 1999-08-15 Jim Meyering * src/tail.c (ENOSYS): Define to a bogus value if not already defined. 1999-08-13 Eli Zaretskii * src/tail.c (parse_options): Warn if --pid=PID is used and kill() returns ENOSYS (e.g. when compiled with DJGPP). 1999-08-13 Jim Meyering * src/tsort.c (usage): Use consistent bug-reporting address. * tests/tsort/Test.pm: New file. * tests/tsort/Makefile.am: New file. * tests/tsort: New directory. * tests/Makefile.am (SUBDIRS): Add tsort. * configure.in (AC_OUTPUT): Add tests/tsort/Makefile. 1999-08-11 Paul Eggert * doc/textutils.texi: Document how to ignore newline during sort. 1999-08-07 Jim Meyering * po/POTFILES.in: Add lots of lib/*.c files. Remove src/system.h. 1999-08-06 Jim Meyering * Version 2.0. * src/sort.c: Include file name in `write error' diagnostics. (write_bytes): Add output_file parameter and use it. Update callers. (mergefps): Likewise. (merge): Likewise. (sort): Likewise. Reported by John Summerfield. 1999-08-05 Jim Meyering * src/tail.c (Follow_mode): Remove comma at end of enumerator list. Reported by Kaveh Ghazi. * config.sub: Update from autoconf. * config.guess: Update from autoconf. 1999-08-04 Jim Meyering * Version 1.22q. * configure.in: Remove inadequate getline-testing code. md5sum would segfault on HPUX because of the getline function in their C library. Use the AM_FUNC_GETLINE test in m4/ instead. 1999-08-03 Jim Meyering * src/tail.c (tail_forever): When following by name and calling recheck because of exceeding max_n_consecutive_size_changes_between_opens, `continue' so we don't fall through and (assuming the file finally grew) get the erroneous `file truncated' message. This was hard to reproduce. 1999-08-01 Jim Meyering * Version 1.22p. * configure.in (AM_C_PROTOTYPES): Move test to precede a few header tests -- merely on principle. 1999-08-01 Paul Eggert * configure.in (AC_SYS_LARGEFILE): Renamed from AC_LFS. 1999-07-16 Paul Eggert * lib/quotearg.c (quotearg_buffer): Don't quote spaces if C quoting style. 1999-07-31 Jim Meyering * src/tail.c (tail_forever): Don't call kill if pid is 0. Detect when `writer_is_dead' also when the writer is some other user. From Karl Heuer. * src/tail.c (parse_options): Warn if --pid=PID is used without -f. 1999-07-30 Jim Meyering * Version 1.22o. * src/tail.c: New option: --pid=PID. Include signal.h for kill prototype. (pid): New global. (long_options): Add `pid'. (usage): Describe it. (tail_forever): Implement it. (parse_options): Handle the new option and required arg. Suggestion and pseudo-code from Karl Heuer. 1999-07-27 Jim Meyering * lib/linebuffer.c: Include now that linebuffer.h uses size_t. This is required on at least SunOS4. From Kaveh Ghazi. 1999-07-26 Jim Meyering * src/sys2.h (PID_T_MAX): Define. * src/tail.c (struct File_spec) [n_stat_calls]: Remove unused member. 1999-07-25 Jim Meyering * src/pr.c (usage): Remove `NEWS'-style paragraph (sorry to have let that in, translators). * Version 1.22n. 1999-06-01 Volker Borchert * tests/Makefile.am: Make envvar-check depend on check-recursive rather than on `check' so that its tests are performed before any real tests. 1999-07-20 Jim Meyering * configure.in (AC_REPLACE_FUNCS) Remove memmove, now that it's AC_REPLACE...'d in m4/jm-macros.m4. * src/wc.c (posixly_correct): Declare global. (write_counts): Use it to select printf formats. (main): Set posixly_correct from the POSIXLY_CORRECT envvar. From Peter Moulder. 1999-07-15 Jim Meyering * tests/md5sum/Makefile.am: Revert the 1999-02-15 change. * tests/md5sum/basic-1, newine-1: Add --text for each individual test. Reported by Eli Zaretskii. 1999-07-12 Jim Meyering * configure.in (ALL_LINGUAS): Add Slovak (sk). 1999-07-11 Jim Meyering * src/tail.c (recheck): Use assert instead of unnecessary close_fd. Remove a couple of unnecessary FIXME comments. 1999-07-10 Jim Meyering * doc/textutils.texi: Document new tail options. * src/tail.c (struct File_spec) [tailable]: Rename from `missing' and document. Change all uses and locals like was_missing to was_tailable. Invert expressions as appropriate. (reopen_inaccessible_files): Rename from allow_missing. (sleep_interval): Describe. (--allow-missing): Deprecate. (--retry): New option, equivalent to --allow-missing. (usage): Document name vs. descriptor differences. Refer to manual for descriptions of --max-unchanged-stats=N and --max-consecutive-size-changes=N. (valid_file_spec): New function. (recheck): Assert valid_file_spec. Remove dead else-if block (suggestion from Eli Zaretskii). Adjust stmts that set f->tailable -- unlike for `missing', tailable doesn't depend on errno == ENOENT. (parse_options): Give a warning if --retry is used when not following by name. * tests/join/Test.pm: New test case (but commented out) for Paul's 1999-06-01 fix. 1999-07-09 Jim Meyering * configure.in (ALL_LINGUAS): Add Greek (el). 1999-07-04 Jim Meyering * tests/join/Test.pm: New test case for Paul's 1999-06-03 fix. 1999-07-03 Eli Zaretskii * src/tail.c (struct File_spec): New member, errnum. (recheck): Record the new value of errno in f->errnum. Don't output an error message unless the new value of errno differs from the old one. Output a message if previously-inaccessible file becomes accessible. (tail_forever): Always recheck files whose fd is negative. If the file cannot be fstat'ed, record the errno value in f[i].errnum. (tail_file): If the file cannot be open, record the errno value in f->errnum. If it can be opened, initialize f->errnum to zero. If it's a non-regular non-fifo file, initialize f->errnum to -1. 1999-06-21 Jim Meyering * doc/textutils.texi: Use lower case characters in sc{} context. Reported by Eli Zaretskii. 1999-06-03 Paul Eggert * src/join.c (xfields): Only s separate fields, not s. 1999-06-01 Paul Eggert * lib/linebuffer.c (readline): Leave room for an extra byte after the newline; comm needs this for memcoll. 1999-06-01 Paul Eggert Add LC_COLLATE support to `join'. * doc/textutils.texi: Describe join and LC_COLLATE. * src/join.c: Add support for LC_COLLATE locale. Include hard-locale.h, linebuffer.h, memcoll.h. (struct line): New member `buf', replacing `beg' and `lim'. All uses changed. (hard_LC_COLLATE): New var. (main): Initialize it. (get_line): Use readline to read the line, instead of doing it by hand. That way, we get a buffer that we can pass to memcoll. (keycmp): Use memcoll to compare lines if hard_LC_COLLATE is nonzero. 1999-05-27 Volker Borchert * tests/Makefile.am: Qualify .env-warn with $(srcdir)/ prefix. 1999-05-25 Paul Eggert Add LC_COLLATE support to `comm', so that `comm' is compatible with `sort' in nontrivial locales. * doc/textutils.texi: Document locale-specific mode for comm. * lib/Makefile.am (libtu_a_SOURCES): Add hard-locale.c, memcoll.c. (noinst_HEADERS): Add hard-locale.h, memcoll.h. * src/comm.c: Include hard-locale.h, memcoll.h. (hard_LC_COLLATE): New variable. (compare_files): Use memcoll to compare if hard_LC_COLLATE. (main): Initialize hard_LC_COLLATE from locale. * src/sort.c: Include hard-locale.h, memcoll.h. (hard_LC_COLLATE, hard_LC_CTYPE, hard_LC_TIME): New variables, replacing `need_locale'. (memcoll): Move to lib/memcoll.c. (keycompare): No need to alloc (0), since our caller now does it. (compare): alloca (0) before returning. (my_setlocale): Remove; hard_locale now dows this. (main): Invoke setlocale, bindtextdomain, and textdomain before invoking anything that might print an error. Use hard_locale to determine which locales are hard. * lib/hard-locale.c, lib/hard-locale.h, lib/memcoll.c, lib/memcoll.h: New files. 1999-05-25 Paul Eggert * lib/linebuffer.c (readline): Append trailing newline to line. * lib/linebuffer.h [struct linebuffer] (size): Declare to be of type size_t, not long. [struct linebuffer] (length): Likewise. * src/comm.c, (writeline): Lines now contain trailing newline. * src/uniq.c (find_field, different): Use size_t, not int, for lengths. (writeline): Lines now contain trailing newline. (check_file): Use size_t, not int, for lengths. * src/nl.c (proc_text, check_section, main): More of the same. * lib/linebuffer.h (struct linebuffer): Use size_t for sizes. src/nl.c (header_del_len, body_del_len, footer_del_len, main): Likewise. src/uniq.c (find_field, different, check_file): Likewise. * lib/linebuffer.c (readline): Silently append trailing newline if needed. Do not bother setting buffer length to 0 at EOF, since it's not part of the spec and nobody relies on it. Do not compute the difference between unrelated pointers. 1999-05-25 Paul Eggert * src/tac.c (memrchr): Ifdef out this unused function. 1999-05-25 Jim Meyering * doc/textutils.texi (Squeezing): Remove misleading square brackets from SET1 in the one-word-per-line example. 1999-05-22 Jim Meyering * lib/Makefile.am (libtu_a_SOURCES): Remove memchr.c. From Ulrich Drepper. * Version 1.22m. 1999-05-22 Paul Eggert * doc/textutils.texi: Document locale-specific changes to `sort', as well as the new, POSIX-compliant definition of line comparison, and -g's more careful treatment of NaNs, infinities and zeros. * src/sort.c (general_numcompare): Put exceptional cases first, not last, to be consistent with -M. 1999-05-21 Paul Eggert * src/sort.c (strtod): Declare if STDC_HEADERS is not defined. (general_numcompare): Use strtod, not xstrtod. Do not consider partial conversions to be errors. Put -infinity at the start, and +infinity at the end; follow +infinity with NaNs (sorted by bit pattern), and finally by conversion errors. 1999-05-21 Jim Meyering * tests/sort/Test.pm (11d): Reverse lines in expected output to reflect latest change. (use-nl): New test from Paul Eggert. 1999-05-20 Paul Eggert * src/sort.c: Treat the trailing newline as part of the line, as required by POSIX.2. (struct line, findlines, compare, checkfp, mergefps, sort): A line now includes its trailing newline. (findlines): Do not replace newline with NUL. (memcoll, keycompare): Work even if the data to be compared are adjacent strings; this is possible now that lines contain the trailing newline. (fillbuf): Always have an unused byte at the end of the buffer, since memcoll and keycompare want to modify a byte after the last line. (sortalloc, mergealloc): Increase by 1, for trailing byte. 1999-05-20 Jim Meyering * tests/sort/Test.pm: Add test case from Paul Eggert. 1999-05-20 Paul Eggert * src/sort.c (keycompare): Ignore any length difference if the localized comparison says the strings are equal. * src/sort.c (memcoll, keycompare, compare): Handle NUL characters properly when comparing with LC_COLLATE semantics. (NLS_MEMCMP): Remove. (memcoll): Renamed from strncoll. Take separate lengths for each string. This function is now invoked only when need_locale. (keycompare): Don't copy strings when ignore and translate are both NULL. 1999-05-18 Paul Eggert * src/sort.c (MONTHTAB_CONST): Renamed from NLS_CONST; the use is also changed. Define to const also if !HAVE_NL_LANGINFO. (usage): `,' -> `;' (English typo). 1999-05-17 Eli Zaretskii * src/cat.c (main): When stdout is in binary mode, make sure all input files are also read in binary mode. 1999-05-16 Jim Meyering * Version 1.22l. 1999-05-11 Paul Eggert * src/sort.c: Don't autodetect the locale of numbers and months, as this conflicts with POSIX.2 and is tricky to boot. (FLOATING_COMMA, NLS_STRNCMP, NLS_MAX_GROUPS, NLS_ONE_CHARACTER_STRING): Remove macros no longer used. (nls_grouping, nls_fraction_found, nls_month_found, nos_monthtab, nls_months_collide, nls_keyhead, us_monthtab): Remove variables no longer used. (struct nls_keyfield): Remove types no longer used. (strncoll_s2_readonly, nls_set_fraction, look_for_fraction, nls_month_is_either_locale, nls_numeric_format): Remove functions no longer used. (monthtab): Now has the role that us_monthtab had, but it's const only if ENABLE_NLS is not defined. (C_DECIMAL_POINT): Renamed from FLOATING_POINT. All uses changed. (MONTHS_PER_YEAR): Renamed from NLS_NUM_MONTHS. All uses changed. (struct_month_cmp): Renamed from nls_sort_month_comp. All uses changed. Use strcmp, not strcoll, since the user doesn't care about collating here. (inittables): Read locale data into monthtab, rather than modifying a separate month table and futzing with indirection. Do not worry about colliding months, since we no longer autodetect month locale. (fraccompare): Don't set no-longer-used variable nls_fraction_found. (getmonth): Use strncmp to compare months, since user doesn't care about collating here. Fix bug where code incorrectly assumed that strlen (monthtab[lo].name) == strlen (monthtab[ix].name). (keycompare, main): Don't autodetect month locale. (compare): Don't use NLS_MEMCP in code that can't be executed if need_locale is false, as NLS_MEMCP is equivalent to memcmp in that case. (sort, insertkey, main): Don't autodetect numeric locale. 1999-05-15 Jim Meyering * tests/join/Test.pm (trailing-sp): New test for this fix. * src/join.c (xfields): Don't interpret a trailing blank as a delimiter when e.g. -t: was specified. From Tim Smithers. 1999-05-12 Jim Meyering * tests/Makefile.am (envvar-check): Renamed from check-local. (check): Depend on envvar-check so the envvar check is performed before all other tests. Reported by Volker Borchert. * tests/.env-warn: Use `%%' place-holder that Makefile.am rule expects, so CDPATH is mentioned in the message. Reported by Volker Borchert. 1999-05-11 Jim Meyering * src/sort.c (usage): Split the --help message into two pieces so that neither is longer than 2048. For Irix4's cc. Reported by Kaveh Ghazi. 1999-05-09 Jim Meyering * lib/regex.c: Update from libc. * Version 1.22k. * Makefile.maint (alpha): Put the announcement in /tmp/announce-$(distdir) * tests/sort/Test.pm (neg-nls): New test. 1999-05-08 Jim Meyering * src/system.h (CHAR_BIT, TYPE_SIGNED, TYPE_MINIMUM, TYPE_MAXIMUM, and all the *_MIN and *_MAX symbols): Remove definitions. * src/sys2.h: Put the definitions here instead (this file is shared between all three *utils packages, while system.h is not). 1999-05-06 Paul Eggert * src/sort.c (fraccompare, numcompare): Merge the NLS and non-NLS versions into a single function. (decimal_point): Now char, since we no longer convert to unsigned char. (th_sep): Now int, since we use a value out of char range to denote the absence of a thousands separator. (IS_THOUSANDS_SEP): New macro. (USE_NEW_FRAC_COMPARE): Remove. (nls_set_fraction): Arg is now char, not unsigned char. Set th_sep to CHAR_MAX + 1 if there is no thousands separator. (numcompare): Don't convert to unsigned char unless necessary. (main): Turn off decimal points and thousand separators if they are multibyte characters, as we don't support that yet. 1999-05-06 Paul Eggert * src/system.h (CHAR_MIN, CHAR_MAX): New macros. (SCHAR_MIN, SCHAR_MAX): Don't assume that char is signed. 1999-05-06 Paul Eggert * src/sort.c (numcompare): Handle comparison of two negative numbers correctly in the ENABLE_NLS case. 1999-05-04 Jim Meyering * src/pr.c (usage): Break the usage message into 3 pieces instead of only 2. The strings had grown to be longer than 2048, which evokes errors when compiling with Irix4's cc. Reported by Kaveh Ghazi. * src/tsort.c (search_item): Use `1' instead of `+1'. The latter elicits a syntax error from SunOS4's cc. From Kaveh Ghazi. 1999-05-03 Jim Meyering * src/ptx.c : Don't include. [!STDC_HEADERS]: Remove definitions of ctype macros. Convert e.g., isspace to ISSPACE to use definitions from sys2.h. Reported by Kaveh Ghazi. * src/sys2.h (TOLOWER): Define. (TOUPPER): Define. * src/join.c (TOLOWER): Remove definition. * src/md5sum.c (TOLOWER): Remove definition. 1999-04-30 Jim Meyering * src/sort.c (usage): Document the differences between the obsolescent, +POS1[-POS2] form, and the POSIX -k option. 1999-04-24 Jim Meyering * configure.in: Use AC_CANONICAL_HOST. * lib/Makefile.am (libtu_a_SOURCES): Add xstrtoumax.c (noinst_HEADERS): Remove xstrtoul.h. * src/csplit.c: Include new "xstrtol.h", not "xstrtoul.h". (struct control) [repeat]: Declare as uintmax_t, not int. (struct control) [lines_required]: Likewise. (handle_line_error): Use human_readable to print lines_required. (parse_repeat_count): Parse a uintmax_t. (parse_patterns): Parse a uintmax_t. * src/tail.c: Include new "xstrtol.h", not "xstrtoul.h". * src/od.c: Likewise. * src/head.c: Include new "xstrtol.h", not "xstrtoul.h". Change all U_LONG_LONG to uintmax_t. (head_lines): Move a couple dcls into an inner scope. (string_to_integer): Rename from string_to_ull. 1999-04-19 Jim Meyering * Makefile.maint (b_host): Remove /pub suffix. 1999-04-18 Jim Meyering * Version 1.22j. * Makefile.maint (my-distcheck): Use AMTAR, not TAR. * src/sort.c (usage): s/DIRECT/DIRECTORY/g Rename global: s/temp_file_prefix/temp_dir/. (NAME_MAX_IN_DIR): Rename from PATH_MAX_IN_DIR. Use _POSIX_NAME_MAX, not _POSIX_PATH_MAX. Guard with #if HAVE_PATHCONF rather than #if HAVE_UNISTD_H. (tempname): Wrap after 99999 only for length-impaired file systems. 1999-04-17 Jim Meyering * src/tail.c (file_lines): Fix serious bug introduced with last changes. From Andreas Schwab. 1999-04-15 Jim Meyering * tests/Makefile.am (EXTRA_DIST): Rename .posix-warn to .env-warn. (check-local): Warn about CDPATH in the same way we warn about POSIXLY_CORRECT, since the cp/same-file test fails at least when using bash with CDPATH set. Reported by Mark Hewitt. * src/pr.c Add comments. (init_header): Tweak white space in Date/Time header. * tests/pr/Test.pm: Updated all tests to reflect the big 1999-02-13 change. From Roland Huebner. 1999-04-12 Jim Meyering * src/cat.c (main): Declare out_dev to be of type dev_t, not `int'. Declare out_ino to be of type ino_t, not `int'. From John Bley. * src/od.c (MIN, MAX): Remove definitions. * src/sys2.h (MIN, MAX): Define here instead. Reported by John Bley. 1999-04-11 Jim Meyering * Version 1.22i. * tests/pr/Test.pm (test_vector): Disable all tests while I wait for an updated suite. Fix the problem whereby `yes > k & sleep 1; tail -2c k' would infloop. * src/tail.c (COPY_TO_EOF): Define. (dump_remainder): Add parameter, n_bytes, and rewrite to use it. Update callers. (file_lines): Rename parameter. (tail_bytes): Remove obsolete comment. Fix the problem whereby `yes > k & sleep 1; tail -1 k' would infloop. * src/tail.c (dump_remainder): Move this function to precede the new use in file_lines. (tail_lines): Don't call dump_remainder here. (file_lines): Call dump_remainder here instead. Reported by Lehti Rami. * lib/readtokens.c (readtoken, readtokens): Protoize. 1999-04-10 Jim Meyering * src/tail.c (xwrite): Use STDOUT_FILENO instead of literal `1'. 1999-04-04 Jim Meyering * src/cat.c: Standardize --help and --version processing. * src/comm.c: Likewise. * src/csplit.c: Likewise. * src/cut.c: Likewise. * src/expand.c: Likewise. * src/fmt.c: Likewise. * src/fold.c: Likewise. * src/head.c: Likewise. * src/join.c: Likewise. * src/md5sum.c: Likewise. * src/nl.c: Likewise. * src/od.c: Likewise. * src/paste.c: Likewise. * src/pr.c: Likewise. * src/ptx.c: Likewise. * src/split.c: Likewise. * src/sum.c: Likewise. * src/tac.c: Likewise. * src/tail.c: Likewise. * src/tr.c: Likewise. * src/unexpand.c: Likewise. * src/uniq.c: Likewise. * src/wc.c: Likewise. 1999-03-29 Jim Meyering * configure.in (GNU_PACKAGE): Remove related code -- now it's in the catch-all for shared autoconf code, m4/jm-macros.m4. (jm_CHECK_ALL_TYPES): Remove explicit AC_TYPE_* macros and use this instead. 1999-03-26 Jim Meyering * lib/Makefile.am (libfu_a_SOURCES): Add version-etc.c. (noinst_HEADERS): Add version-etc.h. 1999-03-25 Jim Meyering * src/tail.c (recheck): Factor out a block of duplicated code. Set f->size to 0 upon encountering a new file so we read it from the beginning rather than from the end of the first line or block. Otherwise, after a log rotation, tail would omit the first line or block of the new file. Reported by Ed Avis. 1999-03-20 Jim Meyering * tests/Makefile.am (SUBDIRS): Temporarily remove pr. 1999-03-13 Jim Meyering * src/tac.c (tac_mem): `#if-0'-out this unused function. (tac_stdin_to_mem): Likewise. * doc/textutils.texi (cut invocation): Describe --output-delimiter. 1999-03-12 Jim Meyering * src/sys2.h (SETVBUF): Define new macro. * src/tail.c (dump_remainder): Don't fflush stdout here. (main) [if forever]: Make stdout unbuffered, instead. Akim Demaille pointed out that when running `echo x>a; tail -f a>>a' , the file `a' didn't grow longer than two lines. Now it grows without bound. 1999-03-07 Jim Meyering * tests/md5sum/newline-1: Test for the actual feature. With help from Eli Zaretskii. * src/pr.c (usage): Add missing \n\. The newline test would always fail on MSDOS/Windows systems -- so move it to a separate file where we can test for that. * tests/md5sum/basic-1: Remove newline test. * tests/md5sum/newline-1: New file. * tests/md5sum/Makefile.am (TESTS): Add newline-1. Pointed out by Eli Zaretskii. 1999-02-13 Roland Huebner * src/pr.c (main): Redefine options -s, -w to be POSIX compliant; introduce new options -J, -S, -W to disentangle -s and -w when used together with the three column options; (add_line_number): Make POSIX compliant; use default number separator TAB with single column output. (add_line_number): Make POSIX compliant; with multicolumn output now prefer `text columns of equal width' rather than a consequent use of `default n-separator TAB'. (add_line_number): Change line number cut-off from lower-oder to higher-oder digit to avoid loss of information; no consequent handling exists in different utilities and other UNIXes. (char_to_clump): Expand input text tabs to 8 spaces, if input_tab_char doesn't equal TAB (adapted to other UNIXes). (usage): Update POSIX compliant options -s, -w; add new options -J, -S, -W. (main): Update the source internal documentation. Some smaller BUGFIXES (print_sep_string, init_header, skip_to_page, reset_status, print_header). * tests/md5sum/basic-1 (backslash): Use .\foo instead of \.foo so we don't tramp on root directory in MSDOS/Windows. Suggestion from Eli Zaretskii. 1999-02-15 Eli Zaretskii * tests/md5sum/Makefile.am (TESTS_ENVIRONMENT): Run md5sum with the --text option (for MSDOS). 1999-03-06 Jim Meyering * src/cut.c (getstr): Change type of `delim' parameter from char to int. (cut_fields): Cast to `unsigned char' before comparing. (main): Cast to `unsigned char' before assigning. From Arne Juul. * tests/cut/Test.pm: Add a test to exercise the bug. * src/ptx.c (swallow_file_in_memory): Use a `%s' format in error call, in case the argument string contains a `%'. (main): Likewise. * src/fmt.c (main): Likewise. * src/sort.c (main): Likewise. 1999-02-13 Eli Zaretskii * src/sys2.h [__DJGPP__]: Include and . * src/ptx.c (swallow_file_in_memory): Slurp up the whole file at once on MSDOS as well, but we have to relax the test for whether reading it succeeded. 1999-03-03 Jim Meyering * src/cat.c: Include long-options.h [long_options]: Remove the "help" and "version" entries. Remove declarations of show_help and show_version. (main): Use parse_long_options, including author name(s). Remove the show_version and show_help blocks. * src/cksum.c: Likewise. * src/comm.c: Likewise. * src/csplit.c: Likewise. * src/cut.c: Likewise. * src/expand.c: Likewise. * src/fmt.c: Likewise. * src/fold.c: Likewise. * src/head.c: Likewise. * src/nl.c: Likewise. * src/od.c: Likewise. * src/paste.c: Likewise. * src/pr.c: Likewise. * src/split.c: Likewise. * src/sum.c: Likewise. * src/tac.c: Likewise. * src/tail.c: Likewise. * src/tr.c: Likewise. * src/unexpand.c: Likewise. * src/uniq.c: Likewise. * src/wc.c: Likewise. * src/ptx.c: Include long-options.h [long_options]: Remove the "help" and "version" entries. Remove declarations of show_help and show_version. (main): Remove `const' attribute from dcl of argv parameter. Call bindtextdomain and textdomain. Use parse_long_options, including author name(s). Remove the show_version and show_help blocks. * src/join.c (main): Include author name argument in call to parse_long_options. * src/md5sum.c (main): Likewise. * src/sort.c (main): Likewise. * src/tsort.c (main): Likewise. 1999-02-07 Jim Meyering * Version 1.22h. * Makefile.maint (my-distcheck): Don't depend on dist, now that this is hooked up to the distcheck rule. * Makefile.am (distcheck-hook): New target and rule -- link to shared rule, my-distcheck, in Makefile.maint. 1999-01-31 Jim Meyering * doc/textutils.texi: Wrap the @top node in @ifnottex instead of @ifinfo so `makeinfo --html ...' works. From Karl Berry. 1999-01-30 Jim Meyering * acconfig.h: Remove lots of `#undef's, now that we use the 3-argument forms of AC_DEFINE* macros. * configure.in: Require autoconf 2.13. Use 3-argument form of AC_DEFINE*. 1999-01-24 Jim Meyering * src/tac.c (DONT_UNLINK_WHILE_OPEN) [__MSDOS__ || _WIN32]: Define. * lib/quotearg.c (quotearg_n_options): Revert type of parameter `n' (and hence that of the local `n1', too) to `int' at Paul's request. 1999-01-18 Akim Demaille * doc/textutils.texi: Harmonization of @samp use for options. 1999-01-17 Jim Meyering * Makefile.am (SUBDIRS): Add djgpp. * configure.in (AC_OUTPUT): Add djgpp/Makefile. * djgpp/: New directory. * man/Makefile.maint ($(man_MANS)): Don't remove the target (the man page) until after we've created its replacement. 1999-01-16 Eli Zaretskii * GNUmakefile: Add SHELL = /bin/sh. * man/GNUmakefile: Likewise. 1999-01-16 Jim Meyering * acconfig.h: Remove @BOTTOM@ section. Instead, add the define and decl via m4/jm-macros.m4. * lib/argmatch.h (XARGMATCH): Define to return a value once again. (XARGCASEMATCH): Likewise. * lib/argmatch.c (EXIT_FAILURE): Define. (ARGMATCH_DIE): Provide default. (__xargmatch_internal): New function. s/rogram_name/program_name. From Akim Demaille. 1999-01-14 Jim Meyering * tests/md5sum/basic-1: Use `f', not `x' as temp. file name to avoid warnings from perl5.004. Reported by Volker Borchert. * lib/long-options.c (parse_long_options): Rename `usage' parameter to avoid shadowing globally scoped function. 1999-01-14 Akim Demaille * acconfig.h: Add a @BOTTOM@ section. (ARGMATCH_DIE) [@BOTTOM@]: Define to usage(1). Declare usage. * src/*.c: Don't prototype usage as static. 1999-01-10 Jim Meyering * Version 1.22g. All of the following new code is protected by `#if DONT_UNLINK_WHILE_OPEN' * src/tac.c (file_to_remove): New global. (fp_to_close): New global. (unlink_tempfile): New function. (record_tempfile): New function. (save_stdin): Call record_tempfile. Use SET_BINARY. (main): Use SET_BINARY and SET_BINARY2. From Eli Zaretskii. 1999-01-09 Jim Meyering * tests/uniq/Test.pm: Add tests from Jochen Hein. * src/uniq.c: New option: --all-repeated (-D). (output_all_repeated) [output_mode]: New enum value. (usage): Describe it. (writeline): Test for new mode. (check_file): Likewise. Based on patches from Jochen Hein and Florin Iucha. (main): Diagnose `too many arguments'. * tests/head/Test.pm (null-1): Add test from Jochen Hein. * src/tail.c (parse_obsolescent_option): Interpret `number' as decimal. (parse_options): Likewise. Reported by Kamal Paul Nigam. * src/tail.c: New option: --max-unchanged-stats=N. New option: --max-n-consecutive-size-changes=N. 1999-01-03 Jim Meyering * src/md5sum.c (usage): Remove third program_name argument -- there were only two `%s' in the format string. 1999-01-02 Jim Meyering * src/tsort.c (tsort): Use a single call to error instead of two to fprintf. (main): Remove `%s: ' prefix on format string. * src/tail.c (parse_options): Use XARGMATCH in place of argmatch. * src/ptx.c (format_vals): New array. (main): Use XARGMATCH in place of argmatch. * lib/argmatch.h (XARGMATCH): Don't return a value; instead, modify a parameter. Add a `Die_stmt' parameter. Add parentheses. * lib/argmatch.c (__xargmatch_internal): Remove now-unused function. (argmatch_to_argument): Add `const' attribute to first parameter. 1999-01-01 Jim Meyering * src/tsort.c: Move inclusion of assert.h to follow that of stdio.h. Some losing systems require this. Use STREQ macro instead of strcmp in a few places. * po/POTFILES.in: Add tsort.c. * man/Makefile.maint ($(man_MANS)): `exit 1' if any of help2man, chmod, or mv fails. Otherwise, the failure could go unnoticed. * man/Makefile.summ: Add entries for ptx and tsort. * src/tsort.c: Include readtokens.h. (zeros): Rename global from `rr'. (getstr): Remove function. (tsort) Use readtoken instead of getstr. * lib/readtokens.c: New file. * lib/readtokens.h: New file. * lib/Makefile.am (libtu_a_SOURCES): Add readtokens.c. (noinst_HEADERS): Add readtokens.h. * man/Makefile.am (man_MANS): Add ptx.1. * man/ptx.x: New file. * src/tail.c (parse_options): Use XARGMATCH in place of argmatch. * man/Makefile.am (man_MANS): Add tsort.1. * man/tsort.x: New (essentially empty) file. * src/tsort.c: Rename globals N and R so they don't shadow locals. (tsort): Rename from `sort'. 1998-11-07 Mark Kettenis * src/Makefile.am (bin_PROGRAMS): Add tsort. * src/tsort.c: New program. * lib/Makefile.am (libtu_a_SOURCES): Add quotearg.c. (noinst_HEADERS): Add quotearg.h. ============================ All of the following are from: 1998-04-17 Eli Zaretskii * src/system.h (SET_BINARY, SET_BINARY2, fileno, setmode): New macros. * src/cat.c (usage) [O_BINARY]: Describe -B,--binary option. (main) [O_BINARY]: binary_files, binary_output, file_open_mode: new variables. Add --binary to long_options[]. Switch stdin and stdout to binary mode unless file contents are not important anyway. Open files in binary mode when required. * src/cksum.c (cksum) [O_BINARY]: Read redirected stdin in binary mode. * src/expand.c (expand): Use binary I/O where appropriate. * src/head.c (head_bytes, head_lines) [O_BINARY]: Use binary I/O. * src/md5sum.c (OPENOPTS) [O_BINARY]: Use binary I/O when non-zero argument. (md5_file) [O_BINARY]: Switch redirected stdin to binary mode. (main) [O_BINARY]: Use binary reads by default on those systems which care about the difference. * src/od.c (skip, read_char, read_block) [O_BINARY]: Switch input stream to binary mode. * src/sort.c (PATH_MAX_IN_DIR) [HAVE_UNISTD_H]: New macro, for max file name characters in a given directory. (tempname): Make sure the temp file name is unique even if long file names aren't supported. * src/split.c (cwrite) [O_BINARY]: Write output in binary mode. (main) [O_BINARY]: Read input in binary mode. * src/sum.c (bsd_sum_file, sysv_sum_file) [O_BINARY]: Read input in binary mode. * src/tac.c (record_tempfile, unlink_tempfile) [DONT_UNLINK_WHILE_OPEN]: New functions, for systems where a file cannot be removed before it is closed. (save_stdin) [DONT_UNLINK_WHILE_OPEN]: Record the temporary file, to be removed before exit. (tac_file, save_stdin, main): Use binary I/O when appropriate. * src/tail.c (tail_lines, tail_bytes) [O_BINARY]: Use binary I/O when appropriate. * src/tr.c (main) [O_BINARY]: Use binary I/O when appropriate. * src/unexpand.c (unexpand): Use binary I/O where appropriate. * src/wc.c (wc): Use binary mode for input. * doc/textutil.texi: Add comments about peculiarities of Textutils operation on MS-DOS/MS-Windows. 1998-12-22 Jim Meyering * configure.in (ALL_LINGUAS): Add chinese (zh). 1998-12-17 Jim Meyering New options for tail: --follow=name, --follow=descriptor, --allow-missing * src/tail.c (Follow_mode): New enum. (n_live_files): New function. (tail_forever): Avoid starvation with --follow=name and a continually-growing unlinked or renamed file. 1998-12-13 Jim Meyering * tests/uniq/Test.pm: New file. * tests/uniq/Makefile.am: New file. * tests/uniq: New directory. * tests/Makefile.am (SUBDIRS): Add uniq. * configure.in (AC_OUTPUT): Add tests/uniq/Makefile. From Jochen Hein. * lib/Makefile.am (lstat.c): Add rule to generate this from xstat.in. (stat.c): Likewise. (EXTRA_DIST): Add xstat.in. * lib/stat.c: Remove file. * lib/lstat.c: Remove file. * lib/xstat.in (xstat@): New file. * configure.in (ALL_LINGUAS): Add Russian (ru). 1998-10-31 Jim Meyering * acconfig.h (stat): New #undef. 1998-10-22 Jim Meyering * src/fold.c (usage): Add mention of --version and --help. Reported by Matej Vela . 1998-10-04 Jim Meyering * lib/fnmatch.h: New file. (unused) * lib/fnmatch.c: New file. (unused) * lib/Makefile.am (noinst_HEADERS): Add fnmatch.h. 1998-10-03 Jim Meyering * man/Makefile.am: Switch to using help2man. (EXTRA_DIST): Add Makefile.summ. * man/help2man: Invoke program with --manhelp option only if --name=STRING not specified. Otherwise, this would fail with `yes'. * man/Makefile.summ: New file. * man/Makefile.maint: Include it. * man/help2man: New file. * man/GNUmakefile: New file. * man/Makefile.maint: New file. * man/*.x: New files. * man/*.1: Remove files. * src/md5sum.c (split_3): Rename local variable, to `escaped_filename'. (main): Output the leading backslash not just when there's a newline in the file name, but also when there's a backslash. Reported by Jim Dennis. * tests/md5sum/basic-1: Add tests with filenames containing newline and backslash characters. (for the bug fixed above) * tests/Makefile.am (EXTRA_DIST): Add Fetish.pm. * tests/Fetish.pm: New file. * tests/md5sum/basic-1: New file: rewrite of old tests to use Fetish.pm. * tests/md5sum/Test.pm: Remove file. * tests/md5sum/Makefile.am: Rewrite. 1998-09-19 Jim Meyering * src/ptx.c (program_name): Declare *not* to be const. 1998-08-29 Jim Meyering * src/cut.c: Don't assume ASCII. * src/pr.c: Likewise. * src/tail.c: Likewise. 1998-08-15 Jim Meyering * src/pr.c (usage): Reformat. * src/ptx.c: Add braces to suppress warning about ambiguous `else'. * lib/bumpalloc.h: Likewise. 1998-08-13 François Pinard * src/ptx.c: New file. * src/Makefile.am (bin_PROGRAMS): Add ptx. * lib/bumpalloc.h, lib/diacrit.h, lib/diacrit.c: New files. * lib/Makefile.am (libtu_a_SOURCES): Add diacrit.c. (noinst_HEADERS): Add bumpalloc.h and diacrit.h. 1998-08-09 Andreas Schwab * src/pr.c (long_options): Add long names for all options. (usage): Update help string. (main): Handle the special options --pages and --columns. 1998-08-02 Jim Meyering * lib/linebuffer.c (readline): Return zero upon error as well as upon end of file. From James Youngman. Ansideclify. 1998-08-01 Jim Meyering * src/sort.c (my_setlocale): Guard definition within #ifdef ENABLE_NLS. From Manfred Hollstein. 1998-07-30 Jim Meyering * tests/cut/Test.pm: Avoid broken pipe message for tests that fail with usage errors. * src/sort.c (usage): Add angle brackets to make `Report bugs...' message consistent with all the rest. 1998-07-28 Andreas Schwab * tests/cut/Test.pm: Avoid broken pipe for 'y' and 'z' tests. * src/sort.c (NEGATION_SIGN): Renamed from NEGATIVE_SIGN to avoid clash with . All uses changed. 1998-07-26 Jim Meyering * intl/localealias.c (read_alias_file): Avoid mixing `char*' and `unsigned char*' variables. Again for irix4. Mostly from Kaveh Ghazi. * src/join.c: Convert some char* dcls to `unsigned char*' and remove a cast -- to placate irix4's cc. * src/fmt.c (check_punctuation): Add cast to placate irix4's cc. Reported by Kaveh Ghazi. * src/md5sum.c (split_3): Add cast to placate irix4's cc. From Kaveh Ghazi. 1998-07-25 Jim Meyering * Version 1.22f. * tests/cut/Test.pm: Add tests for new --output-delimiter option, as well as for NUL input delimiter (--delimiter=''). * src/cut.c (cut_fields): Honor new --output-delimiter option. (main): Fix handling of --delimiter='' (-d ''). Until now, it has never worked as advertised. I guess no one tried it. * tests/sort/Test.pm: Add two tests relating to this. * src/sort.c (main): Stat all non-`-' input file files (and fail if a stat fails) when an output file is specified by `-o' but doesn't exist. Reported by Will Edgington. * tests/sort/Test.pm: New tests of -o. 1998-07-16 Jim Meyering * lib/Makefile.am (noinst_HEADERS): Add lchown.h. * lib/lchown.h: New file, just to define ENOSYS on systems that lack it. * lib/lchown.c: Include lchown.h. 1998-07-04 Jim Meyering * configure.in (AM_WITH_REGEX): Remove. Now the replacement macro, jm_WITH_REGEX, is bundled with the rest in jm_MACROS. * acconfig.h (WITH_REGEX): Remove undef. * src/csplit.c: Remove #ifdef around inclusion. * src/nl.c: Likewise. * src/tac.c: Likewise. * src/csplit.c (extract_regexp): Remove #if !WITH_REGEX...#endif block. * lib/Makefile.am (noinst_HEADERS): Remove rx.h. * lib/rx.c: Remove file. * lib/rx.h: Remove file. 1998-06-29 Jim Meyering * src/wc.c: Update calls to human_readable -- now there's one fewer arg. * lib/Makefile.am (libtu_a_SOURCES): Add argmatch.c. (noinst_HEADERS): Add argmatch.h. 1998-06-28 Jim Meyering * src/sys2.h: Add macro definitions for GNU libc *_unlocked wrappers. 1998-06-27 Jim Meyering * tests/pr/Test.pm: Add two tests for double spacing. * src/pr.c (print_page): If cols_ready_to_print is zero, break out of loop just before the double-space test. Reported by Michael Stutz. 1998-06-18 Jim Meyering * tests/Makefile.am.in (check): Depend on $(maint_gen) so `make maintainer-clean; ./configure; make check' works. 1998-05-25 Jim Meyering * configure.in (_GNU_SOURCE): AC_DEFINE it here. * acconfig.h (_GNU_SOURCE): Remove definition from @TOP@ section. [!_GNU_SOURCE]: Add #undef instead. 1998-05-16 Jim Meyering * configure.in (jm_MACROS): New wrapper macro. Remove uses of most jm_* macros. * src/tac.c (tac_seekable): Fix error in handling regex separators. * tests/tac/Test.pm (opt-b, opt-s, opt_sb, opt_r): New tests. (opt_br): New test -- exercises above-fixed bug. * lib/Makefile.am (EXTRA_DIST): Remove. Automake groks the `LIBOBJS=' lines from the m4/*.m4 macros, so the hack of including some custom-replaced C source file names here is no longer needed. * acconfig.h (chown): Add undef. (D_INO_IN_DIRENT): Likewise. (D_TYPE_IN_DIRENT): Likewise. (ssize_t): Likewise. 1998-05-09 Jim Meyering * src/tac.c (tac_seekable): Rename from tac_stream. Change `FILE *in' parameter to `int input_fd'. Adjust callers. 1998-05-03 Jim Meyering * po/: Update from gettext-0.10.35. * intl/: Likewise. * configure.in: Remove use of AC_LINK_FILES. (AC_OUTPUT): Remove po/Makefile-generating sed command. 1998-04-26 Jim Meyering * tests/tail/Test.pm: Disable test f-1, now that it fails. * src/sort.c (keycompare) (CMP_WITH_IGNORE): Don't return 0 from inside the keyspec-iterating loop. With this change, test 22a passes. Reported by Zvi Har'El. (strncoll): Remove bogus assertion. * tests/sort/Test.pm: Add tests for the above fix. * configure.in: Use jm_ASSERT. * acconfig.h: Add NDEBUG. * src/cut.c: Don't define NDEBUG. * src/csplit.c: Likewise. * src/join.c: Likewise. * src/sort.c: Likewise. * src/tr.c: Likewise. * src/cut.c: Don't define _GNU_SOURCE (now it's in config.h). * src/expand.c: Likewise. * src/fold.c: Likewise. * src/join.c: Likewise. * src/sort.c: Likewise. * src/tr.c: Likewise. * src/unexpand.c: Likewise. * src/uniq.c: Likewise. * src/tail.c (close_fd): New function -- converted from macro. [struct File_spec] (n_stat_calls): New member. [struct File_spec] (n_unchanged_stats): New member. (max_n_unchanged_stats): New global. Initialize new members. (xwrite): New function -- converted from macro. [struct File_spec] (pretty_name): Remove member. (pretty_name): New function. * src/md5sum.c (md5_check): Declare local, `md5num' as _unsigned_ char*. (hex_digits): Declare parameter `s' as _unsigned_ char*. (split_3): Declare parameter `u' as _unsigned_ char**. 1998-04-17 Jim Meyering * src/fmt.c (check_punctuation): Used unsigned char* pointers to avoid new warning. * src/join.c (xfields): Likewise. 1998-04-12 Jim Meyering * src/cat.c: Use STREQ macro rather than strcmp. * src/cksum.c: Likewise. * src/comm.c: Likewise. * src/csplit.c: Likewise. * src/cut.c: Likewise. * src/fmt.c: Likewise. * src/fold.c: Likewise. * src/head.c: Likewise. * src/join.c: Likewise. * src/md5sum.c: Likewise. * src/nl.c: Likewise. * src/paste.c: Likewise. * src/pr.c: Likewise. * src/split.c: Likewise. * src/sum.c: Likewise. * src/tac.c: Likewise. * src/uniq.c: Likewise. * src/wc.c: Likewise. 1998-04-11 Jim Meyering * lib/safe-read.h: New file. * lib/safe-read.c: Include it. * src/cat.c: Include it instead of merely declaring safe_read. * src/csplit.c: Likewise. * src/head.c: Likewise. * src/split.c: Likewise. * src/sum.c: Likewise. * src/tac.c: Likewise. * src/tail.c: Likewise. * src/tr.c: Likewise. * src/wc.c: Likewise. * lib/Makefile.am (noinst_HEADERS): Add safe-read.h. * src/wc.c [HAVE_INTTYPES_H]: Include inttypes.h. Declare counters to be of type uintmax_t. (write_counts): Use human_readable to format potentially-long-long numbers. Suggestion from Rogier Wolff. (wc): Declare per-file counters to be of type uintmax_t. Declare bytes_read to be ssize_t. * lib/Makefile.am (libtu_a_SOURCES): Add human.c. (noinst_HEADERS): Add human.h. * lib/human.c: New file. * lib/human.h: New file. 1998-04-04 Jim Meyering * configure.in (jm_AC_HEADER_INTTYPES_H): Use it. (jm_AC_TYPE_UINTMAX_T): Use it. (jm_PREREQ): Use it. * Makefile.am (ACLOCAL_AMFLAGS): Define this, so automake/aclocal know about the m4/ subdirectory. * Makefile.maint (aclocal-files): Remove now-unnecessary (with automake-1.2h and the above change) aclocal-related rules and includes. 1998-04-03 Jim Meyering * lib/closeout.c: New file. * lib/closeout.h: New file. * lib/Makefile.am (libtu_a_SOURCES): Add closeout.c. (noinst_HEADERS): Add closeout.h. 1998-03-31 Jim Meyering * lib/xstrtol.c: Merge with the version from fileutils. 1998-03-27 Jim Meyering * Makefile.am (AUTOMAKE_OPTIONS): Require 1.2h. 1998-03-23 Jim Meyering * acconfig.h: Remove HAVE_INTTYPES_H, now that m4/inttypes_h.m4 automatically handles it. 1998-03-19 Jim Meyering * src/system.h (SCHAR_MIN): Define. (SHRT_MIN): Define. od.c needs these when compiling with NCR's R2.0c C compiler. (TYPE_MAXIMUM): Cast result to `(t)' so this macro works with `unsigned char'. From Greg Wooledge. 1998-03-15 Jim Meyering * src/tail.c (tail_file): Merge largely-duplicated blocks of code. 1998-03-03 Paul Eggert * src/sort.c (xtmpfopen): Open temporary file exclusively, to foil a common denial-of-service attack. * src/tac.c (save_stdin): Likewise. 1998-02-16 Jim Meyering * configure.in (jm_FUNC_LSTAT): Use it. (jm_FUNC_STAT): Use it. * lib/Makefile.am (EXTRA_DIST): Add lstat.c and stat.c. 1998-02-06 Jim Meyering * configure.in: Don't use AM_MAINTAINER_MODE or AC_PATH_PROG(PERL, perl). (jm_PERL): Use this. 1998-02-04 Jim Meyering * tests/Makefile.am.in (EXTRA_DIST): Remove mk-script.pl. (mk_script): Set to ../mk-script. (x-tests): Use `$(PERL) -w -- $(mk_script)', not ./mk-script. Remove @MAINT@ cruft. (Makefile.am): Likewise. Remove @MAINT@ cruft. Now `missing' will explain the failure when people don't have Perl yet modify a file whose rebuilding would lead to the use of Perl. * tests/Makefile.am (EXTRA_DIST): Add mk-script. * tests/*/mk-script.pl: Remove files. * GNUmakefile: New file. * Makefile.am (EXTRA_DIST): Add GNUmakefile. Don't include Makefile.maint from here. It's included from GNUmakefile. 1998-01-30 Paul Eggert * configure.in (AC_LFS): Put before anything that can affect or use CPPFLAGS, LDFLAGS, or LIBS. 1998-01-25 Jim Meyering * Makefile.maint: New file. * Makefile.am: Move rules common to textutils, fileutils, sh-utils into Makefile.maint. Include Makefile.maint. (EXTRA_DIST): Add Makefile.maint. * src/cat.c (cat): Convert comma-expressions to pairs of semicolon-terminated stmts. Add braces around compound if/else stmts. 1998-01-24 Jim Meyering * src/tail.c (parse_obsolescent_option): Do not interpret `-f -n 1 ...' as obsolescent options. Accept new option: --sleep-interval=SECONDS (-s). (parse_options): Recognize it. (usage): Describe it. (tail_forever): Use it. (dump_remainder): Use it. * tests/tail/Test.pm (f-1): Add test for option-processing of `-f -n 1'. * tests/cut/: Rename directory to remove `-test' suffix. * tests/join/: Likewise. * tests/md5sum/: Likewise. * tests/pr/: Likewise. * tests/sort/: Likewise. * tests/tr/: Likewise. * configure.in (AC_OUTPUT): Reflect renamings in tests/. * tests/Makefile.am (SUBDIRS): Reflect renamings in tests/. * src/system.h (TYPE_MINIMUM): Add extra outer cast to work around bug in Cray C 5.0.3.0 when T == time_t. 1998-01-18 Jim Meyering * src/sort.c (strncoll, strncoll_s2_readonly, look_for_fraction, numcompare): Remove the `unsigned' from some `unsigned char*' parameter types. Add casts via UCHAR where necessary to avoid problems with unwanted sign extension. Based on a patch from Kaveh Ghazi to appease Irix4's cc compiler. 1998-01-17 Jim Meyering * src/split.c (next_file_name): Rewrite. This removes an artificial limit (albeit already high, at INT_MAX :-) on the number of files split could create. Reported by Ralf W. Stephan. 1998-01-16 Jim Meyering * src/sort.c (mergefps): Add braces to avoid ambiguous `else' stmt. (nls_set_fraction): Likewise. * src/sort.c: Guard inclusion of langinfo.h also with HAVE_LANGINFO_H, for Irix-4.0.5. From Kaveh Ghazi. * configure.in: Check for langinfo.h. * lib/getline.c: Make PARAMS-defining conditionals consistent. * lib/linebuffer.h: Likewise. * lib/long-options.h: Likewise. * lib/memcasecmp.h: Likewise. * lib/xstrtod.h: Likewise. * lib/xstrtol.h: Likewise. Suggestion from Kaveh Ghazi. * tests/head/Test.pm (fail-0): Disable test. It depends on sizeof(long) being 32 bits. Reported by Kaveh Ghazi. 1998-01-10 Jim Meyering * Version 1.22d. * src/system.h [HAVE_LIMITS_H]: Include limits.h (TYPE_SIGNED): Define. (TYPE_MINIMUM): Define. (TYPE_MAXIMUM): Define. (CHAR_BIT): Define. (SCHAR_MAX): Define. (UCHAR_MAX): Define. (SHRT_MAX): Define. (INT_MAX): Define. (UINT_MAX): Define. (LONG_MAX): Define. (ULONG_MAX): Define. * src/*.c: Remove definitions of those symbols. * src/csplit.c: Move inclusion of regex.h/rx.h to follow system.h since it now includes limit.h which defines RE_DUP_MAX. * src/nl.c: Likewise. * src/tac.c: Likewise. * lib/xstrtol.c (bkm_scale): Renamed from BKM_SCALE. Rewrite macro as function. Return a value. Update caller. Cast __ZLONG_MAX `__unsigned long int' before casting to double to avoid SunOS /bin/cc compiler bug. 1998-01-08 Jim Meyering * src/tac-pipe.c: New file. But not yet used. * src/Makefile.am (EXTRA_DIST): Add tac-pipe.c. * src/tac.c (tac_stream): Don't perform arithmetic on now-void* pointer result of xrealloc (until recently it was char*). * configure.in (AC_CHECK_FUNCS): Add nl_langinfo. * src/sort.c (inittables): Add && HAVE_NL_LANGINFO to the #if-test guarding the nls month-checking code. (nls_numeric_format): Remove unnecessary (and error-evoking w/SunOS' cc) `unsigned' from dcls of text and lim. (main): Cast lconvp->grouping to `unsigned char*' to appease SunOS's cc. 1998-01-03 Jim Meyering * lib/Makefile.am (AUTOMAKE_OPTIONS): Define to ../src/ansi2knr. * configure.in: Convert the .o suffix on files in LIBOBJS to $U.o so those files will be built via the ANSI2KNR-filtering rules if necessary. 1997-12-25 Jim Meyering * configure.in: Remove AC_DEFINE of _GNU_SOURCE. * acconfig.h (_GNU_SOURCE): Define if not already defined. Put this code in @TOP@ section. (_GNU_SOURCE): Remove #undef. 1997-12-22 Jim Meyering * configure.in: AC_DEFINE _GNU_SOURCE. * acconfig.h: Add _GNU_SOURCE. 1997-12-21 Jim Meyering * configure.in (AC_CHECK_HEADERS): Add stdlib.h. * src/system.h: Merge in things from fileutils' version of this file. * src/csplit.c: s/__P/PARAMS/. * src/fmt.c: s/__P/PARAMS/. * src/od.c: s/__P/PARAMS/. * src/pr.c: s/__P/PARAMS/. 1997-12-14 Jim Meyering * src/sys2.h: s/HAVE_DECLARATION_/HAVE_DECL_/g. 1997-12-13 Jim Meyering * src/fmt.c (main): Add some braces. Check return code from fclose of each input file. Close stdout and check for errors. * src/csplit.c (close_output_file): Check ferror before calling fclose. (main): Close stdout and check for errors. 1997-11-15 Jim Meyering * acconfig.h: Add mktime. * configure.in (jm_FUNC_MKTIME): Use it. * lib/mktime.c: New file. * lib/Makefile.am (EXTRA_DIST): Add mktime.c * intl/Makefile.in (distclean): Don't remove libintl.h here. * Makefile.am (DISTCLEANFILES): Remove it here instead. 1997-11-13 Jim Meyering * lib/strftime.c: Update from FSF. * m4/strftime.m4: Check for POSIX.2's %f format spec. 1997-11-12 Jim Meyering * src/system.h [!HAVE_MEMPCPY] (mempcpy): Define. * configure.in (AC_CHECK_FUNCS): Add mempcpy. 1997-11-09 Jim Meyering * configure.in (jm_FUNC_STRFTIME): Use it. * m4/jm-mktime.m4 (jm_FUNC_MKTIME): Check for localtime_r. * m4/strftime.m4 (jm_STRFTIME_PREREQS): Check for localtime_r. (jm_FUNC_GNU_STRFTIME): Use new macro. (jm_FUNC_STRFTIME): New macro. Likewise. Reported by Noel Cragg. 1997-11-08 Jim Meyering * m4/lfs.m4 (AC_LFS): New file/macro. * m4/Makefile.am (EXTRA_DIST): Add lfs.m4. * configure.in (AC_LFS): Use it. (AC_CHECK_FUNCS): Add fseeko. * src/od.c (fseeko): Define a stub if ! HAVE_FSEEKO. (skip): Use fseeko if available. Don't use lseek; it causes the stdio stream to become out of sync with respect to the underyling file descriptor. From Paul Eggert. 1997-10-16 Paul Eggert * configure.in (AC_CHECK_FUNCS): Add fseeko. * src/od.c (skip): Use fseeko. Don't use lseek; it causes the stdio stream to become out of sync with respect to the underyling file descriptor. 1997-11-06 Jim Meyering * src/sort.c (getmonth): Remove HAVE_ALLOCA #ifdefs. We always have alloca. (keycompare): Don't use variable size arrays (it's a gcc-extension). Rewrite code that increments new lengths when not `ignoring'. 1997-11-02 Jim Meyering * acconfig.h: Add malloc and realloc. * src/wc.c (main): New option, --max-line-length (-L). (wc, write_counts): Implement it. From Bruno Haible. 1997-10-26 Jim Meyering * lib/memcasecmp.c: Convert to upper case before comparing. This makes join -i work with sort -f. Reported by Arthur Pool. 1997-10-25 Jim Meyering * lib/xmalloc.c (xalloc_fail): Renamed from fixup_null_alloc. (xcalloc): #ifdef-out unused function. (xrealloc): Remove code to work around deficient versions of realloc. Now we have an autoconf-enabled replacement version. (xmalloc): Remove code to work around deficient versions of malloc. Now we have an autoconf-enabled replacement version. * lib/memcmp.c (rpl_memcmp): Rename from memcmp. * src/sort.c (NLS_STRNCMP) [!ENABLE_NLS]: s/strcmp/strncmp/. * lib/xmalloc.c (xalloc_fail_func): Initialize to 0, not NULL. * configure.in (jm_FUNC_MALLOC): Use it. (jm_FUNC_REALLOC): Use it. * lib/Makefile.am (EXTRA_DIST): Add malloc.c. * lib/malloc.c: New file. * m4/malloc.m4: New file. * m4/Makefile.am (EXTRA_DIST): Add malloc.m4. * lib/Makefile.am (noinst_HEADERS): Add xalloc.h. (EXTRA_DIST): Add realloc.c. (EXTRA_DIST): Add malloc.c. * src/system.h: Include xalloc.h. Remove dcls of xmalloc, xcalloc and xrealloc. * lib/xalloc.h: New file. * lib/xmalloc.c: Include xalloc.h. Change VOID to void. (xalloc_exit_failure): Renamed extern. (xalloc_msg_memory_exhausted): New extern. (xalloc_fail_func): New extern. (fixup_null_alloc): Use new variables. 1997-10-24 Jim Meyering * src/unexpand.c: Remove old-style xmalloc and xrealloc decls. (add_tabstop): Cast first arg of xrealloc to char*. * src/sort.c: Include xalloc.h. (xmalloc): Remove function. (xrealloc): Remove function. (main): Set xalloc_fail_func to cleanup. Set xalloc_exit_failure SORT_FAILURE. * src/paste.c: Remove old-style xmalloc and xrealloc decls. (paste_parallel): Cast first arg of xrealloc to char*. * src/od.c: Remove old-style xmalloc and xrealloc decls. (decode_format_string): Cast first arg of xrealloc to char*. * src/expand.c: Remove old-style xmalloc and xrealloc decls. (add_tabstop): Cast first arg of xrealloc to char*. * lib/xalloc.h: New file. * src/cut.c (ADD_RANGE_PAIR): Cast first arg of xrealloc to char*. (getstr): Cast xmalloc return value to char*. * src/csplit.c: Include xalloc.h. (xmalloc): Remove function. (xrealloc): Remove function. (main): Set xalloc_fail_func to cleanup. * src/*.c: Remove old-style xmalloc and xrealloc decls. * src/system.h: Add prototyped xcalloc, xmalloc and xrealloc decls. Suggestion from Achim Blumensath. 1997-10-23 Jim Meyering * Makefile.am (aclocal-files): Also depend on m4/Makefile.am. * src/system.h [!HAVE_DECLARATION_FREE]: Declare free. [!HAVE_DECLARATION_MALLOC]: Declare malloc. [!HAVE_DECLARATION_REALLOC]: Declare realloc. [!HAVE_DECLARATION_STPCPY]: Declare stpcpy. [!HAVE_DECLARATION_STRSTR]: Declare strstr. * src/cat.c: Remove stpcpy dcl. * src/csplit.c: Remove malloc and realloc dcls. * src/sort.c: Remove free, malloc, and realloc dcls. * src/tac.c: Remove malloc, and realloc dcls. * src/tr.c: Remove stpcpy dcl. On some systems, strstr and stpcpy are macros, so declaring them unconditionally gets syntax errors. Reported by Mark M. Kettenis. * configure.in: Use jm_CHECK_DECLS. * m4/Makefile.am (EXTRA_DIST): Add decl.m4 and check-decl.m4. 1997-10-22 Jim Meyering * m4/decl.m4: New file. * m4/check-decl.m4: New file. New macro, jm_CHECK_DECLS. 1997-10-16 Paul Eggert * src/od.c (LONG_MAX): Define if not defined. 1997-10-16 Jim Meyering * src/sort.c (look_for_fraction): Patch from Ørn Hansen. (getmonth): Compare the two month names only to the length of the string in the month table. Patch from Ørn Hansen. (NLS_STRNCMP): New macro. (strncoll_s2_readonly): New function. (inittables): Don't use temporary `comp' to hide type of comparator function. (nls_sort_month_comp): Declare parameters to be void* to match comparator function type required for qsort. (getmonth): Use NLS_STRNCMP rather than #ifdef. Use do-while, rather than while-loop. 1997-10-14 Jim Meyering * src/sort.c (_NL_ITEM) [!defined]: Define. From from Ørn E. Hansen. * src/sort.c: Use STREQ in place of most uses of strcmp. (NLS_STRCMP): Define. (getmonth): Remove ifdef and use NLS_STRCMP instead. Use HAVE_ALLOCA, not _HAVE_ALLOCA. (zaptemp): Make parameter `const'. * tests/sort-test/Test.pm: Add tests to exercise new fraccompare. * src/sort.c (CHARS_IN_ABM): Remove definition (inittables): Remove assumption that all abbreviated month names have length 3. (getmonth): Likewise. (main): Add #if's for more efficient code when using the GNU C library. From Ulrich Drepper. * src/sort.c (strncoll): Rename parameter to LEN. (keycompare): Move assignment out of if-expression. * lib/xstrdup.c: New file. * lib/Makefile.am (libtu_a_SOURCES): Add xstrdup.c * src/sort.c: Declare xstrdup. (my_setlocale): New function. (main): Guard against failure of strdup (use xstrdup) and setlocale. * src/system.h (STREQ): Define. * src/od.c (STREQ): Remove definition. * src/sort.c (look_for_fraction): Eliminate arbitrary limit on number of `groups'. Declare as void, not int. Patch from Ørn E. Hansen. (main): When determining whether we're in the C or POSIX locale, don't rely on the form of the string returned by setlocale. Suggestion from Ulrich Drepper. 1997-10-12 Jim Meyering * src/sort.c: Apply big patch from Ørn E. Hansen. Clean up, indent. (NLS_MEMCMP): Define. (keycompare): Use it instead of open-coded #ifdefs. (compare): Likewise. (NLS_MAP): Remove unused definitions. Replace with uses of UCHAR. (nls_locale_map): Remove dcl of unused file-scope array. 1997-10-10 Jim Meyering * tests/sort-test/Test.pm: Add a test. From William Lewis. 1997-10-07 Jim Meyering * src/*.c: Update bug-reporting address. * src/cat.c: Indent cpp directives to reflect nesting. * src/cksum.c: Likewise. * src/csplit.c: Likewise. * src/fmt.c: Likewise. * src/nl.c: Likewise. * src/paste.c: Likewise. 1997-09-27 Jim Meyering * m4/memcmp.m4: Integrate test to detect bug in memcmp from the Next x86 OpenStep C library. Test program from William Lewis. 1997-09-21 Jim Meyering * src/od.c [struct tspec] (hexl_mode_trailer): Rename from `trailer.' (dump_hexl_mode_trailer): Rename from dump_string_trailer. Use fputs and putchar instead of trivial or %-less printfs. (decode_one_format): Parenthesize each field_width assignment in an argument list to make the side effect a little more apparent. (write_block): Use fputs and putchar instead of trivial or %-less printfs. (dump_strings): Cast string_min to off_t to avoid long-standing warning. * src/od.c: Implement new `z' (hexl-mode) modifier. [struct tspec] (trailer): New field. (field_width): Likewise. (dump_string_trailer): New function. (decode_one_format): Save each field_width in the tspec. Patch from John Kodis. 1997-09-14 Jim Meyering * src/od.c (main) [--traditional]: Don't give diagnostic about there being more than three arguments if there are *no* arguments specified. Reported by Jochen Hein. 1997-09-13 Jim Meyering * tests/sort-test/Test.pm: Add tests 19a and 19b. 1997-07-21 Jim Meyering * src/tail.c (parse_obsolescent_option): #ifdef-out portability warning. 1997-07-19 Jim Meyering * src/sort.c (checkfp): Print the `disorder' message. Include both the number and the contents of the first out-of-order line, in addition to the file name. Change meaning of return value. (check): Don't print disorder message here. Adjust test of checkfp's return value. Feature suggestion from Karl Heuer. 1997-07-13 Jim Meyering * doc/Makefile.am (EXTRA_DIST): Remove explicit mention of texinfo.texi. Now, automake includes it automatically. * src/head.c: Include xstrtoul.h. Remove global variable, unit_size. (atou): Remove now-unused function. (parse_unit): Likewise. (string_to_ull): New function. (head): Take new parameter, count_lines. Use it instead of unit_size. Update callers. (head_file): Likewise. (main): Use string_to_ull, not atou/parse_unit. The problem was that overflow wasn't detected, so `head -c 4096m' was treated just like `head -c 0'. Reported by Jerome Abela. * tests/wc: New directory. * tests/head: New directory. * tests/Makefile.am (SUBDIRS): Add head and wc. * configure.in (AC_OUTPUT): Add tests/head/Makefile and tests/wc/Makefile. 1997-07-05 Jim Meyering * src/tail.c (parse_obsolescent_option): If POSIXLY_CORRECT is set, give a diagnostic and fail when there are two or more non-option arguments. 1997-07-04 Jim Meyering * tests/tac/mk-script.pl: Redirect output of cmp to /dev/null. * configure.in (ALL_LINGUAS): Add Norwegian (no). * Makefile.am (aclocal-files): Look in source directory, not build dir. From Andreas Schwab. Sun Jun 15 06:36:41 1997 Jim Meyering * src/pr.c (init_header) [T_BUF_FMT]: Output the 4-digit year (not the 2-digit abbreviation) in each page header. Reported by Noah Friedman. Sat Jun 14 12:29:12 1997 Jim Meyering * src/cut.c (cut_fields): Detect when the input is empty and handle that special case. Before `cut -f1 < /dev/null' would improperly output a single newline. Reported by Phil Richards. Sun Apr 27 15:10:58 1997 Jim Meyering * man/Makefile.am (DISTFILES): Add $(man_MANS). * configure.in (ALL_LINGUAS): Add Czech (cs) and Swedish (sv). Thu Apr 3 21:14:02 1997 Jim Meyering * m4/Makefile.am (EXTRA_DIST): Update file list. * tests/cut-test/Test.pm: Add test from Phil Richards. Sat Mar 22 20:29:10 1997 Jim Meyering * missing: New file -- from the automake-1.1m distribution. Fri Mar 21 23:56:41 1997 Jim Meyering * Makefile.am (AUTOMAKE_OPTIONS): Require automake-1.1l. (aclocal.m4): Use aclocal's new -I option. Thu Mar 13 21:46:04 1997 Jim Meyering * src/tr.c (main): Make sure c1 is not -1 before using it as an array index. Patch from Greg McGary. Although this is truly a bug, I believe it would not cause tr to misbehave on most systems. I could not construct a test case with which this bug causes tr to generate invalid output. * tests/pr-test/Test.pm: Add test that -o 0 works. * src/pr.c (main): Allow use of 0 (zero) as the margin offset argument to the -o option. Patch from Gary Anderson. Fri Feb 28 22:32:51 1997 Jim Meyering * src/uniq.c (usage): Fix typo in --help output. From Andreas Schwab. Tue Feb 25 20:34:51 1997 Jim Meyering * lib/Makefile.am (noinst_HEADERS): Add obstack.h. (libtu_a_SOURCES): Add obstack.c. Sun Feb 16 08:30:29 1997 Jim Meyering * tests/Makefile.am (SUBDIRS): Add tac. * tests/tac: New directory. * configure.in (AC_OUTPUT): Add tests/tac/Makefile. * tests/pr-test/{tt-0FF, tt-bl, tta3-0FF, ttb3-0FF, tt-FF, tt-t, tta3-FF, ttb3-FF}: New files. Renamed (s/T/tt/) to avoid name clashes on case-independent filesystems. * tests/pr-test/Test.pm (Tests '7.*'): Reflect file-renaming. * src/tail.c (parse_obsolescent_option): Reverse order of args in diagnostic. Remove `' quotes in diagnostic. (parse_options): Remove `' quotes in diagnostic. Sat Feb 8 22:43:45 1997 Jim Meyering * src/tail.c (parse_obsolescent_option): Give warning diagnostic for (but now accept) obsolescent usage with more than one file argument. Sun Feb 2 23:06:59 1997 Jim Meyering * src/join.c: Move alloca-related preprocessor code into system.h. * src/od.c: Remove alloca-related preprocessor code. * src/system.h: Add alloca-related preprocessor code. Sat Feb 1 07:21:43 1997 Jim Meyering * tests/pr-test/*: Rename files to avoid exceeding 14-character limit. * tests/pr-test/Test.pm: Reflect renamings. * tests/pr-test/mk-script.pl (spec_to_list): Warn about all filenames that exceed max-length before dying. * src/pr.c (init_parameters): For compatibility: use default separator `TAB' with full length lines. From Roland Huebner. Fri Jan 31 19:53:54 1997 Jim Meyering * src/tac.c: Rename globals buffer and buffer_size to have G_ prefix to avoid shadowing local variables. * lib/long-options.c (parse_long_options): Compare getopt_long return value against -1, not EOF. Use NULL, not `(int *) 0' as last parameter in getopt_long call. * src/pr.c (add_line_number): Rename from `number' to avoid shadowing local variables. * src/*.c: Compare getopt_long return value against -1, not EOF. Use NULL, not `(int *) 0' as last parameter in getopt_long call. (usage): Bracket bug-reporting address with <> and append a period. Wed Jan 29 20:54:24 1997 Jim Meyering * tests/tr-test/Test.pm: Do each test twice: get input via REDIR and PIPE. * tests/tail/Test.pm (test_vector): Do each (non-stdin-requiring) test three times. * tests/pr-test/Test.pm: Remove common_option_prefix flag. (test_vector): Prepend the common option here instead. * tests/md5sum-test/Test.pm: Remove input_via_stdin flag. Do each test twice: get input via REDIR and PIPE. * tests/cut-test/Test.pm: Do each test three times. * tests/cut-test/mk-script.pl: Allow each test to be run any or all of three different ways. Program input may be specified via a file or files listed on the command line, via input redirection (if there's only one file), or via a pipe. Tue Jan 28 20:54:06 1997 Jim Meyering * tests/cut-test/mk-script.pl: Remove `t' prefix on all generated test file names. Change suffixes from (.in, .exp, .out, .err) to (.I, .X, .O, .E). Ensure that no test file (generated or maintainer-supplied) has a name longer than 14 characters. Sun Jan 26 12:49:50 1997 Jim Meyering * Version 1.22. * src/tail.c (parse_options): Add quotes to make messages identical. Sat Jan 25 00:12:29 1997 Jim Meyering * src/tail.c (parse_obsolescent_option): Comment. (parse_options): Remove unnecessary goto and label. * tests/sort-test/mk-script.pl: Interpret `input arg is a hash reference' as meaning that the tested program will read no input. Most of the tests for the date program use this feature. Also for date, the generated script now reflects specification (in Test.pm) of default and per-test environment settings. * src/csplit.c: Reflect changes to xstrtol and xstrtoul interfaces. * src/fold.c: Likewise. * src/head.c: Likewise. * src/join.c: Likewise. * src/nl.c: Likewise. * src/od.c: Likewise. * src/pr.c: Likewise. * src/uniq.c: Likewise. * lib/xstrtoul.h (XSTRTOL_H): Undefine it. * lib/xstrtol.h [!_STRTOL_ERROR]: Define the type `enum strtol_error' only if it hasn't already been defined. (_STRTOL_ERROR): Undefine. * lib/xstrtol.c (__xstrtol): Change interpretation of VALID_SUFFIXES == NULL vs VALID_SUFFIXES == "". Use the former when any suffix is valid, the later when no suffix is allowed. Fri Jan 24 23:36:00 1997 Jim Meyering * lib/long-options.c (parse_long_options): Reset optind to zero before just returning so that getopt internals get initialized from the probably-new parameters when/if getopt is called later. From Ulrich Drepper. Thu Jan 23 19:17:03 1997 Jim Meyering * src/tail.c (parse_obsolescent_option): Recognize and fail for more malformed obsolescent options. Makes for better diagnostics. Wed Jan 22 21:34:50 1997 Jim Meyering * tests/Makefile.am (SUBDIRS): Add tail. * tests/tail: New directory. * configure.in (AC_OUTPUT): Add tests/tail/Makefile. * src/tail.c (parse_options): New function. (parse_obsolescent_option): New function. (main): Use the new functions instead of open-coding them. This better fixes the bug in handling obsolescent `+Nc' options. General cleanup. Tue Jan 21 22:49:00 1997 Jim Meyering * src/tail.c (main): Properly handle what the POSIX spec calls the `obsolescent' usage (e.g., tail +2c). It didn't work. Reported by Karl Heuer. Sun Jan 12 22:13:27 1997 Jim Meyering * tests/tr-test/Test.pm: Rename tests so that the associated files have names that are distinct even on filesystems on which file names are case insensitive. Reported by Fred Fish. * tests/cut-test/Test.pm: Likewise. * tests/tr-test/Makefile.am: Regenerated to reflect renamed tests. * tests/cut-test/Makefile.am: Likewise. Wed Jan 8 16:38:24 1997 Jim Meyering * Version 1.21. * lib/getopt.c (_getopt_internal): Use `_', rather than the (sometimes-)expansion `gettext'. Tue Jan 7 22:50:13 1997 Jim Meyering * lib/getopt.c: New (more POSIX compliant) version from GNU libc. [_]: Define to gettext also if ENABLE_NLS is defined. This is temporary. Thu Jan 2 21:17:50 1997 Jim Meyering * src/pr.c (init_fps): Initialize lines_stored field in three places. This avoids uninitialized memory reads in close_file. Wed Jan 1 17:32:18 1997 Jim Meyering * configure.in (AC_ARG_PROGRAM): Remove explicit use. It's implicit in AM_INIT_AUTOMAKE. From Wayne Stewart. Sun Dec 29 23:42:57 1996 Jim Meyering * src/pr.c (init_header): Plug a small memory leak by using stack rather than heap for a 15-byte temporary buffer. Sat Dec 28 00:03:23 1996 Jim Meyering * src/pr.c (TAB_WIDTH): Parenthesize uses of parameters and renamed from tab_width. (POS_AFTER_TAB): Renamed from pos_after_tab. Define in terms of TAB_WIDTH rather than duplicating the expression. Fri Dec 27 17:29:02 1996 Jim Meyering * tests/join-test/mk-script.pl: Fix a couple of thinkos and typos. Now it works with pr's tests, too. * tests/pr-test/Makefile.am (explicit): Regenerate list of files, this time with duplicates removed. * src/pr.c (usage): Break long string literal into two separate printf statements to accomodate default maximum of 2048 characters for Irix-4.0.5. Reported by Kaveh Ghazi. * Makefile.am (aclocal.m4): No longer depend on acinclude.m4. * tests/Makefile.am (SUBDIRS): Add pr-test. * configure.in: AC_REQUIRE version 2.12 of autoconf. AC_OUTPUT: Add tests/pr-test/Makefile. Sun Dec 22 08:11:27 1996 Jim Meyering * acinclude.m4: Move macros to individual files in new directory, m4/. See ChangeLog entries in sh-utils for the details. * configure.in: Increment version to 1.20a. * tests/md5sum-test/Makefile.am: Use same framework as other tests. * tests/md5sum-test/Test.pm: New file. Sat Dec 21 20:32:58 1996 Jim Meyering * tests/Makefile.am (EXTRA_DIST): Add README and Makefile.am.in. * tests/Makefile.am.in: New file. * tests/sort-test/Makefile.am: Clone the Makefile.am from join-test. Replace only the definition of $x and the auto-generated lists of test files. * tests/cut-test/Makefile.am: Likewise. * tests/tr-test/Makefile.am: Likewise. * tests/join-test/Makefile.am: Add @MAINT@-protected rule for verifying that Makefile.am is consistent with Test.pm. Fri Dec 20 00:08:36 1996 Jim Meyering * tests/join-test/Makefile.am (MAINTAINERCLEANFILES): Use $(in1) and $(in2), not $(in). * tests/join-test/mk-script.pl: New option: --list. Generate lists of files used/generated. Thu Dec 19 23:28:41 1996 Jim Meyering * tests/join-test/mk-script.pl: Rewrite to be more general -- so it can be used for join, cut, sort, tr, and soon, pr test suites. * tests/tr-test/mk-script.pl: Symlink through CVS repo to join's mk-script.pl. * tests/sort-test/mk-script.pl: Likewise. * tests/cut-test/mk-script.pl: Likewise. * tests/tr-test/Test.pm: Adapted for new mk-script.pl. * tests/join-test/Test.pm: Likewise. Tue Dec 17 16:48:51 1996 Jim Meyering * tests/join-test/Makefile.am (.pl): Make generated file read-only so I don't accidentally modify it instead of the one with the .pl suffix. * acinclude.m4 (jm_SYS_PROC_UPTIME): Require AC_PROG_CC, rather than now-obsolete AC_C_CROSS. (jm_FUNC_MKTIME): When redefining, use rpl_ prefix, not gnu_ one since there's nothing GNU-specific about the replacement. Contrast with gnu_ prefix added to strftime. (jm_FUNC_MEMCMP): New macro. * configure.in: Use jm_FUNC_MEMCMP instead of AM_FUNC_MEMCMP. * acconfig.h: Add memcmp. * README: Remove warning about memcmp. The new macro handles it. Mon Dec 16 23:03:27 1996 Jim Meyering * tests/Makefile.am (pkgdata_DATA): Remove unnecessary assignment. From Eric Backus. (AUTOMAKE_OPTIONS): Likewise. Sat Dec 14 14:51:50 1996 Jim Meyering * acinclude.m4 (jm_SYS_PROC_UPTIME): Fix typo in cache variable name for cross compiling. Remove definitions of gettext-related macros. The gettext installation procedure installs corresponding .m4 files so that aclocal will use them. 1996-12-12 Paul Eggert * system.h (ISDIGIT): Replace with smaller, faster edition that yields nonzero only on ASCII digits. (ISDIGIT_LOCALE): New macro, with same meaning that ISDIGIT used to have. * tr.c (is_char_class_member): Use ISDIGIT_LOCALE instead of ISDIGIT to test for characters in CC_DIGIT class. * sort.c (digits): Remove; subsumed by new ISDIGIT. (inittables): Remove initialization of `digits'. (fraccompare, numcompare, main): Use ISDIGIT (x) instead of digits[x]. (fraccompare, numcompare): Avoid casts to unsigned char that are no longer needed. * csplit.c (get_format_width, get_format_prec): Avoid unnecessary comparison of digit to '\0'. Thu Dec 12 23:42:51 1996 Jim Meyering * src/sort.c (usage): Clarify description of -u option. From Karl Berry. Wed Dec 11 19:32:18 1996 Jim Meyering * Version 1.20. Tue Dec 10 00:15:50 1996 Jim Meyering The file name, `build-script.pl' is longer than 14 characters. Rename it. * tests/cut-test/mk-script.pl: Rename from build-script.pl. * tests/cut-test/Makefile.am: Reflect renaming. * tests/tr-test/mk-script.pl: Rename from build-script.pl. * tests/tr-test/Makefile.am: Reflect renaming. * tests/sort-test/mk-script.pl: Rename from build-script.pl. * tests/sort-test/Makefile.am: Reflect renaming. * tests/join-test/mk-script.pl: Rename from build-script.pl. * tests/join-test/Makefile.am: Reflect renaming. Reported by Karl Heuer. * configure.in (VERSION): Bump to 1.19r. (AC_ISC_POSIX): Remove kludgy macro. Use this test instead: (LIBS): Add -lcposix if it contains strerror. Patch from Karl Heuer. Sun Dec 8 07:22:38 1996 Jim Meyering =========== Update for automake-1.1k. * Makefile.am (AUTOMAKE_OPTIONS): Require 1.1k. * configure.in (AM_CONFIG_HEADER): Use it. (AC_OUTPUT): Remove stamp-h timestamping statement. Now, AM_CONFIG_HEADER does it automatically. * lib/Makefile.am (noinst_LIBRARIES): Rename to libtu.a as per new automake requirement. Rename tu_* variables to libtu_a_*. Wed Dec 4 21:03:18 1996 Jim Meyering * src/pr.c: Apply big patch from Roland Huebner. * src/tr.c (main) [!POSIXLY_CORRECT]: Allow the identity mappings: [:upper:] to [:upper:] and [:lower:] to [:lower:]. (main) [POSIXLY_CORRECT]: Give a more specific diagnostic for the invalid identity mappings [:upper:] to [:upper:] and [:lower:] to [:lower:]. (class_ok): Update table to reflect that tr now allows these identity mappings. Suggestion from Risto Kankkunen. Thu Nov 28 00:31:18 1996 Jim Meyering * configure.in (AM_GNU_GETTEXT): Renamed from ud_GNU_GETTEXT for gettext-0.10.25. Tue Nov 26 23:05:14 1996 Jim Meyering * src/sort.c (fraccompare): Add explicit `int' in `register' dcls, to avoid new warning from gcc. Sun Nov 24 21:02:15 1996 Jim Meyering * tests/sort-test/Makefile.am (t): Add names of those 5 tests. * tests/sort-test/Test.pm: Add 5 tests to exercize just-fixed code. * src/sort.c (set_ordering): Revert 1994-05-04 change to this function so that blanks are not unconditionally ignored when computing start and end positions for numeric keys with explicit character offsets. Reported by Markus Demleitner. Sat Nov 23 16:07:08 1996 Jim Meyering * cat.c (usage): Alphabetize option descriptions the way sort -f would. * fmt.c (usage): Likewise. * join.c (usage): Likewise. * md5sum.c (usage): Likewise. * od.c (usage): Likewise. * pr.c (usage): Likewise. * sort.c (usage): Likewise. * split.c (usage): Likewise. * wc.c (usage): Likewise. From Karl Berry. Fri Nov 22 20:20:37 1996 Jim Meyering * src/csplit.c (process_regexp): Update CURRENT_LINE only if the new value would be larger. This avoids the infinite loop otherwise provoked by situations like this: printf "\na\n" | csplit - '/a/-1' '{*}' From Jens Schmidt. Tue Nov 19 23:16:57 1996 Jim Meyering * configure.in (AM_SANITY_CHECK_CC): Remove it. Autoconf-2.11 has this built-in. Sat Nov 2 08:50:01 1996 Jim Meyering * doc/Makefile.am (MAKEINFO): Set to makeinfo --no-split. Otherwise, the generated info files have names longer than the 14-byte max of some old systems. Reported by Karl Heuer. Fri Nov 1 21:33:16 1996 Jim Meyering * lib/strtol.c: Update from GNU libc. * tests/sort-test/Test.pm: Add a test to exercize this fix. * src/sort.c [!ENABLE_ASSERTIONS]: Guard NDEBUG definition. (checkfp): Fix off-by-one error that resulted in writing one byte beyond the end of a malloc'd buffer. It caused `sort -c' to segfault on Linux systems having a relatively recent libc. Before, running the command, perl -e "print 'x' x 30, \"\n\";"|sort -c would provoke the memory overrun (though not necessarily the failure). Add an assertion. Reported by Risto Kankkunen. Thu Oct 31 17:48:24 1996 Jim Meyering * src/sort.c (key_init): New function. (main): Use key_init in the two places where it was open-coded. This fixes a UMR of the general_numeric field. * src/join.c (decode_field_spec): Always give FIELD_INDEX a value. This avoids a spurious UMR report from purify. (prjoin): Add an assertion. (add_field): Update assertion. * src/tr.c (es_free): New function. (parse_str): Use it to plug a small memory leak. Wed Oct 23 22:02:24 1996 Jim Meyering * lib/md5.c: Update from GNU libc. * lib/md5.h: Update from GNU libc. Mon Oct 21 16:48:12 1996 Jim Meyering * src/tr.c (validate): Remove TAB from diagnostic, for consistency. Sun Oct 20 13:44:07 1996 Jim Meyering * lib/Makefile.am (INCLUDES): Add -I../intl. Reported by Eric Backus. * Makefile.am (AUTOMAKE_OPTIONS): Set to `gnits' (EXTRA_DIST): Remove. * configure.in: Remove README_ALPHA related code. Automake takes care of it automatically now when in gnits mode. * lib/md5.c (md5_process_bytes): Cast void* pointer to char* before doing arithmetic with it. Sat Oct 19 23:13:54 1996 Jim Meyering * lib/md5.c: Update from GNU libc. * lib/md5.h: Update from GNU libc. Fri Oct 18 00:08:04 1996 Jim Meyering * configure.in (AC_OUTPUT): Add tests/join-test/Makefile. * tests/Makefile.am (SUBDIRS): Add join-test. Tue Oct 15 23:25:31 1996 Jim Meyering * src/tr.c (get_next): Rewrite to treat lower/upper mapping as a special case. (main): Write separate loops to initialize mapping for lower->upper and upper->lower conversion. Reported by Arne Henrik Juul. Sun Oct 13 12:52:42 1996 Jim Meyering * lib/xstrtol.c (__xstrtol): Return an error for invalid suffix. Before, e.g., `split -b 1M' would be silently accepted and treated like `split -b 1'. Reported by Franc,ois. * src/split.c (usage): Remove parameter, REASON. (main): Update callers to use combination of error (0, 0, ... and usage (EXIT_FAILURE). (main): When a string operand cannot be converted to a number of bytes or lines, include that string in the diagnostic. Sat Oct 12 00:05:11 1996 Jim Meyering * configure.in (AC_REPLACE_FUNCS): Add strpbrk and strtol. * lib/Makefile.am (EXTRA_DIST): Remove. Thu Oct 10 23:41:36 1996 Jim Meyering * lib/md5.c (md5_stream): Apply fix from Ralph Loader via Ulrich Drepper. Wed Oct 9 07:26:40 1996 Jim Meyering * lib/Makefile.am (tu_DEPENDENCIES): Depend on $(tu_LIBADD). Tue Oct 8 21:32:17 1996 Jim Meyering * lib/error.c: Include error.h to align with libit -- under protest, François :-). Sun Oct 6 08:02:28 1996 Jim Meyering * acinclude.m4 (AC_HEADER_SYS_TIME_H, AM_FUNC_MKTIME, AM_FUNC_ERROR_AT_LINE, AM_FUNC_OBSTACK): Remove definitions. They're included in automake-1.1f. * configure.in (AM_FUNC_ERROR_AT_LINE): Use it instead of open-coding it. Sat Oct 5 12:40:22 1996 Jim Meyering * lib/memcpy.c [HAVE_CONFIG_H]: Include config.h. From Karl Berry. Fri Oct 4 07:20:37 1996 Jim Meyering * configure.in (GNU_PACKAGE): New variable. (PACKAGE_VERSION): Remove it. * acconfig.h (GNU_PACKAGE): New variable. (PACKAGE_VERSION): Remove now-unused variable. * lib/long-options.c (parse_long_options): Separate package name and version number to accomodate new --version output requirement. * lib/long-options.h: Update prototype. Thu Oct 3 23:27:31 1996 Jim Meyering * src/md5sum.c (main): Remove dead code that used to print --version output. Tue Oct 1 06:54:22 1996 Jim Meyering * configure.in: Use result of AM_FUNC_STRTOD to set POW_LIBM. * src/Makefile.am (sort_LDADD): Use any library (-lm) that gets substituted for @POW_LIBM@. * tests/md5sum-test/md5-rfc (md5sum): Use ../../src/md5sum. Mon Sep 30 23:35:46 1996 Jim Meyering * tests/md5sum-test/md5-rfc (md5sum): Set to $srcdir/md5sum. * acinclude.m4: Update definitions from gettext-0.24. Sun Sep 29 20:04:53 1996 Jim Meyering * tests/sort-test/build-script.pl: Die if close fails. * tests/tr-test/build-script.pl: Likewise. * configure.in (AC_OUTPUT): Add tests/cut-test/Makefile. * tests/Makefile.am (SUBDIRS): Add cut-test. * tests/cut-test: New directory. Fri Sep 27 22:22:09 1996 Jim Meyering * configure.in (ALL_LINGUAS): Add spanish (es). Thu Sep 26 21:02:54 1996 Jim Meyering * src/md5sum.c (usage): Remove references in --help output to the now-deprecated --string option. Support for it will be removed soon. Using this option can provoke a memory access violation on some systems because of (unavoidable in an efficient implementation) alignment assumptions made by functions in md5.c. md5sum.c could leave the option in and accomodate the alignment restriction, but this option was intended solely for testing, and the tests are now all file-oriented, so it's not necessary. * tests/md5sum-test/md5-rfc: Rewrite to avoid use of --string option. Wed Sep 25 21:43:10 1996 Jim Meyering * tests/sort-test/Makefile.am (t): Factor out .in suffix. Include new test: t16a. * tests/tr-test/Makefile.am (t): Likewise. * src/Makefile.am (EXTRA_DIST): Remove md5-test.rfc. Mon Sep 23 10:00:50 1996 Jim Meyering * tests/md5sum-test/Makefile.am (EXTRA_DIST): Distribute $(TESTS). Sun Sep 22 09:24:22 1996 Jim Meyering * src/sort.c (keycompare): Declare translate to be unsigned char *. Otherwise, sign extension caused misordering when using e.g. -f. Reported by Erick Branderhorst. * tests/sort-test/Test.pm: Add Erick's test for that fix. * tests/Makefile.am (SUBDIRS): Add md5sum-test. Sat Sep 21 13:34:59 1996 Jim Meyering * src/Makefile.am (check): Remove tests for md5sum. * tests/md5sum-test: New directory. Put them here instead. * configure.in (AC_OUTPUT): Add tests/md5sum-test/Makefile. Thu Sep 19 08:54:05 1996 Jim Meyering * tests/: Rename from checks. * Makefile.am (SUBDIRS): Change `checks' to `tests'. * configure.in (AC_OUTPUT): Likewise. * configure.in ($PACKAGE, $VERSION): Don't AC_DEFINE_UNQUOTED these. AM_INIT_AUTOMAKE now does it (as of automake-1.1e). Sun Sep 15 23:08:48 1996 Jim Meyering * Makefile.am (EXTRA_DIST): Remove acinclude.m4. Thu Sep 12 17:05:23 1996 Jim Meyering * lib/Makefile.am (EXTRA_DIST): Add getline.c. (tu_SOURCES): Remove getline.c From Kaveh Ghazi. Sun Sep 8 13:55:18 1996 Jim Meyering * checks/sort-test/build-script.pl: Prefix $in with \$srcdir/ so make check works with VPATH build. From Uli Drepper. Sat Sep 7 12:25:42 1996 Jim Meyering * src/system.h: Add comments justifying IS* versions of ctype.h macros. * src/fmt.c (check_punctuation): Use ISPUNCT instead of ispunct. (get_line): Use ISSPACE instead of isspace. From Bruno Haible. Mon Sep 2 10:34:46 1996 Jim Meyering * src/*.c (usage): Tell where to report bugs. Sun Aug 25 22:50:47 1996 Jim Meyering * checks/sort-test/Makefile.am (EXTRA_DIST): Add $x-tests. ($x-tests): Guard dependencies with @MAINT@. * checks/tr-test/Makefile.am: Likewise. Sat Aug 24 14:25:41 1996 Jim Meyering * src/cat.c (cat): Cast first arg to stpcpy to char* to avoid warnings. Wed Aug 21 22:28:26 1996 Jim Meyering * configure.in (ALL_LINGUAS): Add polish (pl). Sun Aug 18 09:34:42 1996 Jim Meyering * configure.in (AC_OUTPUT): Add checks/sort-test/Makefile. * checks/Makefile.am (SUBDIRS): Add sort-test. Sat Aug 17 18:57:17 1996 Jim Meyering * Makefile.am (EXTRA_DIST): Add acinclude.m4. Tue Aug 13 21:47:23 1996 Jim Meyering * src/Makefile.am (EXTRA_DIST): Remove ansi2knr.1 and ansi2knr.c. Automake includes them automatically. * configure.in (AM_INIT_AUTOMAKE): Use it. (AM_PROG_INSTALL): Remove. AM_INIT_AUTOMAKE does this. (AC_PROG_MAKE_SET): Likewise. Sun Aug 11 20:49:21 1996 Jim Meyering * configure.in: Rename a few macros for automake-1.1c. (AC_REPLACE_GNU_GETOPT): Remove. (AC_PATH_PROG): Find perl -- needed only for checks if you change or remove things with e.g. make maintainerclean. * lib/Makefile.am (tu_LIBADD): Remove @REGEXOBJ@. New AM_WITH_REGEX adds .o files to @LIBOBJ@. (tu_SOURCES): Add getopt.c and getopt1.c. (EXTRA_DIST): Remove getopt.c and getopt1.c. * acinclude.m4: New file -- derived from aclocal.m4. * aclocal.m4: This file is now generated by the aclocal program (which comes with the automake package.) Sun Aug 4 10:50:46 1996 Jim Meyering * checks/: New directory. * checks/Makefile.am: New file. * configure.in (AC_OUTPUT): Add new directories: checks, checks/tr-test. * Makefile.am (SUBDIRS): Likewise. Sat Jul 27 17:22:14 1996 Jim Meyering * src/system.h (bindtextdomain) [!ENABLE_NLS]: Undefine to avoid redefinition warnings on solaris. (textdomain) [!ENABLE_NLS]: Likewise. Thu Jul 25 23:06:35 1996 Jim Meyering * src/unexpand.c (unexpand): Move EOF check to follow code that processes and generates output for pending spaces. Before `printf ' ' |unexpand -t 1,2' generated no output. * src/md5sum.c (md5_check): Remove parameter, BINARY. Rename local TYPE_FLAG to BINARY, so md5_file uses the binary mode from the input stream rather than the one from the command line. (main): Fail with a diagnostic if --binary or --text is specified when verifying checksums. Reported by Eli Zaretskii . Sun Jul 21 11:58:48 1996 Jim Meyering * configure.in: Add check for error_at_line so systems with GNU libc don't compile and link with distributed error.c. * lib/Makefile.am (tu_SOURCES): Remove error.c. (EXTRA_DIST): Add error.c. * configure.in (AC_REPLACE_FUNCS): Replace getline.c. Add related check for the getdelim function. From Ulrich Drepper -- as done in gettext. Sat Jul 20 15:59:36 1996 Jim Meyering * configure.in (PACKAGE_VERSION): Use space instead of hyphen to separate PACKAGE and VERSION. * aclocal.m4 (AM_SANITY_CHECK_CC): New macro. Derived from macros from Bruno Haible and from Cygnus. * configure.in (AM_SANITY_CHECK_CC): Use it. Wed Jul 17 00:30:28 1996 Jim Meyering * install-sh: Get version with MIT copyright. Tue Jul 16 00:09:37 1996 Jim Meyering * aclocal.m4 (fp_C_PROTOTYPES): Improved version from François Pinard. Update all other macros to reference $fp_cv_prog_cc_stdc, rather than $ac_cv_prog_cc_stdc. * src/od.c (OFF_T_MAX): Remove definition. (main): Compare against LONG_MAX rather than OFF_T_MAX. Mon Jul 15 23:42:57 1996 Jim Meyering * Many files: Update FSF address. Sun Jul 14 16:53:50 1996 Jim Meyering * src/unexpand.c (unexpand): Fix bug that contributed to endless loop when invoking `echo ' ' |unexpand -t 1,2': use print_tab_index, not tab_index in inner flush_pend: while loop. From Keith Owens . * src/unexpand.c [HAVE_LIMITS_H]: Include limits.h for INT_MAX. [!INT_MAX]: Define it. (main): Append INT_MAX to command-line-specified tab list to ensure termination in unexpand's inner loop. Derived from a patch from Keith Owens. Thu Jul 11 22:04:36 1996 Jim Meyering * Version 1.19. * configure.in (VERSION): Bump to 1.19. Wed Jul 10 22:57:29 1996 Jim Meyering * aclocal.m4 (fp_PROG_CC_STDC): Include sys/stat.h in test program so that DYNIX/ptx V4.1.3 doesn't use `-Xc -D__EXTENSIONS__' -- with those options on that system, sys/stat.h gets compile errors. With help from Marcus Daniels. * getopt.c: Update from gettext-0.10.23. * getopt1.c: Likewise. * getopt.h: Likewise. Tue Jul 9 18:07:23 1996 Jim Meyering * src/fmt.c (MAXCOST): Use `(unsigned long)1' rather than `(COST)1' so the left operand of the << isn't signed. From Kjetil Torgrim Homme. * po/Makefile.in.in (install-data): Don't install NLS files when they're not requested. From Ulrich Drepper. Reported by Kjetil Torgrim Homme . Fri Jul 5 21:55:58 1996 Jim Meyering * po/Makefile.in.in ($(PACKAGE).pot): Merge from gettext-0.10.23. * ABOUT-NLS: Likewise. * intl/*: Likewise. * aclocal.m4: Likewise. Thu Jul 4 07:24:54 1996 Jim Meyering * src/uniq.c (usage): Remove space before newline in usage message. * src/md5sum.c (MIN_DIGEST_LINE_LENGTH): New macro. [NEWLINE_REPLACEMENT_STRING*]: Remove macros. (main): Output a leading backslash for a line describing a file whose name contains a newline. Then translate each NEWLINE byte in the file name to the string, "\\n", and each backslash to "\\\\". File names that don't contain NEWLINE aren't translated. (split_3): Rewrite to handle file names with embedded newlines. Miles Bader and Jim Blandy suggested this new encoding scheme. * src/md5sum.c (md5_file): Replace obsolete comment with a description of the function. (md5_check): Don't use "s"-adding trick to form the plural of `checksum.' That doesn't work well with translation. Suggestions from Ulrich Drepper. (split_3): Add missing semicolon so it compiles. From Jim Blandy. Wed Jul 3 23:21:09 1996 Jim Meyering * src/md5sum.c (split_3): Correct test for 35-byte line to accomodate fact that leading blanks may be stripped. Tue Jul 2 21:51:40 1996 Jim Meyering * configure.in (ALL_LINGUAS): Add dutch (nl). Mon Jul 1 23:50:19 1996 Jim Meyering * src/md5sum.c [NEWLINE_REPLACEMENT_STRING]: Define. (split_3): Translate NL bytes not to NUL, but to NEWLINE_REPLACEMENT_STRING. Suggested by Ulrich Drepper. (main): Translate back to NL-containing filename. Sun Jun 30 22:42:17 1996 Jim Meyering * src/md5sum.c (split_3): Take an additional parameter, S_LEN. Adapt caller. Map translated NEWLINE-containing filename back into the original NEWLINE-containing name. (md5_check): Translate NEWLINE bytes to NUL bytes in filename. Sat Jun 29 18:59:07 1996 Jim Meyering * configure.in (PACKAGE_VERSION): Add `GNU ' prefix so it appears in the output of --version. Reported by RMS. Wed Jun 26 21:35:10 1996 Jim Meyering * src/Makefile.am (LDADD): Remove sort-specific hack to link with -lm. * aclocal.m4 (AM_FUNC_STRTOD): New macro. * configure.in (AM_FUNC_STRTOD): Use it. (AC_REPLACE_FUNCS): Remove strtod. Suggested by Tom Tromey. * po/Makefile.in.in (install-data): Add `else true;' to avoid letting failing if-condition cause make failure. From Fred Fish (fnf@ninemoons.com). Tue Jun 25 21:55:18 1996 Jim Meyering * Version 1.18a. Tue Jun 18 20:43:20 1996 Jim Meyering * aclocal.m4: Update from gettext-0.10.20. * ABOUT-NLS: Likewise. * intl/*: Likewise. Thu Jun 13 22:31:39 1996 Jim Meyering * configure.in (VERSION): Bump to 1.18a. * po/Makefile.in.in ($(PACKAGE).pot): Reapply change of May 30. Depend on POTFILES only when using maintainer mode. * intl/*: Update from gettext-0.10.18. Tue Jun 11 23:32:10 1996 Jim Meyering * Version 1.18. * po/Makefile.in.in: Update from gettext-0.10.17. * intl/*: Likewise. * ABOUT-NLS: Likewise. Mon Jun 10 18:22:29 1996 Jim Meyering * Makefile.am (cvs-dist): Use -c option in cvs tag command. Fri Jun 7 22:06:46 1996 Jim Meyering * configure.in (VERSION): Set to 1.18. * src/Makefile.am (sort_LDADD): Set this to -lm to get pow when using the strtod supplied with this package. Otherwise, linking failed on Solaris-2.4 systems. Reported by Wayne Stewart . Thu Jun 6 21:57:08 1996 Jim Meyering * po/Makefile.in.in: Update from gettext-0.10.16. * intl/*: Likewise. * ABOUT-NLS: Likewise. * aclocal.m4 (AC_REPLACE_GNU_GETOPT): Move definition to precede gettext-related ones. Update from gettext-0.10.16. Fri May 31 22:04:51 1996 Jim Meyering * Version 1.17. * Makefile.am (EXTRA_DIST): Set to texinfo.tex so that file is distributed. Thu May 30 00:07:48 1996 Jim Meyering * Version 1.16. * tail.c (main): Interpret the old-style +VALUE and -VALUE options like -c VALUE and -c +VALUE (resp) when VALUE has one of the [bkm] suffix multipliers. This makes the code consistent with --help output. Reported by Karl Heuer. * sort.c (limfield): #ifdef-out a block of code that makes GNU sort incompatible with Unix sort when a key-end spec refers to the N'th character in a field that has fewer than N characters. The POSIX spec doesn't appear to specify behavior for this case. From Karl Heuer. * po/Makefile.in.in (stamp-cat-id): Avoid using temp filename longer than 14 characters. (mostlyclean): Likewise. From Karl Heuer. * po/Makefile.in.in ($(PACKAGE).pot): Depend on POTFILES only when using maintainer mode. Reported by Karl Heuer. * po/Makefile.in.in (.po.mo): Reenable dependency and rule. It is required when using native NLS, e.g. on Solaris. From Marcus Daniels. * od.c (decode_one_format): Use %lu (not %d) printf format corresponding to unsigned long, SIZE. * Version 1.15. * intl/Makefile.in (installcheck): New target. * po/Makefile.in.in (installcheck): New target. * od.c (decode_one_format): Take another parameter. Give a specific diagnostic for invalid format spec here rather than an overly general one from caller's caller. (decode_format_string): Update caller. (main): Don't give diagnostic here when decode_format_string fails. Remove assertions that could fail on some Crays because they don't have a two-byte type. Johan Danielsson (joda@pdc.kth.se) reported the failed assertions. Mon May 27 17:43:31 1996 Jim Meyering * cat.c (main): Use STDIN_FILENO and STDOUT_FILENO instead of less portable fileno (stdin) and fileno (stdout). * sort.c (main): Use STDIN_FILENO instead of less portable fileno (stdin). Mon May 20 21:50:23 1996 Jim Meyering (meyering@na-net.ornl.gov) * Makefile.am (distcheck): Remove target and rule -- automake-0.33 adds it automatically. Sun May 19 13:15:49 1996 Jim Meyering (meyering@na-net.ornl.gov) * configure.in (LINGUAS): Add ko. * Makefile.am (distcheck): Remove @MAINT@ prefix from first command. Thu May 16 22:18:41 1996 Jim Meyering (meyering@na-net.ornl.gov) * Makefile.am (distcheck): New target. From Gord Matzigkeit. Fri May 10 20:46:13 1996 Jim Meyering (meyering@na-net.ornl.gov) * po/Makefile.in.in (all-yes): Always depend on CATALOGS. Otherwise they won't be built and install could fail. From Ulrich Drepper. (.po.mo): Disable dependency and rule when not in maintainer mode. * aclocal.m4 (ud_WITH_NLS): Make DATADIRNAME and CATOBJEXT depend on whether we're using GNU gettext. From Ulrich Drepper. Wed May 8 21:10:43 1996 Jim Meyering (meyering@na-net.ornl.gov) * intl/Makefile.in (distdir): Renamed from dist. For compatibility with automake-0.32. * po/Makefile.in.in (distdir): Likewise. * po/Makefile.in.in (all-yes): Depend on GMOFILES. Depend on CATALOGS and GMOFILES files only in maintainer mode. Tue May 7 22:10:20 1996 Jim Meyering (meyering@na-net.ornl.gov) * configure.in: Require autoconf-2.10. * aclocal.m4 (jm_MAINTAINER_MODE): New macro. * configure.in (jm_MAINTAINER_MODE): Use it. Sat May 4 20:40:01 1996 Jim Meyering (meyering@na-net.ornl.gov) * configure.in (LINGUAS): Add de. * po/de.po: New file. * memcasecmp.c [IN_CTYPE_DOMAIN]: Rename from ISASCII. * regex.c [IN_CTYPE_DOMAIN]: Likewise. * system.h [IN_CTYPE_DOMAIN]: Likewise. Sun Apr 28 17:10:03 1996 Jim Meyering (meyering@na-net.ornl.gov) * fmt.c (copy_rest): Don't copy past end of prefix. Output trailing spaces if the prefix had any. Before, fmt would improperly output NUL bytes. The command perl -e 'print "| S";' |fmt -p' | '|tr '\0' @ output `|@S'. Reported by François Pinard. * sort.c (main): Give a better diagnostic for `sort -0'. Reported by Karl Berry. * configure.in: Make fp_C_PROTOTYPES precede AC_C_INLINE. Otherwise, some systems lose because the value AC_C_INLINE choses with plain `cc' is different from that chosen when using the ANSI-mode C compiler. From Kaveh Ghazi. Tue Apr 23 22:05:35 1996 Jim Meyering (meyering@na-net.ornl.gov) * aclocal.m4 (AC_REPLACE_GNU_GETOPT): New macro. * configure.in: Use it. * lib/Makefile.am (tu_SOURCES): Remove getopt.c and getopt1.c. (EXTRA_DIST): Add getopt.c and getopt1.c * src/*.c, src/*.h: Update Copyright years to include 1996. Sun Apr 21 08:04:51 1996 Jim Meyering (meyering@na-net.ornl.gov) * getline.c: New version from gettext-0.10.12. * getline.h: Likewise. * configure.in (AC_CHECK_FUNCS): Add getdelim. Suggested by Ulrich Drepper. * md5sum.c (md5_check): Remove spurious newline at end of error format string. * od.c (decode_one_format): Use printf's L modifier for long doubles, not `l'. From Eric Backus. * Makefile.am (LDADD): Put @INTLLIBS@ before package library. * od.c: Include . [!BITSPERBYTE]: Define. [OFF_T_MAX]: Use BITSPERBYTE, not 8, and cast the result to off_t. * configure.in (AC_CHECK_HEADERS): Add values.h. Fri Apr 19 23:48:53 1996 Jim Meyering (meyering@na-net.ornl.gov) * Makefile.am (cvs-dist): New rule. Based on the one from Tom Tromey's automake. Thu Apr 18 22:13:14 1996 Jim Meyering (meyering@na-net.ornl.gov) * tail.c (main): Make code clearer: use new variable `n_files' in place of `argc - optind'. Use `file' instead of `argv + optind'. Sat Apr 13 13:08:29 1996 Jim Meyering (meyering@na-net.ornl.gov) * lib/Makefile.am (noinst_HEADERS): Include memcasecmp.h, not memcasecmp.c. * memcasecmp.c: Unprotoize. [ISASCII]: Define. [ISUPPER]: Define. * join.c (decode_field_spec): Rename local variable `valid' to invalid (and reverse sense) to avoid conflict with the macro definition in /usr/include/locale.h on SunOS 4.1.3. Tue Apr 9 22:43:57 1996 Jim Meyering (meyering@na-net.ornl.gov) * uniq.c: Add new option --ignore-case (-i). Include memcasecmp.h. (different): Compare with memcasecmp if ignoring case. (main): Handle 'i'. * join.c (memcasecmp): Remove static definition of function. Include memcasecmp.h instead. * memcasecmp.c: New file. * memcasecmp.h: New file. * lib/Makefile.am (tu_SOURCES): Add memcasecmp.c. (noinst_HEADERS): Add memcasecmp.h. Thu Apr 4 17:05:33 1996 Jim Meyering (meyering@na-net.ornl.gov) * tr.c: The following commands all failed unnecessarily. tr -cs '[:upper:]' '[X*]' tr -dcs '[:lower:]' n-rs-z tr -ds '[:xdigit:]' [:alnum:]' tr -dcs '[:alnum:]' [:digit:]' tr -dc '[:upper:]' Reported by Ross Ridge (ross@worf.mks.com). (validate): Add missing conjunct (translating) in test for `translating and complementing character classes' error. Before, valid uses of tr could fail. E.g. `tr -dcs '[:cntrl:]' '[:alnum:]''. (homogeneous_spec_list): New function. (validate): Use it to relax the old (overly restrictive) restriction that prohibited use of complemented character classes when translating. Now, that is allowed as long as the translation maps many to one. (get_spec_stats): Rename and redefine global has_char_class from has_upper_or_lower. Wed Apr 3 07:08:57 1996 Jim Meyering (meyering@na-net.ornl.gov) * aclocal.m4: Update NLS macros from gettext-0.10.11. * src/Makefile.am (INCLUDES): Add -I$(srcdir) to get system.h when srcdir != builddir. From Kaveh Ghazi. Wed Mar 27 23:33:36 1996 Jim Meyering (meyering@na-net.ornl.gov) * cksum.c (cksum): Use format "%lu %ld" (instead of "%10lu %8ld") to be POSIX compliant. From Stephen Gildea . Sun Mar 24 08:47:40 1996 Jim Meyering (meyering@na-net.ornl.gov) * join.c (main): Use EXIT_FAILURE, not 2, as exit code in error call. * od.c (skip, main): Likewise. * tr.c (main): Likewise. * cat.c cksum.c comm.c csplit.c cut.c expand.c fmt.c fold.c head.c join.c md5sum.c nl.c od.c paste.c pr.c split.c sum.c tac.c tail.c tr.c unexpand.c uniq.c wc.c: Exit with status EXIT_SUCCESS or EXIT_FAILURE, rather than 0 or 1. Sat Mar 23 23:14:40 1996 Jim Meyering (meyering@na-net.ornl.gov) * system.h [EXIT_FAILURE]: Undefine and define to 1 also if it is defined to zero. Suggested by Paul Eggert. * sort.c [SORT_FAILURE]: New macro. Be careful to exit with 1 only when -c is used and the input is not properly sorted. In all other cases, use SORT_FAILURE as required by POSIX. (main): Change some `error (1, ...' to use SORT_FAILURE. Upon successful termination, exit with EXIT_SUCCESS instead of `0'. Replace all uses of `2' (as exit code) with SORT_FAILURE. Thu Mar 21 22:47:50 1996 Jim Meyering (meyering@na-net.ornl.gov) * csplit.c: Include assert.h, but disable assertions. (process_line_count): Replace if-abort with a slightly relaxed assertion. Before, `echo |csplit - 1 1' would abort. Reported by Samuli.Karkkainen@hut.fi. (parse_patterns): Disallow uses like `csplit FILE 0' with zero line number, `csplit FILE 2 1' with decreasing line numbers, and warn about uses like `csplit FILE 3 3' that have equal line numbers. * sort.c (main): Declare to be of type int, not void. From Peter Seebach . * cat.c cksum.c comm.c csplit.c cut.c expand.c fold.c head.c join.c nl.c paste.c pr.c split.c sum.c tac.c tail.c tr.c unexpand.c uniq.c wc.c (main): Likewise. Sat Mar 16 16:30:07 1996 Jim Meyering (meyering@na-net.ornl.gov) * configure.in (PACKAGE_VERSION): New macro. * acconfig.h (PACKAGE_VERSION): Add it. * src/Makefile.am: Remove rules for generating version.c. Remove references to version.o, version.h, and stamp-v. * version.c: Remove file. * version.h: Remove file. * cat.c cksum.c comm.c csplit.c cut.c expand.c fmt.c fold.c head.c join.c md5sum.c nl.c od.c paste.c pr.c sort.c split.c sum.c tac.c tail.c tr.c unexpand.c uniq.c wc.c: Don't include version.h. (main): Use PACKAGE_VERSION instead of version_string. Mon Mar 11 16:07:08 CST 1996 Jim Meyering (meyering@na-net.ornl.gov) * sort.c (write_bytes): Renamed from xfwrite. Remove SIZE parameter. It was always 1. Update all callers. Rename NELEM parameter to N_BYTES and change its type to size_t. Sat Mar 9 13:30:12 1996 Jim Meyering (meyering@na-net.ornl.gov) * cat.c cksum.c comm.c csplit.c cut.c expand.c fmt.c fold.c head.c join.c md5sum.c nl.c od.c paste.c pr.c sort.c split.c sum.c tac.c tail.c tr.c unexpand.c uniq.c wc.c (main): Initialize for internationalized message support: call setlocale, bindtextdomain, and textdomain. * system.h: Add definitions and includes for NLS. * Makefile.am (SUBDIRS): Add intl and po. * src/Makefile.am (datadir): Define. (localedir): Define. (DEFS): Add LOCALEDIR definition. (LDADD): Add @INTLLIBS@. * aclocal.m4: Add NLS-related macro definitions from gettext distribution. * acconfig.h (ENABLE_NLS, HAVE_CATGETS, HAVE_GETTEXT, HAVE_LC_MESSAGES, HAVE_STPCPY): New macros. Add PACKAGE and VERSION. * configure.in (PACKAGE): Define with AC_DEFINE_UNQUOTED. (VERSION): Likewise. (ALL_LINGUAS): Define. (AC_CHECK_HEADERS): Add float.h. (ud_GNU_GETTEXT): Use it. Link an nls file. (AC_OUTPUT): Reflect addition of two new directories, intl and po. Create po/Makefile. From François Pinard. * tr.c (star_digits_closebracket): Declare formal param, IDX, and local I to be of type size_t (rather than int) to avoid warnings from gcc -Wall. Wed Feb 28 20:32:48 1996 Jim Meyering (meyering@na-net.ornl.gov) * src/Makefile.am: Adapt for automake-0.30. (noinst_HEADERS): Rename from HEADERS. (INCLUDES): Remove -I.. and -I$(srcdir). automake-0.30 adds these to DEFS automatically. * lib/Makefile.am (noinst_HEADERS): Rename from HEADERS. Wed Feb 21 18:56:27 1996 Jim Meyering (meyering@na-net.ornl.gov) * join.c (xfields) [without -t]: Ignore leading blanks. From David Dyck (dcd@tc.fluke.COM). Sun Feb 18 12:07:27 1996 Jim Meyering (meyering@na-net.ornl.gov) * join.c [TOLOWER]: Define. (usage): Briefly describe new --ignore-case (-i) option. (memcasecmp): New function. (main): Set IGNORE_CASE for -i. (keycmp): Compare with memcasecmp if IGNORE_CASE is set. Suggestion and an initial patch from Alberto Accomazzi . Sat Feb 17 18:54:40 1996 Jim Meyering (meyering@na-net.ornl.gov) * sort.c: Add -z option. Change all occurrences of '\n' to `eolchar'. (usage): Describe it. (main): Set eolchar to 0 for -z. From Mark W. Eichin . Fri Feb 16 21:44:14 1996 Jim Meyering (meyering@na-net.ornl.gov) * join.c (main): Mark two strings for translation. * md5sum.c (md5_check): Mark a string for translation. From François Pinard. Mon Feb 12 21:12:28 1996 Jim Meyering (meyering@na-net.ornl.gov) Update for automake-0.29. * Makefile.am (CONFIG_HEADER): Don't define. Automake now does it. * src/Makefile.am: Likewise. * lib/Makefile.am: Likewise. * doc/Makefile.am (info_TEXINFOS): Renamed from TEXINFOS. * man/Makefile.am (man_MANS): Renamed from MANS. * fmt.c [TRUE]: Undefine before defining to avoid warning on NeXT. [FALSE]: Likewise. From Derek Clegg (dclegg@next.com). Sat Feb 3 16:20:40 1996 Jim Meyering (meyering@na-net.ornl.gov) * Version 1.14. Wed Jan 31 23:26:13 1996 Jim Meyering (meyering@na-net.ornl.gov) * tail.c (dump_remainder): Flush stdout when tailing-forever on multiple files. From Carlos Canau . Fri Jan 5 18:30:28 1996 Jim Meyering (meyering@na-net.ornl.gov) * aclocal.m4: For SVR4 systems, use -Xc -D_POSIX_C_SOURCE in CFLAGS, not just -Xc. From François Pinard. * doc/Makefile.am (DIST_OTHER): Don't set to mdate-sh. mdate-sh is distributed automatically. From Tom Tromey. Thu Jan 4 20:32:52 1996 Jim Meyering (meyering@na-net.ornl.gov) * nl.c: Include rx.h after limit.h (not before) to avoid warning about the redeclaration of RE_DUP_MAX in limit.h. rx.h protects its definition. From Kaveh Ghazi. * src/Makefile.am (DIST_OTHER): Include md5-test.rfc so make check works again. Reported by François Pinard and Kaveh Ghazi. * rx.c [SYNTAX]: Rename from SYNTAX_TABLE. [SYNTAX_TABLE]: Remove extern dcl of re_syntax_table. (RE_SEARCH_2_FN): Cast string1 and string2 to avoid pointer type mismatch errors from e.g. SunOS's /bin/cc. Reported by Kaveh Ghazi. Wed Jan 3 23:34:39 1996 Jim Meyering (meyering@na-net.ornl.gov) * configure.in (VERSION): Set variable directly rather than from a separate file. * Makefile.am (DIST_OTHER): Remove VERSION. * src/Makefile.am (version.c): Depend on Makefile, not ../VERSION. (stamp-v): New target; depend on Makefile. (DISTCLEANFILES): Add stamp-v, version.c. Tue Jan 2 00:08:08 1996 Jim Meyering (meyering@na-net.ornl.gov) * sort.c (keycompare): Rewrite cases handling ignored characters so that they really are ignored. Reported by Carl Johnson . Fri Dec 22 21:34:55 1995 Andreas Schwab * configure.in: Touch stamp-h only if config.h is remade. * csplit.c (extract_regexp) [!WITH_REGEX]: Intialize syntax_parens field. Mon Dec 18 21:42:44 1995 Jim Meyering (meyering@comco.com) Use automake to generate Makefile.in files. * Makefile.am: New file. * doc/Makefile.am: New file. * lib/Makefile.am: New file. * man/Makefile.am: New file. * src/Makefile.am: New file. Thu Dec 14 21:13:22 1995 Jim Meyering (meyering@comco.com) * fmt.c (main): Use xstrtol instead of atoi. * fold.c (main): Use xstrtol instead of atoi. (fold_file): Make a couple locals `int', not size_t, and cast fwrite's third parameter to size_t. * nl.c (main): Use xstrtol instead of atoi. Rename misleading --first-page=N option to --starting-line-number=N. (usage): Update. * uniq.c (main): Use xstrtol instead of atoi. Fri Dec 8 20:48:00 1995 Jim Meyering (meyering@comco.com) * aclocal.m4 (fp_C_PROTOTYPES): Set ANSI2KNR to ./ansi2knr. Thu Dec 7 19:55:03 1995 Jim Meyering (meyering@comco.com) * aclocal.m4 (ud_WITH_REGEX): New macro. * configure.in: Use it. * rx.c, rx.h: New files. * lib/Makefile.in (REGEX): Set to @REGEXOBJ@, not regex.o. (SOURCES, DISTFILES): Add rx.c, rx.h. * acconfig.h (WITH_REGEX): Add it. * csplit.c [!WITH_REGEX]: Include rx.h. * nl.c [!WITH_REGEX]: Likewise. * tac.c [!WITH_REGEX]: Likewise. From Ulrich Drepper and François Pinard. Tue Dec 5 17:49:53 1995 Jim Meyering (meyering@comco.com) * sort.c (main): Diagnose invalid arguments to -k, then fail. * sort.c (numcompare): Properly handle all cases in which signs differ. From Karl Heuer. * Makefile.in (mandir): @mandir@ is just $(prefix)/man. Append `/man1'. Reported by Karl Heuer. * sum.c (bsd_sum_file): Remove misleading comment. From Karl Heuer. * src/Makefile.in [.c._o]: Don't create $<-tmp. From Kaveh Ghazi. Fri Dec 1 01:02:36 1995 Jim Meyering (meyering@comco.com) * lib/Makefile.in (default): New target. * sort.c (numcompare): Properly order `-' and valid integers; treat `-' as if it were zero. Reported by Ross Alexander . * Makefile.in (DISTFILES): Remove unused mkdep-Makefile. (.PHONY): Depend on default. Wed Nov 29 21:50:53 1995 Jim Meyering (meyering@comco.com) * aclocal.m4 (jm_PROG_C_MKDEP): Remove, for now. (jm_AUTODEPS): Likewise. (jm_WITH_GNU_MAKE): Likewise. * configure.in (jm_AUTODEPS): Remove. * sort.c (begfield): Replace loop with equivalent if-else. (limfield): Fix bug that made field specs like -k1.5 return a pointer into LINE beyond the first field. Don't try to trim trailing blanks here; it's not possible because we don't know where the field begins. Replace loop with equivalent if-else. (trim_trailing_blanks): New function. (findlines): Call it. (keycompare): Likewise. (main): Don't decrement T2 corresponding to `x' in field specs like -ku.v,w.x. Sun Nov 26 15:50:23 1995 Jim Meyering (meyering@comco.com) * csplit.c [!UINT_MAX]: Define. [!INT_MAX]: Define. Include xstrtoul.h. (string_to_number): Remove function. (check_for_offset): Use xstrtoul instead of string_to_number. (parse_repeat_count): Likewise. (parse_patterns): Likewise. (main): Likewise. Wed Nov 22 23:12:47 1995 Jim Meyering (meyering@comco.com) * Makefile.in (all et. al.): If make was invoked with -k and a sub-make fails, fail after the loop rather than exiting right away. Otherwise, make's -k option could be ineffective. Sat Nov 18 12:04:30 1995 Jim Meyering (meyering@comco.com) * src/Makefile.in (version.c): Depend on ../VERSION, not Makefile. Update for autoconf-2.5: * src/Makefile.in (libdir): Set to @libdir@. (bindir): Set to @bindir@. * doc/Makefile.in (infodir): Set to @infodir@. * man/Makefile.in (exec_prefix): Remove it. (mandir): Set to @mandir@. (bindir): Remove it. Thu Nov 16 21:25:45 1995 Jim Meyering (meyering@comco.com) * Makefile.in (default): New default target. Depend on `all' to work around bug in AIX-3.2.5's /bin/make. Reported by Andreas Luik . Wed Nov 15 23:48:31 1995 Jim Meyering (meyering@comco.com) * mkdep-Makefile (.deps/empty.P): Depend on BUILT-HEADERS. * aclocal.m4 (jm_PROG_C_MKDEP): Rename from jm_PROG_MKDEP. * src/Makefile.in (C_MKDEP): Rename from MKDEP. * lib/Makefile.in (C_MKDEP): Likewise. * mkdep-Makefile (C_MKDEP): Likewise. Add comments. * csplit.c [__P]: Don't define it here. * pr.c [__P]: Don't define it here. * fmt.c [__P]: Don't define it here. * system.h [__P]: Define it here. * aclocal.m4 (jm_AUTODEPS): Rename from jm_WITH_AUTODEPS. Revamp, with suggestions from Franc,ois Pinard. (jm_WITH_GNU_MAKE): New macro. (jm_PROG_MKDEP): New macro. * configure.in (jm_PROG_MKDEP): Use it. * src/Makefile.in (MKDEP): Set it. * lib/Makefile.in (MKDEP): Likewise. * mkdep-Makefile: Use MKDEP, not DEP_CC. (DEP_CFLAGS): Remove it. Tue Nov 14 23:10:54 1995 Jim Meyering (meyering@comco.com) * sort.c: Add support for sorting numbers in scientific notation. Include xstrtod.h. (struct keyfield): Add field: general_numeric. (usage): Describe -g option. (general_numcompare): New function. (keycompare): Use new comparison function when general_numeric flag is set. (set_ordering): Honor `g' flag. (main): Initialize and use new field. From Marcus Daniels . * configure.in (AC_REPLACE_FUNCS): Add strtod. * lib/Makefile.in (SOURCES): Add xstrtod.c. (OBJECTS): Add xstrtod.o. (DISTFILES): Add xstrtod.h. * csplit.c (cleanup): Don't exit. (interrupt_handler): Rewrite to reset default signal handler then repost caught signal. (xmalloc, xrealloc, read_input, write_to_file, handle_line_error, process_line_count, process_regexp, create_output_file, close_output_file): Call cleanup_fatal, not cleanup. Add const attribute to lots of parameters. (cleanup_fatal): Renamed from cleanup. Call cleanup, then exit. Update callers. * tac.c (cleanup): Don't exit. (cleanup_fatal): New function. Call cleanup, then exit. (sighandler): New function. (cleanup_fatal): Renamed from cleanup. Call cleanup, then exit. Update callers. (tac_stdin): Set up sighandler as signal handler, not `cleanup'. Sat Nov 11 15:46:02 1995 Jim Meyering (meyering@comco.com) * od.c (skip): Cast fseek's offset argument to `long'. Erik Bennett reported that this is necessary on BSDI systems. And if offset doesn't fit in a long, then try using lseek instead. * sort.c (xmalloc): Declare static. (xrealloc): Likewise. * tac.c (cleanup): Add signal number parameter. Update callers. * cat.c (main): Don't die when dev/ino of input and output are the same and the file descriptors correspond to stdin and stdout. This is necessary on at least Digital UNIX (aka OSF1) 3.2C. Otherwise, running `cat << X > $$' would fail with `cat: -: input file is output file'. From Jarkko Hietaniemi . * Makefile.in (DISTFILES): Add @README_ALPHA@. (config.status) Depend on VERSION. * configure.in: Set README_ALPHA if the version number ends in [a-z]. Fri Nov 10 21:13:42 1995 Jim Meyering (meyering@comco.com) * fold.c (usage): Clarify meaning of --spaces. From Karl Berry. src/Makefile.in (.c._c): Don't redirect directly to $@. (._c._o, .c._o): Remove temporary src file. Tue Nov 7 23:55:24 1995 Jim Meyering (meyering@comco.com) * VERSION: New file. * configure.in (VERSION): New variable. (PACKAGE): New variable. * Makefile.in (DISTFILES): Add VERSION. (distdir): Use @PACKAGE@-@VERSION@, not distname. Remove all references to distname. * src/Makefile.in (version.c): Generate it. (SOURCES): Remove now-generated version.c. (distclean): Delete version.c. * doc/Makefile.in (version.texi): Use @PACKAGE@-@VERSION@, not version.c * */Makefile.in (distdir): Use @PACKAGE@-@VERSION@, not distname. Remove all references to distname. * split.c: New option: --verbose. (main): Recognize it. (usage): Document it. From Marty Leisner . Sun Nov 5 06:59:27 1995 Jim Meyering (meyering@comco.com) * src/Makefile.in: Reflect protoization of csplit.c, fmt.c, pr.c, and unexpand.c. * pr.c: Protoize. (store_columns): Remove spurious 2nd argument in call to read_line. [__P]: Define. Use it in prototyped forward decls. * csplit.c: Move prototypes after struct declarations. Remove prototypes for xmalloc, xrealloc. Protect prototypes with __P. (new_control_record): Cast arg 1 to realloc. * fmt.c [__P]: Test PROTOTYPES, not __STDC__. Protoize. Sat Nov 4 23:51:30 1995 Jim Meyering (meyering@comco.com) * csplit.c (close_output_file): Set output_stream to NULL so we don't recurse endlessly between this function and cleanup. * unexpand.c: Reorder functions to obviate forward dcls. Remove forward dcls. Protoize. Add `const' attribute to some parameters. Mon Oct 30 23:15:56 1995 Jim Meyering (meyering@comco.com) * wc.c: Reorder functions to obviate forward dcls. Remove forward dcls. Protoize. Add `const' attribute to some parameters. * uniq.c: Likewise. * split.c: Likewise. * split.c (isdigits, convint): Remove these. (main): Use xstrtol instead. Sun Oct 29 14:04:36 1995 Jim Meyering (meyering@comco.com) * paste.c (main): Move function body to end of file. Remove forward dcls and protoize. * sum.c (main): Likewise. * nl.c: Protoize. Reorder functions to obviate forward dcls. Remove forward dcls. * tac.c: Likewise. * src/Makefile.in (OBJECTS): Reflect that all of the above now use prototypes. * sum.c (bsd_sum_file): Give file name parameter const attribute. (sysv_sum_file): Likewise. * tac.c (xwrite): Give output buffer parameter const attribute. (output): Likewise for both parameters. (tac): Likewise for file name parameter. (tac_file): Likewise. Sat Oct 28 16:02:39 1995 Jim Meyering (meyering@comco.com) * md5sum.c (md5_check): Use the same message format when there is a single file and it gets a read error or checksum mismatch as when there are more. Write that warning to standard error, not standard output. Suggestions from Greg Troxel (gdt@b-sgi.bbn.com). Thu Oct 26 00:11:35 1995 Jim Meyering (meyering@comco.com) * sort.c (xtmpfopen): New function to set proper permissions on temporary files. Use it instead of xfopen. Reported by Erik Corry (erik@kroete2.freinet.de). Mon Oct 23 23:17:04 1995 Jim Meyering (meyering@comco.com) * aclocal.m4 (jm_WITH_AUTODEPS): New directive. * configure.in: Use it. * lib/Makefile.in (INCLUDE, COMPILE): New variables. (.c.o): Rewrite to be more like src/Makefile.in. * src/Makefile.in (distclean): Remove .deps. Include @top_srcdir@/mkdep-Makefile. * mkdep.Makefile: New file. * Makefile.in (DISTFILES): Add mkdep-Makefile. * doc/Makefile.in (mostlyclean): Also remove *.info. Thu Oct 19 17:38:28 1995 Jim Meyering (meyering@comco.com) * od.c: Indent cpp directives. * tr.c (BEGIN_STATE): Use INT_MAX - 1 so as not to interfere with large repeat counts. (struct E_string): New struct. (find_closing_delim): Take E_string arg rather than char* and length. (find_bracketed_repeat): Likewise. (star_digits_closebracket): New function. (append_char_class): No longer give diagnostic. (append_equiv_class): Likewise. (build_spec_list): Give them here instead, allowing string1 like [:*][:upper:] that got errors before. Take E_string arg rather than char*. Convert switch stmts into if-then so as to use ES_MATCH. Tue Oct 17 22:48:00 1995 Jim Meyering (meyering@comco.com) * cat.c: Protoize. Reorder functions to obviate forward dcls. Remove forward dcls. * expand.c: Likewise. * fold.c: Likewise. * head.c: Likewise. * od.c: Likewise. * sort.c: Likewise. * tail.c: Likewise. * src/Makefile.in (OBJECTS): Reflect that all of the above now use prototypes. * head.c: Add `const' attribute to dcls of lots of formals. * sort.c: Add `const' attribute to dcls of lots of formals. Move struct dcls to precede dcls of file-scope variables. (monthtab, keyhead): Separate variable dcls from type dcls. Fri Oct 13 20:38:39 1995 Jim Meyering (meyering@comco.com) * join.c (add_field_list): Allow SPACE and TAB as well as `,' as list item separators, per POSIX spec. (prjoin): Rewrite loop that iterates over field specs. Handle case in which file spec is 0. Sun Oct 8 22:41:15 1995 Jim Meyering (meyering@comco.com) * strpbrk.c: New file. * lib/Makefile.in (SOURCES): Add strpbrk.c. Remove bcopy.c. Sat Oct 7 22:27:01 1995 Jim Meyering (meyering@comco.com) * join.c (add_field): No longer return a value. (decode_field_spec): New function. (add_field_list): Rewrite to be more strict. Before, `,1.2' was accepted as valid. (main): Use xstrtol instead atoi. Combine nearly identical -a and -v cases. Close input files. Fri Oct 6 23:10:01 1995 Jim Meyering (meyering@comco.com) * join.c (main): Report `too few/many non-option arguments' then print --help output rather than just the latter. Properly interpret obsolescent usage like `join -o 2.1 2.1 2.2'. (usage): Describe POSIX -1, -2 options and deprecate -j* ones * join.c (usage): Move to precede all other functions. * join.c: Protoize. Tue Oct 3 22:44:05 1995 Jim Meyering (meyering@comco.com) * csplit.c, sort.c, tac.c: Test SA_INTERRUPT, not _POSIX_VERSION, to determine whether `sigaction' functions are available. Reported by Paul Nevai . Fix suggested by Karl Berry. * md5sum.c (main): Declare counter, N_STRINGS, to be an integral type, not `char'. * cut.c: Convert many declarations using `int' to use `unsigned int' to avoid warnings from `gcc -Wall' about comparing signed and unsigned types. (set_fields): Use memset (not open coded loop) to initialize array. Thu Sep 28 23:16:05 1995 Jim Meyering (meyering@comco.com) * Version 1.13. * Makefile.in (DISTFILES): Remove README.alpha for major release. * od.c (decode_one_format): Remove spurious semicolon. From John Kodis. Tue Sep 26 23:05:01 1995 Jim Meyering (meyering@comco.com) * man/Makefile.in (install-data, uninstall): Use sed not basename. The GNU Coding Standard suggests that only a select set of relatively standard utilities be used in Makefiles. basename is not among them. Suggested by Ulrich Drepper. Sun Sep 24 08:36:47 1995 Jim Meyering (meyering@comco.com) * cksum.c: Protoize. * cut.c: Protoize. * src/Makefile.in (OBJECTS): Reflect that cksum.c and cut.c use prototypes. (cksum): Depend on cksum$O. (cut): Depend on cut$O. Sat Sep 23 15:43:46 1995 Jim Meyering (meyering@comco.com) * Version 1.12.2. * Makefile.in (DISTFILES): Add README.alpha. * lib/Makfile.in (GETOPT, MEMCHR, REGEX): New variables. (OBJECTS): Use them instead of hard-coding object file names. Suggested by Ulrich Drepper. * md5sum.c (md5_check): Distinguish between open/read failure and checksum mismatch. Mon Sep 18 23:15:05 1995 Jim Meyering (meyering@comco.com) * md5sum.c: Protoize. * src/Makefile.in (OBJECTS): Change md5sum.o to md5sum$O. (join.o, md5sum$O, sort.o) Depend on ../lib/long-options.h. (md5sum): Depend on md5sum$O. * md5sum.c (main, usage): Remove -h, -s, -v short options. Rename --verbose to --warn, --quiet to --status. (main): Handle --help and --version using parse_long_options. (md5_check): Check ferror. * sort.c (tempname): Replace `16' with a more readable expansion. Make sure that counter never exceeds 99999. (checkfp): Rearrange loop to avoid duplicate test. Move a couple dcls from function scope into inner block. Tue Aug 8 21:49:27 1995 Jim Meyering (meyering@comco.com) * md5sum.c (main): Fail if either --verbose or --quiet is specified when not verifying checksums. * md5sum.c (md5_check): Fail if no valid line is found. Don't use the word `fail' unless there were failures -- instead, say `all N tests passed.' * md5sum.c (main) [handling --string option]: Don't output nonstandard `b' binary flag. From Greg Troxel (gdt@bbn.com). * md5-test.rfc: Remove now-inconsistant `b' marker. (usage): Clarify help message. With suggestions from Greg Troxel. Mon Aug 7 23:27:54 1995 Jim Meyering (meyering@comco.com) * sort.c (mergefps) [lint]: Initialize SAVEALLOC to avoid spurious compiler warning. From Ulrich Drepper. * pr.c (read_line) [lint]: Initialize CHARS to avoid spurious compiler warning. From Ulrich Drepper. * Makefile.in (DISTFILES): Don't distribute unneeded COPYING.LIB. From Franc,ois. Mon Aug 7 00:02:59 1995 Jim Meyering (meyering@comco.com) * system.h (_): Define macro -- as empty for now. * src/*.c: Annotate localizable strings with _(...). From Franc,ois. Sat Aug 5 23:51:25 1995 Jim Meyering (meyering@comco.com) * Version 1.12.1. * configure.in: Add fp_C_PROTOTYPES. Fri Aug 4 23:21:46 1995 Jim Meyering (meyering@comco.com) * src/Makefile.in (DISTFILES): Include ansi2knr.c and ansi2knr.1. Add suffixes, dependencies, and rules to convert ANSI source to K&R when required. Mark tr.o and comm.o as the only such files (for now). * tr.c: Protoize. * comm.c: Protoize. * Makefile.in (DISTFILES): Add aclocal.m4 and acconfig.h. (configure): Depend on aclocal.m4. (stamp.h.in): Depend on aclocal.m4 and acconfig.h. Sun Jul 30 00:01:58 1995 Jim Meyering (meyering@comco.com) * md5sum.c (md5_check): Use getline instead of fgets. * lib/Makefile.in (SOURCES): Add getline.c. (OBJECTS): Add getline.o. (DISTFILES): Add getline.h. (getline.o): Depend on getline.h. * md5.c [memcpy]: Define to bcopy if needed. (md5_buffer): Explicitely cast SWAP expressions to unsigned to avoid warning. * lib/Makefile.in (strtoul.o): Depend on strtol.c. * sort.c (check): Print `disorder on...' message on standard error, not stdout. This is a minor concession to POSIX which says `no output shall be produced...'. * sort.c (main): Fix -k so it works. -k 2,3 was being treated like -k 2. Reported by Marcus Daniels . Wed Jul 26 22:54:48 1995 Jim Meyering (meyering@comco.com) * man/Makefile.in (install, uninstall): Strip off `.1', not just `1' before applying $(transform) to manual name. From Karl Berry. * md5sum.c (have_read_stdin): New global variable. (md5_file, md5_check): Set it. (main): Use it. * md5sum.c [OPENOPTS]: Depend explicitly on BINARY. (md5_file): Take a new parameter, MD5_RESULT, and no longer generate output. (md5_check): Invoke md5_file instead of calling fopen directly. When giving a diagnostic for a line with invalid format, also report the line number. (main): Generate output after a successful md5_file call. [in many places]: Upon detection of an error, rather than exiting immediately, issue a diagnostic, note that an error occurred and exit later. Mon Jul 24 21:48:26 1995 Jim Meyering (meyering@comco.com) * tail.c (tail_bytes) [from_start]: For regular files, seek relative to the initial input file pointer position, not necessarily from the beginning of the file. [!from_start]: Don't back up past the initial position of the input file pointer. (tail_lines): Call file_lines only if FD refers to a regular file with its file pointer positioned at beginning of file. Otherwise, call pipe_lines. This is a kludge. Once there's a decent test suite, fix this properly. Before, (echo 1; echo 2) > k; sh -c 'read x; tail' < k would output both lines of the input file even though the first had already been read. Reported by John Roll (john@panic.harvard.edu). * md5sum.c [_LIBC || STDC_HEADERS] (TOLOWER): Define to tolower. * csplit.c (main): When too few arguments are given, don't just issue the usage message; also report why. Sat Jul 22 00:16:01 1995 Jim Meyering (meyering@comco.com) * md5sum.c (md5_check): New function -- extracted from main. Fri Jul 21 01:21:49 1995 Jim Meyering (meyering@comco.com) * md5sum.c (md5_file): New function -- extracted from main. (main): Call the new function instead of doing all that in an if-stmt. Always use "%s" format rather than raw filename as format argument in printf-style functions like error. Otherwise, filenames containing `%' are likely to lose. Remove all mention of `old format'. (split_3): No longer parse the `new format.' There is only one valid format now: the compatible one. * lib/Makefile.in (SOURCES): Add md5.c. (OBJECTS): Add md5.o. (DISTFILES): Add md5.h. (md5.o): Depend on md5.h. * system.h [!EXIT_FAILURE, !EXIT_SUCCESS]: Define them. * lib/md5.c (md5_stream): Check for read failure and return indication of success rather than second argument. * lib/md5.h (md5_stream): Update prototype. [__P]: Define macro. * md5sum.c: Use EXIT_FAILURE and EXIT_SUCCESS. (main): Report an error if md5_stream fails. Otherwise, running `md5sum dir-on-mounted-filesystem' always reported the checksum for an empty file. Now it gets the `is a directory' error. Greg McGary (gkm@magilla.cichlid.com) reported that the released version got stuck in an infinite loop with such arguments. * lib/md5.c: New file. * lib/md5.h: New file. * md5sum.c: Rewritten to use library interface. From Ulrich Drepper. Sun Jul 16 18:42:34 1995 Jim Meyering (meyering@comco.com) * configure.in: Add AC_C_INLINE: md5sum.c uses inline keyword. Tue Jul 11 21:56:02 1995 Jim Meyering (meyering@comco.com) * md5sum.c Add #else block after #ifdef VMS. (main): Rename local variable. Detect and report failed fclose calls on stdout and checkfile_stream. * tr.c (unqote): Comment out unreachable break stmts. (append_range): Declare FIRST and LAST parameters to be unsigned int, not size_t. (find_closing_delim): Clean up interface, separating boolean success indicator and index. (find_bracketed_repeat): Likewise. (build_spec_list): Adapt to use new interfaces. Declare CLOSING_DELIM_IDX and CLOSING_BRACKET_IDX to be of type size_t, not int. (main): Declare NR and CHARS_READ to be of type long, not int. * sort.c (main): Comment out unreachable break stmt. * tail.c (tail): Explicitly cast to long the N_UNITS argumennt in call to tail_lines. Otherwise, NetBSD lost because N_UNITS (of type off_t, a 64-bit type on NetBSD) wasn't converted to the target type, long. Reported by Dan Hagerty (hag@gnu.ai.it.edu). * comm.c: Reorder functions so main is last, so no forward dcls are needed. Sun Jul 9 11:41:15 1995 Jim Meyering (meyering@comco.com) * sort.c (main): When using obsolescent +pos -pos syntax, allow omission of field spec only when character offset is specified. Otherwise, -. +. would be accepted. Sat Jul 8 13:48:06 1995 Jim Meyering (meyering@comco.com) * sort.c (main): Allow field spec, M, to be omitted in +M.N or -M.N position specification. For compatibility with existing practice (what vendors?), omitting M is now equivalent to specifying `0' for M, but only when using the +pos -pos syntax, not with the -k option. From Rick Sladkey (jrs@world.std.com). Thu Jul 6 23:44:20 1995 Jim Meyering (meyering@comco.com) * od.c (dump): Correct loop-termination criterion. Before, running `printf 1234| ./od --width=4 --read-bytes=4' printed output for 8 bytes -- the last four were garbage. This happened only when the dump limit, N, was specified (with --read-bytes=N) and N was a multiple of bytes_per_block (usually 16, but 4 in this example). From Andreas Schwab. Sun Jul 2 22:12:40 1995 Jim Meyering (meyering@comco.com) * lib/Makefile.in (check): Depend on libtu.a. * md5sum.c (split_3): Rewrite to parse Plumb/Lankester format as well. (main): Print each sum line in (de facto) standard Plumb/Lankester format. New option: --quiet. Check option no longer takes an argument. When checking, exit status reflects success. Sat Jun 24 16:18:01 1995 Jim Meyering (meyering@comco.com) * md5sum.c [__GNUC__]: Don't conditionalize use of "inline". Autoconf (via config.h) defines it away for compilers that don't grok it. * cat.c, comm.c, join.c, sort.c, uniq.c: Undefine min/max before defining to avoid redefinition warning on some systems. * csplit.c (read_input): Rename paramater MAX to MAX_N_BYTES. This avoids potential conflict with max macro from some system's header files when using bad pre-ANSI compilers. * uniq.c (usage): Fix typo in text of --help output. From Steve McConnel * md5sum.c [uint32]: When testing for 32-bit arithmetic type, also check `unsigned long', as a concession to systems with int's smaller than 32 bits. * configure.in (AC_REPLACE_FUNCS): Remove memcmp (again!) since AC_FUNC_MEMCMP already handles it. This avoids warnings from ranlib on nextstep systems. From Kaveh Ghazi. * csplit.c (interrupt_handler): Declare to have a single integer parameter. Otherwise, some compilers fail with a type mismatch error in sa_handler assignment. * fmt.c [word]: Redefine. Otherwise, systems (Unicos for one) with headers that typedef `word' get syntax errors because of the variable by the same name. * join.c [join]: Redefine to avoid conflict with prototype in some system header file. Also for Unicos. * md5sum.c (main): Remove spurious colon in getopt spec string. From Ken Pizzini . Tue Jun 20 06:59:16 1995 Jim Meyering (meyering@comco.com) * md5sum.c (md5_file): Initialize two elements of LEN portably, rather than with ANSI aggregate initialization. Reported by Edzer Pebesma . Sun Jun 18 09:15:57 1995 Jim Meyering (meyering@comco.com) * Version 1.12. * md5sum.c (process_buffer): Copy values from struct into individual local variables before performing computation on them, then copy results back into returned struct. Solely for better optimization by compilers that can't keep struct members in registers. From Ulrich Drepper. Thu Jun 15 22:09:14 1995 Jim Meyering (meyering@comco.com) * md5sum.c: Remove unnecessary uses of `defined' in #if* tests. Don't use #elif. Some older compilers don't grok it. (split_3): New function to parse out sum, flag, and filename when reading check file. (hex_digits): Remove length parameter since string parameter is now nul-terminated. (main): Don't allocate separate arrays for filename and sum when checking. Get pointers into line buffer with split_3 instead of using sscanf. (main): Treat `no file args' just like `-' when not checking or using strings. (hex_digits): Unprotoize. * md5sum.c (main): Don't use a separate loop to convert to lower case. Suggested by Ulrich Drepper. (CLOSE): New macro. (md5_file): Use it. From Ulrich Drepper. * md5sum.c: Don't include values.h. [__P]: Undefine it. * md5sum.c: [UNIX || unix]: Also test __UNIX__, __unix__, and _POSIX_VERSION. Reported by Arne H. Juul. Tue Jun 13 21:20:10 1995 Jim Meyering (meyering@comco.com) * tr.c [NDEBUG]: Define it to disable assertions. Some systems have a broken header. (is_char_class_member): Set a variable in each branch of switch stmt and return that value after the switch (rather than returning directly from every branch). (unquote): More int->size_t changes. (build_spec_list): More int->size_t changes. Be very careful about comparison now that variables are unsigned: Use i + 2 < len rather than i < len - 2. The latter didn't work for len < 2. Caught early thanks to a thorough regression test suite. Mon Jun 12 23:07:59 1995 Jim Meyering (meyering@comco.com) * sort.c (xfclose): Don't try to flush stdin, only stdout. Otherwise, at least Ultrix-4.3's fflush would return EOF. Reported by Jim Blandy (jimb@cyclic.com). * tr.c (non_neg_strtol): Don't compare signed and unsigned. [struct Spec_list] (indefinite_repeat_element): New member. Use size_t rather than int or unsigned long where reasonable. (get_spec_status): Make interface cleaner. (get_s1_spec_status): New function. (get_s2_spec_status): New function. (validate): Use new functions instead of get_spec_status. Sun Jun 11 00:39:50 1995 Jim Meyering (meyering@comco.com) * md5sum.c: New file. From Ulrich Drepper. * md5-test.rfc: New file. * Makefile.in (PROGS): Add md5sum. (check): Run a recursive make in each subdirectory. * man/Makefile.in (MANFILES): Add md5sum.1. * src/Makefile.in (SOURCES): Add md5sum.c. (OBJECTS): Add md5sum.o. (PROGS): Add md5sum. (check): Add basic checks for md5sum. (md5sum): Add dependency and link rule. (DISTFILES): Add md5-test.rfc. * tr.c (look_up_char_class, append_char_class, append_equiv_class): Use const attribute when appropriate. Sun May 28 14:48:58 1995 Jim Meyering (meyering@comco.com) * join.c: Overhauled to make -a1 and -a2 options work. Passed all tests in a fairly thorough test suite. Reported by Michael Hasselberg (mikelh@zonta.ping.de). Sat May 27 00:35:47 1995 Jim Meyering (meyering@comco.com) * tail.c (main): Remove dcl of and statements that set now-unused variable. * system.h [!STDC_HEADERS && HAVE_MEMORY_H]: Include memory.h. Without this, SunOS doesn't get type for memchr. Reported by Kaveh Ghazi. Thu May 25 00:06:50 1995 Jim Meyering (meyering@comco.com) * tail.c (tail_lines): Change one more `long' to `off_t'. Otherwise, tail didn't work on NetBSD. From Arne H. Juul. * csplit.c [struct control]: Change type of `repeat' to unsigned long to avoid conversion warning. Sun May 21 07:50:00 1995 Jim Meyering (meyering@comco.com) * Makefile.in (.PHONY): TAGS is not a phony target. From Franc,ois. * xstrtol.c (__strtol): Remove prototype altogether. Move inclusion of xstrtol.h back so that it follows system headers. * Version 1.11.2. Sat May 20 06:44:14 1995 Jim Meyering (meyering@comco.com) * join.c (ADD_FIELD): Declare in K&R style, not ANSI. From Kaveh Ghazi. * xstrtol.c: Define NDEBUG for systems with losing assert.h. (__strtol): Add __unsigned to prototype and use __P. Move inclusion of xstrtol.h to precede prototype. From Kaveh Ghazi. * memmove.c: Include config.h for definition of const. From Kaveh Ghazi. * All Makefile.in (install-exec, install-exec): New targets. From Karl Berry. * tr.c (look_up_char_class): Take a string length argument -- required because the argument is a string that is *not* NUL-terminated. Use strncmp, not strcmp and add string lengths comparison. (append_char_class): Update caller. Before this change, tr '[:upper:]' '[:lower:]' read one uninitialized byte. From Andreas Schwab. * split.c (next_file_name): Move dcls of file-scope variables into this function. Don't rely on arithmetic being two's complement. * paste.c [CLOSED, ENDLIST]: Don't cast constants to FILE pointers. Instead, declare two FILE structs and use their addresses. * csplit.c (record_line_starts): Remove set-but-not-used, file-scope variable. Sat May 13 08:57:20 1995 Jim Meyering (meyering@comco.com) * all source files (usage): Include one- or two-line synopsis in --help output. From Karl Berry. * lib/Makefile.in [.c.o]: Remove -I. since safe-l?stat.h are no longer used. Wed May 10 22:26:35 1995 Jim Meyering (meyering@comco.com) * Version 1.11.1 Sun May 7 13:35:49 1995 Jim Meyering (meyering@comco.com) * all Makefile.in (maintainer-clean): Renamed from realclean per GNU Standards. Thu May 4 23:15:58 1995 Jim Meyering (meyering@comco.com) * sort.c (main): Use stat, not safe_stat. The latter was unnecessary. Don't include safe-stat.h. * lib/Makefile.in (DISTFILES): Remove safe-xstat.hin. Remove all associated rules. safe-l?stat.h are no longer used. * sort.c Don't hard-code /tmp. [DEFAULT_TMPDIR]: Use this instead. * tac.c (save_stdin): Likewise. * sort.c (set_ordering): Properly parse arguments like `+x.yn'. Reported by John Salmon (johns@mullet.anu.edu.au). Patch from Mike Haertel. Wed Apr 26 23:48:13 1995 Jim Meyering (meyering@comco.com) * sort.c (tempname): AND-off high bits of pid so that its decimal string representation is no longer than five digits. From Hans Verkuil (hans@wyst.hobby.nl). Thu Apr 20 23:09:33 1995 Jim Meyering (meyering@comco.com) * expand.c (add_tabstop): Give correct size when reallocating tab_list buffer. From Geoff Odhner (geoff@franklin.com). Reproduce with expand --t=`perl -e "print join (',', (1..300));"`. Tue Apr 18 22:57:43 1995 Jim Meyering (meyering@comco.com) * configure.in (AC_OUTPUT): Use echo, not date, to avoid creating unnecessary conflicts for people using version control software like RCS and CVS. (AC_ARG_PROGRAM): Use it. Thu Apr 13 23:22:57 1995 Jim Meyering (meyering@comco.com) * head.c: Remove block of redundant comments describing options. (usage): Remove reference to -l option. It's not valid. * od.c (write_block): Declare index I as unsigned int, not int. (get_lcm): Likewise. (dump_strings): Likewise. (main): Likewise. (dump): Cast BYTES_PER_BLOCK to off_t in MIN expression. (dump_strings): Declare bufsize to be of type size_t, not int. * split.c (next_file_name): Don't reuse X as an index. Declare new variable I instead. Sat Mar 25 15:07:23 1995 Jim Meyering (meyering@comco.com) * csplit.c: Declare as `unsigned int' all variables that were just `unsigned'. (make_filename): Change type of arg to unsigned int. (split_file): Change type of indices to unsigned int. Thu Mar 16 22:05:05 1995 Jim Meyering (meyering@comco.com) * od.c (main): Detect and fail when argument to -N is too large to fit in an off_t. Sun Mar 12 12:02:39 1995 Jim Meyering (meyering@comco.com) * configure.in (AC_REPLACE_FUNCS): Add memcmp.c. Remove check for sizeof long. * linebuffer.h: Undefine __P before defining it. * long-options.h: Likewise. * xstrtol.h: Likewise. * memcpy.c: New file. * lib/Makefile.in (SOURCES): Add memcpy.c. Fri Mar 10 21:14:11 1995 Jim Meyering (meyering@comco.com) * src/*.c: Update Copyright dates. Sat Feb 11 07:31:08 1995 Jim Meyering (meyering@comco.com) * system.h: Remove index/rindex and bcmp/bcopy/bzero references. Separate errno declaration from STDC_HEADERS. * wc.c (wc): For efficiency, handle separately the cases in which words need not be counted. Suggested by Karl Heuer. * wc.c (wc): Don't overcount the number of bytes when reading from a regular file on stdin with file pointer not at BOF. From Karl Heuer . Before, the command `(dd ibs=99k skip=1; wc -c) < /etc/group' made wc wrongly output the size of the file. Now it outputs `0'. Sat Jan 28 07:07:23 1995 Jim Meyering (meyering@comco.com) * cat.c (cat): Use memmove rather than bcopy. * csplit.c (load_buffer): Use memcpy rather than bcopy. (extract_regexp): Use strrchr, not rindex. * fmt.c (isopen, isclose, isperiod): Use strchr, not index. (flush_paragraph): Use memmove instead of bcopy. * fold.c (fold_file): Use memmove instead of bcopy. * od.c (decode_format_string): Use memmove instead of bcopy. [EQUAL_BLOCKS]: Use memcmp, not bcmp. (parse_old_offset): Use strchr, not index. (dump): Use memset, not bzero. * sort.c (checkfp, mergefps): Use memcpy instead of bcopy. (fillbuf): Use memmove instead of bcopy. * split.c (main): Use memset, not bzero. (line_bytes_split): Use memmove instead of bcopy. * tac.c (tac): Use memmove instead of bcopy. (output): Use memcpy instead of bcopy. * tail.c (pipe_lines, pipe_bytes): Use memcpy instead of bcopy. * tr.c (substr): Use memcpy instead of bcopy. (card_of_complement, set_initialize): Use memset, not bzero. * tr.c (substr): Don't allocate a byte for trailing NUL in result since the result needn't be NUL-terminated. Don't NUL terminate it. * configure.in (AC_REPLACE_FUNCS): Check for memmove and memcpy, not bcopy. (AC_CHECK_FUNCS): Add strchr and strrchr. * lib/Makefile.in (SOURCES): Add memmove.c and memcpy.c. * memmove.c: New file. * memcpy.c: New file. Tue Jan 24 22:18:19 1995 Jim Meyering (meyering@comco.com) * fmt.c (copy_rest): Replace `||' with `&&'. Before, a paragraph not followed by a newline would be followed by data copied from uninitialized storage. Repeat by `printf abc|fmt|od -ac'. Reported by Franc,ois Pinard. * od.c: Rename macros with FP_ prefix. Use FLOAT_ instead. AIX-2.2.1 declares a typedef, FP_DOUBLE, that conflicted. Be careful to use off_t instead of explicit `unsigned long'. Otherwise, systems like 4.4BSD lose on calls to fseek that expect off_t (which is long long on that system). (parse_old_offset): Use xstrtoul (main): Likewise. * src/Makefile.in (od.o, tail.o): Depend on ../lib/strtol.h. Mon Dec 19 22:13:55 1994 Jim Meyering (meyering@comco.com) * tail.c (numerous functions): Give char* filename arguments the const attribute. * lib/Makefile.in (SOURCES): Remove xwrite.c. It was no longer used. Add xstrtol.c and xstrtoul.c. (OBJECTS): Likewise for .o files. (DISTFILES): Add error.h and xstrtol.h and xstrtoul.h. * lib/Makefile.in (DISTFILES): Remove safe-xstat.cin. (distclean): Remove references to safe-l?stat.c. Remove all related rules and dependencies. * src/Makefile.in: $(OBJECTS): Depend on ../lib/error.h. Thu Dec 15 23:14:04 1994 Jim Meyering (meyering@comco.com) * tail.c (tail_file): Fix call to error with %s in format, but no corresponding argument. * src/*.c: Include "error.h" instead of simply declaring `void error ();'. * od.c (main): Use %lu, not %d for unsigned long. (my_strtoul, uint_fatal_error): Remove functions. Use xstrtoul and STRTOL_FATAL_ERROR instead. Sun Dec 11 17:24:09 1994 Jim Meyering (meyering@comco.com) * sort.c (main): Fix interpretation of field offsets when specified via -k option. They were being interpreted as zero-indexed. POSIX says they are 1-based indices. From Rik Faith . (keycompare): Don't ignore characters at the end of words when otherwise they would compare equal. From Rik Faith. * tail.c (): Change the type of n_units/n_lines/n_bytes to off_t because n_bytes is used as the option to fseek and off_t may be long long (e.g. on BSD4.4). The only place where a 32-bit limit is imposed is in the string-to-integer converstion of xstrtol. (main): Use xstrtol instead of manual string to integer conversion. * csplit.c (get_format_width, get_format_prec): Use ISDIGIT instead of comparisons against '0' and '9'. * tr.c [!ULONGMAX]: Define it. [!LONGMAX]: Define in terms of ULONGMAX instead of as a 32-bit constant. * full-write.c: Declare (or not) errno based on definedness or errno, not STDC_HEADERS. * safe-read.c: Likewise. * xwrite.c: Likewise. Thu Nov 17 06:37:33 1994 Jim Meyering (meyering@comco.com) * tail.c (dump_remainder): Flush standard output just before sleeping so that `tail -f' will output partial lines sooner. This applies only when following the end of a single file. From Leonard N. Zubkoff . * tail.c (file_lines, pipe_lines, pipe_bytes, start_bytes, start_lines, dump_remainder): Use STDOUT_FILENO instead of `1' in XWRITE calls. * tac.c (output): Use STDOUT_FILENO instead of `1' in xwrite call. * system.h (STDIN_FILENO, STDOUT_FILENO, STDERR_FILENO): Define if not already defined. Wed Nov 16 07:01:38 1994 Jim Meyering (meyering@comco.com) * tr.c: Include instead of "getopt.h". * sort.c (tempname): Merge two very similar sprintf stmts into one. * tail.c: Use BUFSIZ instead of BUFSIZE. * cut.c: Declare DELIM to be `int' rather than unsigned char. * fmt.c: Use index, not strchr -- per GNU Standards. Reported by Kaveh Ghazi. Always include . (main): Include program name in --version output. Sun Nov 06 00:17:21 1994 Jim Meyering (meyering@comco.com) * Version 1.11. Sat Nov 05 15:14:44 1994 Jim Meyering (meyering@comco.com) * memcmp.c, srtoul.c: Use up-to-date versions. Reported by Franc,ois Pinard. * src/Makefile.in (DEFS): Remove -Dlint... again. * src/Makefile.in (PROGS): Add fmt! Reported by Andreas Stolcke (stolcke@ICSI.Berkeley.EDU). Fri Nov 04 17:26:16 1994 Jim Meyering (meyering@comco.com) * Version 1.10. Thu Nov 03 23:23:08 1994 Jim Meyering (meyering@comco.com) * linebuffer.h: Use __P instead of _P since the latter conflicts with a definition in some . From Kaveh Ghazi. Wed Nov 02 17:36:34 1994 Jim Meyering (meyering@comco.com) * tr.c: Add const attribute where appropriate. * tail.c (BUFSIZ): Use definition from . Tue Nov 01 23:54:09 1994 Jim Meyering (meyering@comco.com) * tail.c (XWRITE): Allow n_bytes == 0 but call fwrite only if n_bytes is positive. Turn off assertions. `tail /dev/null' was getting a failed assertion. Reported by Doug McLaren (dougmc@comco.com). * cut.c (cut_bytes): Print at least a newline for every line of input. (cut_fields): Print a newline even for lines whose only selected field is empty. But print nothing when using -s without -f1. And print nothing for empty input. Reported by Richard_Sharman@software.mitel.com. Tue Nov 01 06:05:23 1994 Jim Meyering (meyering@comco.com) * fmt.c: Change #ifdef __STDC__ to #if defined (__STDC__) && __STDC__. * linebuffer.h: Likewise. Mon Oct 31 06:53:32 1994 Jim Meyering (meyering@comco.com) * configure.in (AC_REPLACE_FUNCS): Remove memcmp since AC_FUNC_MEMCMP already handles it. From Kaveh Ghazi. * {lib,src}/Makefile.in (DEFS): Don't use -Dlint. Some systems get failures when it's defined. From Kaveh Ghazi. * cat.c (cat): Conditionalize test for errno == ENOSYS. It's not defined on some Next and Alliant systems. From Kaveh Ghazi. * sort.c join.c (main): Include version.h and call parse_long_options with the proper number of arguments. Reported by Franc,ois Pinard. Sun Oct 30 01:30:41 1994 Jim Meyering (meyering@comco.com) * (configure.in): Add AC_OFF_T and rearrange by type along the same lines as autoscan. * fmt.c, fmt.1: New files. Both from Ross Paterson (rap@doc.ic.ac.uk). * Makefile.in (PROGS): Add fmt. * src/Makefile.in (SOURCES, OBJECTS): Add fmt.c and fmt.o respectively. (fmt): Add a link rule. * man/Makefile.in (MANFILES): Add fmt.1. Sat Oct 08 10:44:30 1994 Jim Meyering (meyering@comco.com) * Makefile.in (All of them): Update from the ones in sh-utils. From now on, rules and definitions that are comon to the fileutils, textutils, and sh-utils will get ChangeLog entries only in the sh-utils. Tue Oct 04 20:42:46 1994 Jim Meyering (meyering@comco.com) * od.c (my_strtoul): Set errno to zero before calling strtoul. Otherwise, od can fail complaining about a valid integer argument. From Andreas Schwab. Sun Oct 02 17:57:09 1994 Jim Meyering (meyering@comco.com) * tr.c (main): Give an error and fail when squeezing repeats and no non-options are given. Reported by Tony Robinson (ajr@eng.cam.ac.uk). * sort.c (main): Temporarily copy each input file that might be another name for the output file. When in doubt (e.g. a pipe), copy. This allows the dangerous (historical -- in cnews) idiom `cat file | sort -o file' to work properly. Don't rely on this behavior in new shell scripts. From Paul Eggert (eggert@twinsun.com). Mon Sep 26 18:01:47 1994 David J. MacKenzie (djm@churchy.gnu.ai.mit.edu) * lib/xwrite.c safe-read.c memchr.c linebuffer.c full-write.c: Remove CONFIG_BROKETS ifdef. * src/Makefile.in lib/Makefile.in: Don't define it. Sat Sep 24 21:26:27 1994 Jim Meyering (meyering@comco.com) * full-write.c (full_write): Declare argument LEN to be size_t. Mon Sep 12 13:35:27 1994 Jim Meyering (meyering@comco.com) * tail.c (tail_file): [For multiple files only]: Report truncation of monitored file and reset current file size. From Franc,ois Pinard. Sat Aug 27 16:57:20 1994 Jim Meyering (meyering@comco.com) * system.h [BROKEN_STAT_MACROS]: Remove unnecessary #ifdef's. From Franc,ois Pinard. Tue Jul 26 11:33:53 1994 Jim Meyering (meyering@comco.com) * system.h: Remove `|| defined(_POSIX_VERSION)' from test that decides whether to include . From Francois Pinard. Wed Jul 13 12:33:34 1994 Jim Meyering (meyering@comco.com) * tr.c (substr): Fix off-by-one allocation error. Wed Jun 22 01:02:50 1994 Jim Meyering (meyering@comco.com) * tac.c (tac_file): Use O_RDONLY instead of zero in call to open. (tac): Fix typo that had `1' as arg #2 of error after failed read. Use errno instead. From Michael I Bushnell . Thu May 26 08:46:32 1994 Jim Meyering (meyering@comco.com) * configure.in [INSTALL]: Revert change of March 25. autoconf has been fixed. Mon Apr 11 17:25:43 1994 Jim Meyering (meyering@comco.com) * join.c (prline): Remove unused function. Indent. Fri May 6 05:44:24 1994 Jim Meyering (meyering@comco.com) * tail.c (tail_forever): Don't print headers when asked not to. From Karsten Thygesen . * wc.c (wc): Optimize for the case when only bytes are being counted and the input is not a regular file. From Jeff Moore . Sun Apr 24 10:54:08 1994 Jim Meyering (meyering@comco.com) * configure.in [AC_SIZEOF_TYPE]: Update for autoconf-1.9. * memchr.c: Test SIZEOF_LONG instead of LONG_64_BITS. Mon Apr 11 17:55:52 1994 Jim Meyering (meyering@comco.com) * csplit.c (struct control): Add new field, repeat_forever. Remove now-unneeded definition of INT_MAX. (process_regexp): Test repeat_forever instead of comparing `repeat' to INT_MAX. (split_file): Don't even reference `repeat' count if repeat_forever is set. (new_control_record): Initialize repeat_forever field. (parse_repeat_count): Set it instead of setting repeat count to INT_MAX. (process_line_count): Abort for internal error instead of calling handle_line_error. * long-options.c (parse_long_options): Take a new argument indicating the utility name (e.g. "join", not "gjoin") for version output. Print both the utility and the package names, e.g. `join - GNU textutils-1.9.1' instead of just the package name. * join.c (main): Call parse_long_options with program name arg. * sort.c (main): Ditto. * All source except sort.c and join.c (main): Change --version output to include utility name as well as package name and version info. Wed Mar 30 08:53:21 1994 Jim Meyering (meyering@comco.com) * configure.in: Use AC_SET_MAKE. * Makefile.in: Edit MAKE assignments into @SET_MAKE@. Mon Mar 28 09:55:05 1994 Jim Meyering (meyering@comco.com) * tr.c (main): In addition to the --help usage pointer, give an explicit error message for too many arguments. Fri Mar 25 18:11:19 1994 Jim Meyering (meyering@comco.com) * configure.in: Use AC_VERBOSE, AC_CHECKING, and AC_WARN instead of explicit echo commands. [INSTALL]: When AC_PROG_INSTALL resorts to setting this to `./install.sh', make it `../install.sh' instead since that script will be invoked only from subdirectories. [AC_SIZEOF_TYPE]: Use this instead of deprecated AC_LONG_64_BITS. * cat.c (cat): The command `:|cat -s' failed on Irix5 because ioctl (d, FIONREAD,... now returns ENOSYS for pipes -- under Irix4 it returned EINVAL. Detect this. (main) [lint]: Initialize variables to suppress `used before initialized' warnings. * csplit.c: Change long option name from --suffix to --suffix-format. * od.c (skip): Correct a comment. (main): Don't output anything to stdout if the sole input file doesn't exist. Reported by Bauke Jan Douma . * od.c (dump) [lint]: Initialize a variable to suppress `used before initialized' warning. (main) [lint]: Ditto. * paste.c (paste_parallel) [lint]: Ditto. Fri Jan 28 11:02:21 1994 Jim Meyering (meyering@comco.com) * configure.in: Don't set LDFLAGS since linking now uses both LDFLAGS and CFLAGS. Wed Jan 18 16:14:00 1994 Jim Meyering (meyering@comco.com) * unexpand.c (unexpand): Don't segfault when given a name of a nonexistent file. * expand.c (expand): Ditto. Both from Bauke Jan Douma . Thu Jan 13 17:27:38 1994 Jim Meyering (meyering@comco.com) * src/Makefile.in: Change all link commands to use both $(CFLAGS) and $(LDFLAGS). * csplit.c (load_buffer): Rather than incrementing buffer size, double it each time we fail to find a newline. Otherwise, csplit would run out of memory when processing files with very long lines. Mon Jan 10 01:20:38 1994 Jim Meyering (meyering@comco.com) * man/Makefile.in (manprefix): Use binprefix as the default. Sat Jan 08 22:22:45 1994 Jim Meyering (meyering@comco.com) * cat.c: Use full_write and safe_read instead of write and read resp. * csplit.c: Ditto. * head.c: Ditto. * split.c: Ditto. * tac.c: Ditto. * sum.c: Use safe_read instead of read. * tr.c: Ditto. * wc.c: Ditto. * tail.c: Replace calls to xwrite with calls to fwrite. Use fopen/fclose instead of open/close. (write_header): Use a single call to printf instead of four to xwrite. Use safe_read instead of read. * head.c: Ditto. Tue Dec 28 15:49:32 1993 Jim Meyering (meyering@comco.com) * install.sh: New file. Makefile.in [DISTFILES]: Add it. Wed Dec 22 18:52:44 1993 Jim Meyering (meyering@comco.com) * memcmp.c: Use the latest version from the GNU C library. * cut.c [ADD_RANGE_PAIR]: New macro. (set_fields): Collect the list of all selected ranges before allocating and initializing the boolean lookup table. (cut_bytes, cut_fields): Complete rewrite. Avoid copying into buffer whenever possible. Properly handle input without trailing newline. (getstr): New function. Copied from getline.c, but with minor changes. * sort.c (main): Properly handle -Tdir. Before, `sort -T/var/tmp' gave `sort: unrecognized option `-v''. Reported by Kristoffer Rose (kris@diku.dk). * cut.c (main): Give separate errors for `-s without -f' and `-d without -f'. (main): Now -d '' means `use the NUL byte as the delimiter'. Before, it got an error. Mon Dec 20 23:29:30 1993 Jim Meyering (meyering@comco.com) * configure.in [test for 8-bit clean memcmp]: Add a test to detect losing memcmp from SunOS4.1.x. From Robert H. de Vries . Sat Dec 18 01:12:24 1993 Jim Meyering (meyering@comco.com) * configure.in (AC_OUTPUT): Put `touch stamp-config' in second arg so it goes in config.status. This eliminates unnecessary second run of configure. Thu Dec 02 23:53:03 1993 Jim Meyering (meyering@comco.com) * configure.in (AC_HAVE_FUNCS): Add isascii. * system.h [!defined (isascii) || defined (STDC_HEADERS)]: This failed on AIX PS/2 1.3 systems because isascii is a function and it is used in definitions (with the necessary side effect of assigning to a global variable) of the is* macros. Also test HAVE_ISASCII and redefine ISASCII(c) instead of isascii. Reported by Minh Tran-Le (tranle@intellicorp.com). Fri Nov 19 22:41:48 1993 Jim Meyering (meyering@comco.com) * configure.in: Don't run the test for 8-bit clean memcmp if we already know we'll be using our replacement. Fri Nov 12 00:44:49 1993 Jim Meyering (meyering@comco.com) * Version 1.9.1. * tac.c (usage): Fix typo in long usage. * sort.c (usage): Likewise. From Philippe.Schnoebelen@imag.fr. * cut.c [FATAL_ERROR]: New macro. (main): Call both error and usage (2) through FATAL_ERROR in many places. Before, only `error (2...' was called. (invalid_list): Removed. This function is subsumed by FATAL_ERROR. * cut.c (cut_fields): Properly handle input lacking a terminating newline. Before, the command `printf "a\tb" |cut -f 1' generated no output. Tue Nov 09 17:26:25 1993 Jim Meyering (meyering@comco.com) od.c (usage): Change --backward-compatible to --traditional in the long usage message. Sun Nov 07 00:50:05 1993 Jim Meyering (meyering@comco.com) * Version 1.9. Sat Nov 06 22:51:31 1993 Jim Meyering (meyering@comco.com) * cut.c (cut_fields): Revert July 27 change. Instead, strip off trailing delimiter unless given a range of fields like `3-' that extends to end of line. (set_fields): Moved local `eol_range_start' to file-scope. Reported by Arne H. Juul (arnej@solan.unit.no). Tue Nov 02 00:53:41 1993 Jim Meyering (meyering@comco.com) * fold.c (fold_file): Don't get stuck in an endless loop when width is smaller than 8 and there are TABs in the input. Sat Oct 30 15:31:28 1993 Jim Meyering (meyering@comco.com) * join.c: Remove now-unused dcls of show_help and show_version. Fri Oct 29 13:58:50 1993 Jim Meyering (meyering@comco.com) * csplit.c [INT_MAX]: Make sure it's defined. [not HAVE_LIMITS_H]: Don't include limits.h. From Kaveh R. Ghazi (ghazi@noc.rutgers.edu). Wed Oct 27 01:13:52 1993 Jim Meyering (meyering@comco.com) * Version 1.8.1. Mon Oct 25 20:16:33 1993 Jim Meyering (meyering@comco.com) * sort.c: Accept but ignore -y0 for compatibility with Solaris 2. From Chuck Hedrick (hedrick@klinzhai.rutgers.edu). * sort.c (main): Handle --help and --version options. * Makefile.in (sort.o): Depend on long-options.h. Sun Oct 24 00:31:02 1993 Jim Meyering (meyering@comco.com) * csplit.c, cut.c, expand.c, fold.c, head.c, nl.c, od.c, paste.c, split.c, tac.c, tail.c, unexpand.c, uniq.c: Use the preferred `--longopt=arg' syntax in --help message rather than `--longopt arg'. From Francois Pinard. * tail.c: Don't include . It hasn't been necessary since the Dec 12, 1992 change. * join.c (main): Accept `-v 1' again. Adding the --version long option had broken it, although -v1 still worked. Call parse_long_options instead of adding "help" and "version" to join's longopt strct. * Makefile.in [SOURCES, OBJECTS, DISTFILES]: Add long-options.c and long-options.h. * (join): Depend on and link with long-options.o. * (join.o): Depend on long-options.h. * od.c: Change --compatible (-C) to --backward-compatible (-B). * csplit.c: Change --abandon-null-files to --elide-empty-files. Sat Oct 23 01:00:12 1993 Jim Meyering (meyering@comco.com) * tr.c (get_next, string2_extend): Add default: label to switch stmt. * cat.c, cksum.c, comm.c, csplit.c, cut.c, expand.c, fold.c, head.c, join.c, nl.c, od.c, paste.c, pr.c, split.c, sum.c, tac.c, tail.c, tr.c, unexpand.c, uniq.c, wc.c (usage): Add long --help. Exit successfully for --help, non-zero for usage error. From Francois Pinard. * configure.in: Add AC_STAT_MACROS_BROKEN. * system.h [AC_STAT_MACROS_BROKEN]: Test it. Fri Oct 22 23:26:17 1993 Jim Meyering (meyering@comco.com) * linebuffer.c, memchr.c [HAVE_CONFIG_H, CONFIG_BROKETS]: Include or "config.h". * cat.c, cksum.c, comm.c, csplit.c, cut.c, expand.c, fold.c, head.c, join.c, nl.c, od.c, paste.c, pr.c, sort.c, split.c, sum.c, tac.c, tail.c, tr.c, unexpand.c, uniq.c, version.c, wc.c: Ditto. * configure.in: Use AC_CONFIG_HEADER. * Makefile.in [DIST]: Add config.h.in. * Makefile.in: Convert so make may be run from subdirectories. Add dependencies on config.h. * src/Makefile.in: Ditto. * lib/Makefile.in: Ditto. * man/Makefile.in: Ditto. Wed Oct 20 20:05:48 1993 Jim Meyering (meyering@comco.com) * memchr.c (memchr): Do the 32-bit assignment only if !LONG_64_BITS. In the 64-bit assignment, be careful to cast the shift operand to long. Abort if sizeof (unsigned long) > 8. Tue Oct 19 22:37:58 1993 Jim Meyering (meyering@comco.com) * csplit.c: Allow repeat counts to be specified via `{*}'. New option --suffix=format. Supercedes --digits option. New option --abandon-null-files. From Ronald F. Guilmette (rfg@netcom.com). * csplit.1: Updated. From Ronald F. Guilmette. * csplit.c: Remove register keyword (replace with `int' in two cases). * csplit.c: [MAX]: Macro renamed from max and moved to top of file. (bytes_to_octal_digits): New static array. (get_format_flags): Combine '+' and ' ' cases of switch stmt. Return count if for loop terminates. (get_format_width): Use `bytes_to_octal_digits' instead of just 11. (get_format_prec): Make sure is_negative is defined before used. (get_format_conv_type): Give a different error message if there is no conversion specifier. Test ISPRINT (ch) instead of `ch < '~' && ch > ' ''. (max_out): Use `%%' rather than `%' in format string. Sat Oct 16 10:45:17 1993 Jim Meyering (meyering@comco.com) * pr.c (main): When argc == 1, don't try to xmalloc (0). Tue Oct 12 00:53:26 1993 Jim Meyering (meyering@comco.com) * xwrite.c [HAVE_CONFIG_H, CONFIG_BROKETS]: Include or "config.h". Sat Oct 09 23:37:43 1993 Jim Meyering (meyering@comco.com) * configure.in: Remove AC_UNISTD_H; add unistd.h to AC_HAVE_HEADERS. Thu Sep 9 21:52:10 1993 Jim Meyering (meyering@comco.com) * src/*.c: Print version on standard output, not stderr. * configure.in: Add AC_LONG_64_BITS. * memchr.c: Use #ifdef LONG_64_BITS instead of `if (sizeof(longword) > 4)'. Tue Jul 27 22:19:39 1993 Jim Meyering (meyering@comco.com) * cat.c (cat): Don't fail just because ioctl (d, FIONREAD,... can't always handle devices. Ignore errno == EINVAL and errno == ENODEV. `cat -v /dev/null' was failing on many systems. * cut.c (cut_fields): Don't strip off trailing delimiter e.g. `echo 'a:b:c:' | cut -d: -f3-' should print `c:', not just `c'. From William Dowling . Thu May 27 01:37:51 1993 Jim Meyering (meyering@comco.com) * Version 1.8. * memchr.c: De-ansify the fixed version from glibc. It is supposed to work on systems with 64-bit long ints. Mon May 24 00:32:43 1993 Jim Meyering (meyering@comco.com) * Version 1.7. Sat May 22 02:13:12 1993 Jim Meyering (meyering@comco.com) * Version 1.6. Fri May 21 22:57:53 1993 Jim Meyering (meyering@comco.com) * pr.c (cols_ready_to_print): New function. It replaces a global variable by the same name and computes on the fly the value that was supposed to be maintained through that variable. This should put to rest the `pr -2a' bug. (main): Don't let getopt_long reorder arguments so we can distinguish between `pr -1 -2' and `pr -12'. Rework handling of -n and +n options. * fold.c: Declare xmalloc. Wed May 19 19:12:18 1993 Karl Berry (karl@owl.hq.ileaf.com) * sort.c (main): fflush before exit, so a closed stdout doesn't lose. Tue May 18 23:49:26 1993 Jim Meyering (meyering@comco.com) * mkinstalldirs: New file. * Makefile.in (installdirs): Use it to create installation directories. Fri May 14 23:45:52 1993 Jim Meyering (meyering@comco.com) * all source: With --version, print version and exit immediately. Thu May 13 01:03:16 1993 Jim Meyering (meyering@comco.com) * Makefile.in (installdirs): New rules for creating installation directories. (install): depend on it. * tail.c (main): Remove --compatible since `+1f' may be used to get the BSD `-0f' behavior portably. * fold.c (main): Turn -N arguments, where N is a digit, into -wN. From Ian Lance Taylor (ian@cygnus.com). Mon May 10 22:33:44 1993 Jim Meyering (meyering@comco.com) * tail.c (main): New option -C, --compatible to make `tail -0f' work like `tail +1f' for compatibility with BSD tail. Thu May 6 23:28:56 1993 Jim Meyering (meyering@comco.com) * pr.c (print_page): Decrement cols_ready_to_print even when lines_to_print == 0. The command `echo |pr -2ta' *still* didn't terminate. Add parentheses for clarity. Mon May 3 23:57:47 1993 Jim Meyering (meyering@comco.com) * Version 1.5. * od.c (main): Remove set-but-not-used variable, `usage_error'. * split.c (main): Call usage with an argument. Sun May 2 16:05:04 1993 Jim Meyering (meyering@comco.com) * configure.in: Reverse if-else branches in AC_TEST_PROGRAM test for 8-bit clean memcmp. * cat.c (usage, main): Remove unused argument. * csplit.c (remove_line): Remove set-but-not-used variable, `line_got.' (check_for_offset): Remove unused argument `argnum.' * configure.in: Add AC_SIZE_T. Sat May 1 09:03:19 1993 Jim Meyering (meyering@comco.com) * uniq.c (main): Interpret non-option arguments with a leading `+' only if we haven't seen `--'. * tr.c (main): Change variable name to avoid shadowed dcl. * wc.c (write_counts): Ditto. * sum.1: Make documentation agree with the code. * od.c: Depend only on HAVE_LONG_DOUBLE for long double support. Fri Apr 30 20:16:03 1993 Jim Meyering (meyering@comco.com) * configure.in [AC_HAVE_HEADERS]: Add limits.h. * configure.in [CFLAGS, LDFLAGS]: Assign reasonable defaults. * od.c (parse_old_offset): Don't use prototype in function definition. Remove unnecessary conjunct from test for hexadecimal prefix. * od.c: Depend on __GNUC__ || HAVE_LONG_DOUBLE rather than __STDC__ for long double support; there are compilers (Stardent Vistra svr4) without long double but still define __STDC__. Thu Apr 29 02:01:27 1993 Jim Meyering (meyering@comco.com) * src/*.c and man/*.c except for sort: Add --help and --version options. * pr.c: Convert to use getopt_long. * lib/memcmp.c: Use version from glibc. [WORDS_BIGENDIAN]: Test this instead of the using glibc's __BYTE_ORDER macro. * configure.in: Use AC_WORDS_BIGENDIAN to set it. * od.c (parse_old_offset): Allow `0X' as well as `0x' prefix to indicate hex. * Makefile.in (dist): Depend on Makefile so that changes to Makefile.in (like adding new files to DISTRIB) are reflected in the new distribution. Fri Apr 23 21:53:47 1993 Jim Meyering (meyering@comco.com) * {lib,src}/Makefile.in [.c.o]: Make CPPFLAGS and CFLAGS follow other options so users can use them to override DEFS. * src/Makefile.in [.c.o]: Add -I$(srcdir) for version.h. * Makefile.in [CFLAGS, LDFLAGS]: Don't set to -g explicitely; configure's AC_SUBST will set these. * Makefile.in [Makefile]: Add dependencies and a rule to remake it. [targets that cd then run make in subdirectories]: Don't depend on `cd ..'; use a subshell instead. * Makefile.in [info, install-info, dvi, check, installcheck]: New targets but no rules; comply with standards. * configure.in [CFLAGS]: Default to -g -O if CC is gcc and the user hasn't specified CFLAGS. [LDFLAGS]: Substitute into Makefile. Tue Mar 30 21:36:11 1993 Jim Meyering (meyering@comco.com) * od.c: New option --compatible, -C. (parse_old_offset): New function. * od.1: Document new option. Mon Mar 29 21:27:56 1993 Jim Meyering (meyering@comco.com) * cut.c, expand.c, join.c, nl.c: Always call error with errno (not zero) after failed fclose or non-zero ferror. Sun Mar 28 16:59:31 1993 Mike Haertel (mike@cs.uoregon.edu) * configure.in: Add check for working memcmp; use GNU's if the system's doesn't grok the 8th bit. * memcmp.c: Fix it so it groks the 8th bit. TODO: We really need to provide a fast memcmp, since most machines will have a broken memcmp. Probably should get the one from glibc. * sort.c (mergefps): Maintain keybeg and keylim when copying the current line to `saved'. (numcompare): Skip white space here since -n no longer implies -b. (getmonth): Skip white space here since -M no longer implies -b. (compare): Completely overhauled to make the 8th bit work right, also to properly handle the global reverse option. (set_ordering): -n no longer implies -b, according to Posix. For consistency, -M also no longer implies -b. (main): Correct treatment of -r and global keys. (findlines): Clear keybeg and keylim if no keys are used. (sort): Avoid overwriting tempfiles[] array bounds. Sun Mar 21 22:29:29 1993 Jim Meyering (meyering@comco.com) * pr.c (close_file): Reverse May 13, '92 change, but add the condition that cols_ready_to_print not be decremented when printing across. The command `echo |pr -2ta' didn't terminate. (print_page): Rewrite conditional (that had side effects in second conjunct) using nested if statements to make it clear that we do indeed want those semantics. Tue Jan 19 13:35:24 1993 David J. MacKenzie (djm@kropotkin.gnu.ai.mit.edu) * Version 1.4. * system.h: Try BBSIZE if BSIZE isn't defined. From Tony Robinson . Sat Dec 12 12:37:00 1992 David J. MacKenzie (djm@kropotkin.gnu.ai.mit.edu) * tail.c (tail_forever): Use an array of file descriptors instead of forking processes. (dump_remainder): Return number of bytes read. (tail_file): Fill in the new array. From Ian Lance Taylor. Fri Dec 11 17:18:16 1992 David J. MacKenzie (djm@kropotkin.gnu.ai.mit.edu) * system.h: Only define index, bcmp, etc. if not already defined. Tue Dec 8 10:31:14 1992 Jim Meyering (meyering@idefix.comco.com) * tr.c (is_char_class_member): Remove unreached return after abort. Sun Dec 6 22:34:52 1992 Jim Meyering (meyering@idefix.comco.com) * csplit, cut.c, expand.c, fold.c, head.c, join.c, od.c, pr.c, sort.c, split.c, tail.c, tr.c, unexpand.c, uniq.c: Remove inclusion of and definitions of is* ctype macros to system.h. Change a few more uses of is* ctype macros to (protected) upper case versions. * system.h: Add isascii-protected ctype IS* macros. Wed Dec 2 12:28:10 1992 Jim Meyering (meyering@idefix.comco.com) * all files using getopt.h: Convert static declarations of struct option to use new macros from getopt.h: no_argument, required_argument, and optional_argument. Tue Dec 01 10:57:24 1992 Jim Meyering (meyering@idefix.comco.com) * od.c, pr.c, sort.c: Make uses of ctype.h macros consistent. * tr.c (main): Close stdin and stdout and check return status. Tue Nov 24 09:26:08 1992 David J. MacKenzie (djm@goldman.gnu.ai.mit.edu) * system.h, csplit.c: Use HAVE_FCNTL_H and HAVE_STRING_H instead of USG. Tue Nov 24 08:42:30 1992 Jim Meyering (meyering@idefix.comco.com) * tr.c: Define new macros ISPRINT, ISALNUM, ISXDIGIT, ... that use isascii if it is defined. (is_char_class_member, make_printable_char, make_printable_str, non_neg_strtol): Use new macros instead of lower case ones from . * od.c (print_ascii, dump_strings): Use ISDIGIT and ISPRINT macros like pr.c. Suggested by David J. MacKenzie. * od.c (print_ascii, dump_strings): Check for isascii before isprint. (dump_strings): Free malloc'd buffer before returning. (skip): Return non-zero if an error occurred, zero otherwise. Exit only if asked to skip beyond end of combined input. (check_and_close): New function. (skip, read_block, read_char): Call check_and_close when done processing the file associated with in_stream. (skip, read_block, read_char): Set have_read_stdin. (main): Close stdin (if it was ever read) and check for errors just before exiting. (write_block, dump_strings, dump): Don't test return value from functions that operate on streams. Rely on later ferror tests. Sat Nov 21 12:41:49 1992 David J. MacKenzie (djm@goldman.gnu.ai.mit.edu) * sort.c (main, usage): Add -T option. Thu Nov 19 14:33:40 1992 David J. MacKenzie (djm@goldman.gnu.ai.mit.edu) * tail.c (tail_forever, sigusr1, kill_kids): New functions to do -f on multiple files. (main): Call tail_forever. From Ian Lance Taylor. Tue Nov 10 14:29:11 1992 David J. MacKenzie (djm@goldman.gnu.ai.mit.edu) * cut.c (cut_fields): Add cast. * od.c (decode_one_format): Remove '#' from pre_fmt_string; many systems don't support it. Conditionalize long double support on __STDC__, not __GNUC__. From Ian Lance Taylor. Mon Nov 9 00:24:41 1992 Jim Meyering (meyering@idefix.comco.com) * sort.c (numcompare, keycompare): Add parentheses suggested by gcc -Wall. Put braces around individual monthtab initializers. * cksum.c: Declare error. Make checksum table `const.' * pr.c: Remove comment and dcl of unused variable, `print_a_number'. * split.c (main): Add `default: abort();' to enumeration switch. * All files: Make all functions and extern variables static. Make all longopts arrays const as well as static. Make a couple statically initialized aggregates `const.' Sun Nov 8 19:46:59 1992 David J. MacKenzie (djm@goldman.gnu.ai.mit.edu) * od.c (main): Make old-style format options accumulate. From Jim Meyering. Sat Nov 7 00:26:14 1992 David J. MacKenzie (djm@goldman.gnu.ai.mit.edu) * tr.c (look_up_char_class): Add cast. * nl.c (build_type_arg), csplit.c (extract_regexp), tac.c (main): Add `const' to variable receiving value from re_compile_pattern. * wc.c (wc): If doing only -c, use st_size for regular files. * fold.c (fold_file): Was folding 1 column too early. From Eric Backus . * memset.c: New file. Fri Nov 6 20:14:51 1992 David J. MacKenzie (djm@goldman.gnu.ai.mit.edu) * cksum.c: New file. Tue Oct 13 16:24:06 1992 David J. MacKenzie (djm@goldman.gnu.ai.mit.edu) * tac.c (tac_stdin): Handle SIGPIPE. * sort.c (main): Handle SIGTERM. * od.c: New file. * system.h [USG || STDC_HEADERS]: Define bcmp. Sat Oct 3 20:41:24 1992 David J. MacKenzie (djm@goldman.gnu.ai.mit.edu) * sort.c (main): Handle SIGPIPE. From trq@dionysos.thphys.ox.ac.uk. Tue Sep 29 01:10:05 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) * paste.c (main): Don't write on a string constant. Mon Aug 24 00:02:45 1992 Jim Meyering (meyering@churchy.gnu.ai.mit.edu) * tr.c: Minor cleanup. Replaced some assert(0) with abort(). Tue Jul 7 02:14:19 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) * cmp.c, cmp.1: Move to diff distribution. Fri Jul 3 16:37:59 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) * system.h: Change FOO_MISSING to HAVE_FOO. Wed May 13 20:05:41 1992 David J. MacKenzie (djm@churchy.gnu.ai.mit.edu) * pr.c (COLUMN): Add structure member to remember filename. (main, init_fps, open_file, close_file): Use it. (close_file): Don't decrement cols_ready_to_print when closing a file. From cdl@mpl.UCSD.EDU (Carl Lowenstein). Mon May 11 19:17:33 1992 David J. MacKenzie (djm@churchy.gnu.ai.mit.edu) * cmp.c: --show-chars -> --print-chars. * pr.c: Rename some variables. Sat May 9 18:39:47 1992 David J. MacKenzie (djm@wookumz.gnu.ai.mit.edu) * system.h: Define DEV_BSIZE if not defined. Wed Apr 22 02:15:09 1992 David J. MacKenzie (djm@churchy.gnu.ai.mit.edu) * system.h, tac.c: SIGTYPE -> RETSIGTYPE. Fri Apr 17 10:42:23 1992 David J. MacKenzie (djm@wookumz.gnu.ai.mit.edu) * sort.c (main): Don't stop processing args when we hit "-"; treat it like a regular filename. From ian@airs.com (Ian Lance Taylor). * pr.c (print_page): Fix off by one line count when ^L is in input. From Andreas Schwab (schwab@ls5.informatik.uni-dortmund.de). Mon Apr 6 20:52:29 1992 Jim Meyering (meyering@churchy.gnu.ai.mit.edu) * tr.c (validate): Change error message so it doesn't mention actual name of --truncate-set1 option. From David MacKenzie. Sun Apr 5 14:22:42 1992 Jim Meyering (meyering@hal.gnu.ai.mit.edu) * tr.c (string2_extend, validate): Give an error message when translating without --truncate-set1, with empty string2, and with non-empty string1. "tr 1 ''" produced a failed assertion. Mon Mar 30 02:20:56 1992 David J. MacKenzie (djm@wookumz.gnu.ai.mit.edu) * system.h: Change how ST_BLKSIZE is calculated to allow for non-POSIX systems that don't define BSIZE in sys/param.h. Sat Mar 28 11:18:01 1992 David J. MacKenzie (djm@wookumz.gnu.ai.mit.edu) * sum.c (main, bsd_sum_file): Don't print filename if BSD algorithm is used and only one file was given. Wed Mar 25 11:34:41 1992 Jim Meyering (meyering@wombat.gnu.ai.mit.edu) * tr.c (get_spec_stats): Fix assertion to allow ranges like a-a with starting character equal to ending character. This is contrary to the POSIX spec, but what is already implemented in find_closing_delim. Mon Mar 16 00:15:11 1992 David J. MacKenzie (djm@wookumz.gnu.ai.mit.edu) * Version 1.3. * sort.c (numcompare, checkfp): Add parens to placate gcc2. * sort.c (mergefps): For -u, output the first, not last, of the lines that compare equal. From Mike Haertel. Tue Mar 10 10:51:38 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) * tr.c: Remove initial capitals and periods from error messages. Sun Mar 8 22:03:45 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) * sum.c (main): Add -r option for SYSV compat. Thu Feb 27 22:26:25 1992 David J. MacKenzie (djm@wookumz.gnu.ai.mit.edu) * sort.c (compare): If -s given, leave lines in their original order. (main): Recognize -s. (usage): Document -s. From Mike Haertel. Tue Feb 18 20:29:45 1992 Randall Smith (randy at geech.gnu.ai.mit.edu) * sort.c (sort): Check for complete parsing of buffer into lines before nixing temp files. Mon Feb 17 10:35:58 1992 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * sum.c (sysv_sum_file): Use %lu instead of %u to print a long. Not that it matters for GNU . . . * tr.c (unquote, make_printable_str): Use \007 instead of ANSI \a. (append_normal_char, append_range, append_char_class, append_repeated_char, append_equiv_class, spec_init): Initialize `next' field of new `struct List_element'. From rommel@informatik.tu-muenchen.de (Kai-Uwe Rommel). Sat Feb 8 17:16:49 1992 David J. MacKenzie (djm at apple-gunkies.gnu.ai.mit.edu) * join.c (get_line): Renamed from getline to avoid GNU libc conflict. Sun Feb 2 21:22:01 1992 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * Version 1.2. * nl.c: Support multiple files and "-" for stdin. (main): Check for read and write errors. (nl_file): New function. Wed Jan 29 10:09:10 1992 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * tr.c (main): -t option was called -b in getopt spec. (validate): Don't warn that set1 is longer than set2. * tr.c: Rename --sysv-string2-truncate to --truncate-string1. Fri Jan 17 16:29:05 1992 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * nl.c: New program from bin-src. * nl.c (main): Use a struct linebuffer for line_buf. (process_file): Use readline instead of fgets, to preserve NULs. (check_section): Use memcmp instead of strncmp. (proc_text): Print line_buf with fwrite instead of printf. * nl.c (main): Usage message if too many args given. Check for error in closing input file. Lengths of section delimiter strings were 1 too large. Take separator_str into account in length of print_no_line_fmt. (build_print_fmt): Allocate space for print_fmt, in case separator_str is long. (proc_text): A blank line is one that contains nothing, not even nonprinting characters. Fri Jan 17 01:04:22 1992 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * All programs: Document `--' instead of `+' to introduce long-named options, in usage messages. * sum.c (bsd_sum_file): Renamed from sum_file. (sysv_sum_file): New function. (main): Recognize an option to select between the algorithms. Sun Jan 5 17:41:18 1992 Jim Meyering (meyering at churchy.gnu.ai.mit.edu) * pr.c (close_file, print_page): Fixed bug that had the command yes |head |pr -t printing "yyyyyyyyyy". * (print_page): Fixed bug that had pr -3 -a printing two too few trailer lines per page. * (main): Added restriction that -a and -m are incompatible. Although the POSIX spec doesn't explicitly say they shouldn't be used together, it says -a modifies the -column option and that -column shouldn't be used with -m. Thu Jan 2 15:23:59 1992 David J. MacKenzie (djm at albert.gnu.ai.mit.edu) * nl.c: Include regex.h after, not before, sys/types.h. Thu Jan 2 12:18:10 1992 Tom Lord (lord at geech.gnu.ai.mit.edu) * sort.c (fillbuf) return bytes buffered instead of bytes read. Fri Dec 27 22:53:36 1991 Jim Kingdon (kingdon at geech.gnu.ai.mit.edu) * sort.c (LINEALLOC): New #define. (struct lines): New field ``limit''. (initlines): Set it from new arg ``limit''. (sort, mergefps, checkfp): Pass new arg to initlines(). (findlines): Don't realloc past lines->limit. Tue Dec 24 01:24:03 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * tac.c, sort.c, csplit.c, system.h: Change POSIX ifdefs to HAVE_UNISTD_H and _POSIX_VERSION. * xwrite.c: Change POSIX ifdef to HAVE_UNISTD_H. Sat 14 Dec 1991 11:46:42 Jim Meyering (meyering at wombat) * tr.c: Fixed an inaccurate comment on posix_pedantic. Thu 12 Dec 1991 21:15:20 Jim Meyering (meyering at hal) * tr.c: Changed underscores to hyphens in long option name "sysv_string2_truncate". Wed Dec 11 13:33:34 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * tac.c (main): Set obscure_syntax to tell re_search to allocate memory for the group registers. Fri Dec 6 18:26:27 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * tac.c, sort.c, csplit.c [POSIX]: Use sigaction instead of signal, which POSIX doesn't have. * sort.c: Replace inthandler and huphandler with sighandler. * csplit.c (main): Only handle signals if they were not being ignored. * tr.c: POSIX_ME_HARDER -> POSIXLY_CORRECT. Wed Dec 4 00:47:47 1991 Jim Meyering (meyering at wombat) * tr.c (unquote): Reformat code so it doesn't go beyond column 80. * tr.c (squeeze_filter): Comment a little on why it's better to step through the input by two. * tr.c (set_initialize): Write a comment describing the function. * tr.c: Eliminated the variable `portability_warnings' and replaced references to it by references to `!posix_pedantic'. One of the uses of portability_warnings had been wrong. Tue Dec 3 14:03:35 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * tr.c: New program. Sun Dec 1 15:07:35 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * linebuffer.[ch] (freebuffer): New function (used by cron). Thu Oct 17 22:30:22 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * system.h, configure, Makefile.in: Don't include memory.h if STDC_HEADERS, removing need for MEMORY_H_MISSING. Thu 17 Oct 1991 16:53:07 Jim Meyering (meyering at wombat) * pr.c (print_page): REALLY fixed `extra newline at EOF' problem. Somehow, part of my patch didn't make it last time. Sat Oct 12 12:04:47 1991 David J. MacKenzie (djm at churchy.gnu.ai.mit.edu) * tail.c (pipe_lines, pipe_bytes): Initialize `first->next'. * cmp.c (cmp): Print EOF message to stderr, not stdout, for POSIX 1003.2.11.2. * sort.c (xfwrite): fwrite never returns < 0, so check if number written is number we asked to write. (fillbuf, main): fread never returns < 0, so check ferror instead. From Rainer Orth. Tue Oct 8 18:07:08 1991 Jim Meyering (meyering at churchy) * pr.c (print_page): Really fixed `extra newline at EOF' problem. * (store_columns): Fixed bug that caused `pr -b -2' to coredump on files of certain lengths. Fri Oct 4 22:30:25 1991 Jim Meyering (meyering at churchy) * pr.c (print_page): Fixed to not add single spurious newline at EOF when using -t. Wed Oct 2 01:02:05 1991 David J. MacKenzie (djm at apple-gunkies) * pr.c (print_page): Don't pad the page if -t given. * csplit.c (load_buffer), sort.c (mergefps): Use bcopy, not memcpy. Thu Sep 26 12:35:00 1991 David J. MacKenzie (djm at churchy.gnu.ai.mit.edu) * Version 1.1. * configure, system.h: Include memory.h if it works. * split.c: Allow `b' unit as well as `k' and `m'. * head.c, tail.c: Replace -b +blocks option with specifying units (b, k, or m) after the number. (parse_unit): New function. * fold.c (main): Check that -w arg is a number. * cut.c: +delimiter takes an arg. Mon Sep 16 14:52:38 1991 David J. MacKenzie (djm at churchy.gnu.ai.mit.edu) * pr.c (close_file): Don't close an already closed file. Thu Sep 12 00:14:43 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * memchr.c: New file. * configure: Check if it's needed. * csplit.c, gcsplit.1: New program. * pr.c (cleanup): Only free buffers if they were allocated. * sort.c [!USG && !STDC_HEADERS]: Declare memchr. Wed Sep 11 20:54:16 1991 Jim Meyering (meyering at churchy) * pr.c: The following 3 bugs appeared (at least) when printing a single file with the options `-3 -f'. * (print_white_space): Single spaces were being replaced with tabs. * (print_page): Some lines were getting too much white space at the beginning because spaces_not_printed wasn't being reset to 0. * (read_line): The single space between a truncated column on its left and the column on its right was omitted. Fixed so that previous value of input_position is restored before returning FALSE. Sat Sep 7 03:22:18 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * configure: Only remove /etc from PATH when it's not part of a larger name. Wed Sep 4 17:09:24 1991 David J. MacKenzie (djm at apple-gunkies) * linebuffer.c (readline): Fix incorrect recalculation of `end'. * head.c, tail.c: Replace `mode' variables and bitmasks with separate variables for each option. Mon Sep 2 04:00:37 1991 David J. MacKenzie (djm at apple-gunkies) * wc.c: New program. Sun Sep 1 01:18:38 1991 David J. MacKenzie (djm at apple-gunkies) * fold.c (fold_file): Read in an int, not a char, for EOF comparison. * configure: Check whether st_blksize is missing. * tac.c (save_stdin): Put copy of pipe input in TMPDIR if defined, instead of /tmp. Thu Aug 29 14:48:15 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * xwrite.c [POSIX]: unistd.h might require sys/types.h. Wed Aug 28 11:57:39 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * paste.c (main): Consider -d "" to be like -d "\0", for POSIX (if I read it right). * sort.c, join.c: New programs. * cut.c (set_field): Allow blanks as well as commas to separate numbers in ranges. Sun Aug 25 19:57:40 1991 Jim Meyering (meyering at apple-gunkies) * pr.c: Failure to open an input file is no longer a fatal error. A message is printed for each failed open. When printing in parallel, each failed open results in one fewer output column. Added POSIX -r option to suppress the message. * pr.c: Added variables: failed_opens, ignore_failed_opens. These changes were based in part on work by David MacKenzie. Sat Aug 24 15:27:39 1991 Jim Meyering (meyering at pogo) * pr.c: Complain if user gives both -m and -[0-9]+ options. Wed Aug 21 22:04:57 1991 David J. MacKenzie (djm at apple-gunkies) * Version 1.0. Mon Aug 19 00:16:51 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * expand.c: Rename some variables. (expand): Don't access value off end of array. * unexpand.c: Rename some variables. (unexpand): Don't access value off end of array. Instead of copying tabs verbatim and flushing pending spaces when one is reached, count them as the proper number of pending spaces. Instead of changing tabs to single spaces if the tabstop list is exhausted, print the rest of the line unchanged (for POSIX). Sat Aug 17 01:49:41 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * cut.c (cut_file), paste.c (paste_parallel, paste_serial): Clear EOF and error conditions on stdin so it can be reused. * expand.c, unexpand.c (parse_tabstops): Allow blanks as well as commas to separate tabstops, for POSIX. * expand.c (expand), unexpand.c (unexpand): Don't line-buffer the output; send it directly to stdout. * unexpand.c (main): Make -t stupidly imply -a for POSIX. (unexpand): If a tab stop list was given and we move past its end, copy the rest of the line verbatim. * split.c (convint): New function to allow 'm' and 'k' after byte counts. (main): Use it. Change -c option to -b for POSIX. Fri Aug 9 02:47:02 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * pr.c: Protect isdigit with isascii, if required. Tue Aug 6 21:42:25 1991 David J. MacKenzie (djm at wheat-chex) Most of the following is from Paul Eggert: * cat.c (main): If stdin is read, check close at end. * cmp.c (main): Check for stdin being closed. Check for close errors on stdin and stdout. (cmp): Return a value instead of exiting. * cut.c (cut_file): New function, from code in main. Check for read errors. (main): Check for close errors. * expand.c, unexpand.c (main): Check for close errors. (next_file): Go to next file if one can't be opened. Check for close errors. * head.c (main), tail.c (main): If stdin was read, check for close errors. * head.c (head_file), tail.c (tail_file): Check for close errors. * paste.c (main, paste_parallel, paste_serial), tac.c (main): Check for close errors. Close stdin if it was read. * split.c (main): Check for close errors. * configure, Makefile.in's: Support +srcdir option. Make config.status. Fix up clean targets. Wed Jul 31 01:32:59 1991 David J. MacKenzie (djm at hal) * linebuffer.h (struct linebuffer): Add a field to indicate the number of valid chars in the line. * linebuffer.c (initbuffer, readline): Fill it in. * uniq.c, comm.c: Use it. * pr.c (main): Check stdin and stdout fclose for errors. (init_parameters): If there's no room for header and footer, omit them rather than dying (for POSIX). (init_header): Take a file descriptor as additional arg. (init_fps): Change callers. Note when stdin is read. (open_file): For filename "-" use stdin. (close_file): Don't close stdin. Check close for errors. (print_char, char_to_clump): Use isprint instead of explicit comparisons. * memcmp.c: New file (needed for comm). * bcopy.c: New file (needed for fold). * system.h: Don't define bcopy as memcpy. * configure: Check for bcopy and memcmp. * uniq.c (main): Use "-" instead of NULL to mean stdin or stdout. (check_file): Use "-" instead of NULL to mean stdin or stdout. Check readline return instead of for NUL character to detect eof. Check fclose for errors. (find_field): Use linebuffer length, not NULs, to detect end of line. (different): New function, replaces compare. Uses memcmp instead of strncmp. (writeline): Use fwrite instead of fputs so NULs are preserved. * comm.c (compare_files): Return an error indication. Don't take a filename of NULL to mean stdin. Use memcmp instead of strcmp to allow for NULs. Check fclose for errors. (writeline): Use fwrite instead of fputs so NULs are preserved. * sum.c (sum_file): Take an arg indicating whether to print the filename, and don't take NULL meaning stdin. Set a flag when we read stdin. Check fclose return for errors. (main): If stdin was read, check fclose return for errors. Use filename of "-" if no args given. Thu Jul 25 15:17:10 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * fold.c: Rewritten from scratch for POSIX. Wed Jul 24 01:55:41 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * split.c (line_bytes_split): Use xmalloc instead of alloca. * system.h: Don't declare alloca. * tac.c, tail.c: Use SEEK_ instead of L_ for lseek. * system.h: Define SEEK_ macros if not defined. * pr.c: Rename variable `truncate' to avoid library function conflict. Tue Jul 23 13:21:48 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * linebuffer.c, linebuffer.h: New files. * comm.c, uniq.c (initbuffer, readline): Functions removed (use versions in linebuffer.c). Mon Jul 22 13:23:53 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * sum.c (sumfile): Always print 5 digits for second number, too. Interpret "-" to mean stdin. Sat Jul 20 14:24:40 1991 David J. MacKenzie (djm at bleen) * uniq.c: Use isblank instead of isspace, to support POSIX.2. * comm.c, pr.c, uniq.c (concat, fatal, error, pfatal_with_name, xmalloc, xrealloc): Functions removed. Sat Jul 13 02:04:53 1991 David J. MacKenzie (djm at geech.gnu.ai.mit.edu) * nl.c: Add long-named options. Doc fixes. Sat Jul 6 02:19:09 1991 David J. MacKenzie (djm at geech.gnu.ai.mit.edu) * expand.c, unexpand.c [STDC_HEADERS]: Include stdlib.h. * xwrite.c [POSIX]: Include unistd.h. [STDC_HEADERS]: Don't declare errno. Sun Jun 30 23:35:16 1991 David J. MacKenzie (djm at geech.gnu.ai.mit.edu) * uniq.c: Add long-named options. Remove marginally useful -z option (zero padded repeat counts). Thu Jun 27 16:31:45 1991 David J. MacKenzie (djm at geech.gnu.ai.mit.edu) * tail.c (tail_file), tac.c (save_stdin, tac_file), split.c (cwrite), head.c (head_file), cat.c (main): Check close return value for delayed error report due to NFS. Tue Jun 11 00:12:15 1991 David J. MacKenzie (djm at geech.gnu.ai.mit.edu) * cat.c: Replace "uchar" with "unsigned char", to avoid problems with various systems' typedefs. Thu Jun 6 12:54:26 1991 David J. MacKenzie (djm at geech.gnu.ai.mit.edu) * cat.c (cat): Interpret ENOTTY return from FIONREAD ioctl to mean operation is unsupported, for HP-UX 7.0. Sun Apr 14 21:49:17 1991 Richard Stallman (rms at mole.gnu.ai.mit.edu) * sum.c: Always print five digits for first number. Fri Mar 15 16:16:54 1991 David J. MacKenzie (djm at geech.ai.mit.edu) * cat.c, cmp.c: Don't use fileno(); not needed. Thu Jan 10 02:16:55 1991 David J. MacKenzie (djm at albert.ai.mit.edu) * tac.c, tail.c: Change _POSIX_SOURCE to POSIX. Thu Dec 27 00:06:45 1990 David J. MacKenzie (djm at egypt) * cut.c (cut_file_bytes, cut_file_fields): Make inbufp and outbufp global. (enlarge_line): Adjust inbufp and outbufp. Sun Sep 9 16:54:19 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * cat.c: Declare free returning void, not int, so it doesn't bomb on Xenix. Mon Sep 3 22:23:57 1990 David J. MacKenzie (djm at coke) * tac.c: Print error messages before calling cleanup, not after. Tue Aug 28 18:05:24 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * tac.c (cleanup): Return SIGTYPE, not int. Tue Aug 7 12:51:18 1990 David J. MacKenzie (djm at apple-gunkies) * cut.c (main, usage): Add -b and -n options for POSIX. (set_fields): Don't allow SPC or TAB as number separators. * paste.c (paste_parallel): If open of any file fails, quit (for POSIX). Mon Aug 6 22:14:13 1990 David J. MacKenzie (djm at apple-gunkies) * pr.c: Add POSIX -F option (same as -f). * uniq.c (check_file): Allow '-' to mean stdin or stdout. Mon Aug 6 14:43:30 1990 David J. MacKenzie (djm at pogo.ai.mit.edu) * head.c, tail.c: Change `chars' to `bytes' globally. (main, usage): Use POSIX.2 draft 10 option syntax. Sun Aug 5 11:51:12 1990 David J. MacKenzie (djm at pogo.ai.mit.edu) * cat.c (main): Don't delay error messages, so they appear where expected. (main, simple_cat, cat): Make errors in input files nonfatal. Sat Aug 4 10:11:30 1990 David J. MacKenzie (djm at pogo.ai.mit.edu) * cat.c: Remove -c option added for POSIX draft 9, since POSIX draft 10 removed it. * tac.c (tac_stdin): Use fstat instead of lseek to determine whether stdin is seekable, because lseek silently fails on some special files, like tty's. tail.c (tail_chars, tail_lines): Use fstat instead of lseek; don't turn off -f for non-regular files (assume the user knows what he's doing; it might work for fifo's and sockets). * paste.c (main): If no files given, use stdin. Don't let collapse_escapes write on string constant (delim default). (paste_parallel): Don't close stdin. * cut.c (main): Use standard input for filename of "-". * comm.c (compare_files): Allow '-' to mean stdin. Fri Aug 3 13:38:28 1990 David J. MacKenzie (djm at pogo.ai.mit.edu) * cut.c (enlarge_line): Take an arg giving the required amount of space. Change callers. (main): Don't allow -t'' without -f. Make `delim' unsigned to fix sign extension problem in comparison. Tue Jul 17 12:36:11 EDT 1990 Jay Fenlason (hack@ai.mit.edu) * pr.c Deleted excess whitespace from ends of lines. Modified to work with current version of getopt, which returns 1 instead of 0 for non-options. Reversed the meaning of the -f option, to be compatable with real pr. Sun Jul 8 00:39:31 1990 David J. MacKenzie (djm at apple-gunkies) * cmp.c (main, usage): Rename -L option to -c and don't have it imply -l. (printc): Take an arg to specify number of chars to pad to, for column alignment. (cmp): Respect flag_print_chars in default output format. Align columns for cmp -cl. Sat Jul 7 17:23:30 1990 David J. MacKenzie (djm at apple-gunkies) * cmp.c: For +show-chars, have getopt return 'L' so `flag_print_chars' gets set. Fri Jun 29 01:04:19 1990 David J. MacKenzie (djm at apple-gunkies) * tac.c (main): Initialize fastmap and translate fields of regex before compiling it. Fri Jun 22 00:38:20 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * tac.c: Change +regexp to +regex for consistency with GNU find. Wed Jun 20 01:46:09 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * cat.c (cat): If FIONREAD is available, only use it if it is supported by the filesystem that the file is on. Sun Jun 3 20:26:19 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * cat.c (main): Add a variable to control whether the check for input file == output file is made, because no values of st_dev and st_ino should be assumed to be available for this purpose. Only do the check for regular files. * tac.c: Use bcopy instead of memcpy. Thu May 31 00:55:36 1990 David J. MacKenzie (djm at apple-gunkies) * head.c: Use longs instead of ints for file offsets, for 16 bit machines. Tue May 22 00:56:51 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * cmp.c: Change some ints to longs for 16 bit machines. (bcmp_cnt): Make char-pointer counting slightly simpler. Sat May 12 01:16:42 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * cat.c (main): Allow input file to be output file for devices (ttys, etc.). Check whether input file is output file when reading standard input. Print any error messages for standard input. * cmp.c (bcmp_cnt): Handle int comparisons correctly on 16 bit machines as well as 32 bit ones. * cmp.c, tail.c: Use longs instead of ints for file offsets. Fri May 11 02:11:03 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * cmp.c: Fix some exit statuses for POSIX. Tue May 8 03:41:42 1990 David J. MacKenzie (djm at abyss) * tac.c: Use regular expressions as the record boundaries. Give better error messages. Reformat code and make it more readable. (main): Use getopt_long to parse options. (tac_stdin): Do not make a temporary file if standard input is a file. (tac_file): New function. (tac): Take an open file desc as an arg. (output): Rewrite to use its own efficient buffering. (xmalloc, xrealloc, xwrite): New functions. Sun Apr 8 20:33:20 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * head.c, tail.c: Use `error' instead of `fatal_perror' and `nonfatal_perror'. Remove some unnecessary info from messages. Wed Mar 21 09:30:18 1990 David J. MacKenzie (djm at pogo.ai.mit.edu) * comm.c (main): Pass the list of files to compare_files as a char ** instead of a char *. (compare_files): Make arg a char **. * uniq.c: Declare some functions as void. Change global vars `countmode' and `mode' from ints to enums. (main): Use getopt to parse options and support POSIX options. Don't use integer_arg to parse numbers, since `-#' can't be parsed that way using getopt. (find_field): Use isspace for finding fields boundaries. Tue Mar 20 14:28:25 1990 David J. MacKenzie (djm at pogo.ai.mit.edu) * comm.c (main): Call usage if given bad option or wrong number of args. Exit with 0 status normally. (usage): New function. Declare some other functions as void. Wed Mar 14 10:48:40 1990 David J. MacKenzie (djm at rice-chex) * cmp.c (main, cmp, usage): Replace -q +quick option with -L +show-chars option to add ASCII representation of bytes to -l format. Tue Mar 13 00:50:14 1990 David J. MacKenzie (djm at rice-chex) * cmp.c (cmp): Change EOF message for POSIX compatibility. For -l format, clear bits > FF. Mon Mar 5 17:21:00 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * tail.c: Move global `errors' into main instead of having nonfatal_perror set it. (tail, tail_chars, tail_file, tail_lines, pipe_chars, pipe_lines): Return an error status. (file_lines, start_chars, start_lines): Reverse the meaning of the return value. (tail_lines, tail_chars): Account for that reversal. Mon Mar 5 00:34:36 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * head.c: Move global `errors' into main and have the various functions return an error status instead of setting it in nonfatal_perror. Sat Mar 3 11:27:27 1990 Torbj|rn Granlund (tege at echnaton) * cmp.c (cmp): Call function bcmp_cnt for flag == 0 (i.e. no options specified), to compare the two blocks and count newlines simultaneously. * cmp.c New function: bcmp_cnt. * cmp.c (main): Test if output is redirected to /dev/null, and assume `-s' if this is so. Tue Feb 20 17:09:19 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * cat.c: Change `argbad' from a char to a short, so it will work on machines with unsigned chars. Sat Feb 10 02:16:40 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * cmp.c (cmp): Rename `r' to `first_diff', and `x' to `smaller'. Remove unneccessary variable `c1'. If -l was given, increase `char_number' by the number of bytes read, after producing output, rather than by the offset of the first differing bytes, before producing output. Replace if-else-if constructions with case statements for clarity. (bcmp2): Rename `n' to `nread'. Wed Dec 20 01:32:06 1989 David J. MacKenzie (djm at hobbes.ai.mit.edu) * nl.c (proc_text): Use re_search instead of re_match. Tue Dec 19 01:26:34 1989 David J. MacKenzie (djm at hobbes.ai.mit.edu) * nl.c: Indent. Un-nest statements. Use GNU regexp functions instead of System V ones. Move function declarations together. (quit): Remove useless function. (program_name): New variable for error messages. (main): Use perror in error message. (xmalloc): New function to replace myalloc. (myalloc): Function removed. Global: use program_name and xmalloc. Sun Dec 17 00:36:36 1989 David J. MacKenzie (djm at hobbes.ai.mit.edu) * uniq.c: Declare some functions. (main): Initialize infile and outfile. Call usage if given invalid args. Normally exit with 0 status instead of garbage. (usage): New function to print usage message and exit. (check_file): Remove unused variable. (readline): Compare against EOF, not < 0. (xmalloc, xrealloc): Return char *, not int. Ok to return 0 if 0 bytes requested. (lb1, lb2): Remove unused global vars. (concat): Remove unused function. Sat Dec 16 15:15:50 1989 David J. MacKenzie (djm at hobbes.ai.mit.edu) * comm.c: Remove unused global variables lb1, lb2. (main): Remove unneeded variable. (compare_files): Remove unused arg. (readline): un-nest assignment. Test against EOF instead of < 0. (error): Print to stderr, not stdout. (xmalloc, xrealloc): Return char * instead of int. Returning 0 is ok if 0 bytes requested (ANSI C). Local Variables: version-control: never End: diff --git a/contrib/gnu-sort/NEWS b/contrib/gnu-sort/NEWS index fab67541e476..a1722997d637 100644 --- a/contrib/gnu-sort/NEWS +++ b/contrib/gnu-sort/NEWS @@ -1,452 +1,457 @@ Changes in release 2.1 +[2.0.22] +* `od -t f8' works once again [bug introduced in textutils-2.0.8] +* various portability fixes, and general clean-up +* various minor, corner-case bug fixes [2.0.21] * split accepts new option -a or --suffix-length. * split no longer generates longer suffixes than requested; instead, it reports an error when suffixes are exhausted. POSIX requires this behavior. * The _POSIX2_VERSION environment variable lets you select which version of POSIX the utilities should conform to. Its default value is system dependent. Set _POSIX2_VERSION=199209 to cause the utilities to support obsolete usage like "sort +1". * The following obsolete usages are no longer supported when conforming to POSIX 1003.1-2001, which prohibits most digit-string options: expand -N (instead, use expand -t N) head -N (instead, use head -c N or head -n N) fold -N (instead, use fold -w N) split -N (instead, use split -l N) tail -N (instead, use tail -c N or tail -n N) unexpand -N (instead, use unexpand --first-only -t N) uniq -N (instead, use uniq -f N) The following obsolete usages (options without arguments) are no longer supported when conforming to POSIX 1003.1-2001, which prohibits most options with optional arguments: od -s (instead, use od --strings) od -w (instead, use od --width) pr -S (instead, use pr --sep-string) [2.0.20] * tr no longer gets failed a assertion for [==] or [::] * The following obsolete usages are no longer supported when conforming to POSIX 1003.1-2001, which prohibits most options with leading "+": sort +POS1 -POS2 (instead, use sort -k) tail +N (instead, use tail -c +N or tail -n +N) uniq +N (instead, use uniq -s N) * Warnings are issued for obsolete usages on older hosts, unless POSIXLY_CORRECT is set in the environment. * sort -m no longer segfaults when given an empty file * sort -S now accepts 'K' as a synonym for 'k'. * wc recognizes all locale-defined white-space characters, not just those in the "C" locale. [2.0.19] * portability tweak to make lib/regex.c compile * split translatable strings only in the middle of sentences [2.0.18] * sort could segfault on systems without a working mkstemp function and with a gettimeofday function that clobbers the static buffer that localtime uses for it's return value -- introduced in 2.0.17 [2.0.17] * csplit no longer gets a failed assertion for this: printf 'a\n\n'|csplit - '/^$/' 2 * sort detects physical memory attributes more portably * tail no longer gets a segfault on Linux's /proc/ksyms * sum -s produces the proper 16-bit checksum for large files (this fixes a bug that was introduced in 2.0f) * uniq is now about 3 times faster than the version from 2.0 on Linux systems; the code uses lock-avoiding variants of common I/O functions [2.0.16] * tail -F no longer segfaults [2.0.15] * `head -c N' and `od -N N' now read no more than N bytes of input * tail accepts new option: -F, equivalent to `--follow=name --retry', for compatibility with the FreeBSD and NetBSD versions of tail. * fmt no longer segfaults when using a maximum line width larger than 32767 * uniq's --all-repeated option has new modes to delimit groups of duplicate lines: --all-repeated={precede,separate,none(default)} [2.0.14] * sort now accepts long options like "--reverse" and "--". * sort now checks option syntax as POSIX requires, except that (as usual for GNU) options can follow file names unless POSIXLY_CORRECT is set. For example, invalid positional combinations like "sort +1 -r -2" are now rejected as per POSIX. * The next POSIX standard will require that obsolescent 'sort' positional options like +1 be treated as file names, not options. Please use 'sort -k' instead. [2.0.13] * pr accepts new -D or --date option, to specify date format. * The following changes are required by POSIX: - If POSIXLY_CORRECT is set, dates in pr headers now look something like 'Dec 4 23:59 2001', with the exact appearance affected by LC_TIME. - pr -h now affects only the center header string, not the entire header. - pr no longer truncates headers. * Spacing in pr headers has been adjusted slightly. * `fmt --prefix=S' now works when S contains a byte with the high bit set [2.0.12] * sort has improved performance when using very little main memory * sort has improved memory management * sort is no longer susceptible to certain denial of service attacks * sort no longer suffers from a race condition whereby an interrupt received during cleanup could cause it to fail to remove temporary files. This problem could arise only on hosts without sigaction. [2.0.11] * sort accepts new -S SIZE option, to specify main-memory usage. [2.0.10] * od is faster and more portable than it was in 2.0.9 * tail avoids an uninitialized memory reference [2.0.9] * od now prints valid addresses for offsets of 2^32 and larger, and allows the byte offset (-j) and byte count (-N) arguments to be 2^32 and larger. * tail now works with line and byte counts of 2^32 and larger, on systems with large file support * join now works with an 8-bit delimiter * fix a compilation failure on some Solaris systems with wc.c [2.0.8] * od now supports 8-byte integers, assuming they're printable with e.g., %lld * new program: sha1sum * wc accepts new -m option: count (potentially multi-byte) characters * wc's `--chars' option is now equivalent to -m, not --bytes as it used to be * `cat -n' works properly when processing 2^31 or more lines [2.0g] * sort's --help output now warns that it is locale-aware * tail: fix a buffer underrun error that occurred on an empty pipe, also thanks to bounded pointers * pr: fix a bounds violation found by Greg McGary's bounded-pointers-enabled gcc It could have caused (with low probability) the columns on the last page of output *not* to be `balanced' when they should have been. * sort: if the -T tmpdir option is given multiple times, all the given directories are used; this can improve performance for huge sort/merges. [2.0f] * all programs fail when printing --help or --version output to a full device * cut no longer gets a segfault under some circumstances * unexpand accepts new option: --first-only [2.0e] * `tail -f directory' no longer gets a failed assertion * sort: big performance improvement when sorting many small files; from Charles Randall * configure and portability changes in m4/ and lib/ [2.0d] * preliminary sort performance improvements * tsort now works more like the traditional UNIX tsort. Before it would exit when it found a loop. Now it continues and outputs all items. * unexpand no longer infloops on certain sequences of white space * unified lib/: now that directory and most of the configuration framework is common between fileutils, textutils, and sh-utils [2.0c] * include lib/nanosleep.h. [2.0b] * portability tweaks for error.c vs. systems with deficient strerror_r [2.0a] * `tail --follow=name' no longer gets a failed assertion for a dev,inode-reusing race condition * sort and comm no longer consider newlines to be part of the line, as this requirement will likely be removed from POSIX.2. This undoes some changes made for textutils 1.22m and 1.22n. * tail's (short only) -f option no longer accepts an optional argument, so e.g., `tail -fn 2 file' works again. * tail no longer refuses to operate on certain types of files * fixed bug in tsort's handling of cycles + Changes in release 2.0 [1.22q] * HPUX portability fix: md5sum would dump core due to use of libc's getline [1.22p] * portability fixes from Paul Eggert based largely on tar-1.13 reports * `tail --pid=PID' now works even when PID belongs to some other user [1.22o] * tail accepts new option: --pid=PID [1.22n] * tail accepts the following new options (some of which were added in 1.22g): --retry --follow[={name|descriptor}] --max-unchanged-stats=N --max-consecutive-size-changes=N --sleep-interval=S * wc uses the POSIX-mandated output format when POSIXLY_CORRECT is set * To maintain compatibility with sort, comm and join now obey the LC_COLLATE locale, and comm now considers newlines to be part of the lines. * use lib/memchr.c only if it's not provided by the system -- this means that on systems with a fast library memchr function you may notice an improvement. If you use a system with a buggy or signifcantly slower memchr, please report it. [1.22m] * sort now considers newlines to be part of the line, as required by POSIX.2. E.g. a line starting with a tab now sorts before an empty line, since tab precedes newline in the ASCII collating sequence. * sort handles NUL bytes correctly when configured/compiled with --enable-nls * fix typos in my version of AC_SEARCH_LIBS. * fix dates on config files so builders don't need autoconf/automake [1.22l] * sort no longer autodetects the locale of numbers and months, as that conflicts with POSIX.2 * `join -tC' now works when input contains trailing spaces * portability tweaks for Irix's cc [1.22k] * `sort -n' works with negative numbers when configured/compiled with --enable-nls * head accepts byte and line counts of type uintmax_t (so up to 2^64 - 1) [1.22j] * tail: fix bug introduced in 1.22i [1.22i] * tail now terminates in `yes > k & sleep 1; tail -2c k' * `tail -f' now ensures that stdout is unbuffered * fix a bug in cut to allow use of 8-bit delimiters * pr accepts POSIX compliant options -s and -w, the new capital letter options -J, -S and _W turn off the unexpected interferences of the small letter options -s and -w if used together with the column options. * pr output has been adapted to other UNIXes in some cases. [1.22h] * portability tweaks * Window/NT/DOS support [1.22g] * uniq accepts new option: --all-repeated (-D). * Windows/DOS portability fixes * new program: tsort * tail has several new options * md5sum can handle file names with embedded backslash characters * pr accepts long option names (see `pr --help') * new program: ptx (moved to this package from being its own distribution) [1.22f] * cut accepts new --output-delimiter=STR option * `sort -o no-such-file no-such-file' now fails, as it should * fix pr bug: pr -td didn't double space * fix tac bug when using -b, -r, and -s SEPARATOR * fix sort bug whereby using key-local `d' option would cause following key specs to be ignored when any two keys (in the `d'-modified test) compared equal. [1.22e] * remove maintainer mode [1.22d] * wc accepts new option: --max-line-length (-L) * sort can sort according to your locale if your C library supports that [1.22c] [1.22b] * od supports a new trailing `z' character in a type specification: $ od -tx1z . 0000000 be ef c6 0f fd f9 d7 e0 ec cb f3 c6 00 db e8 00 >................< 0000020 00 00 d2 00 00 00 00 00 00 00 00 00 00 00 00 00 >................< 0000040 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 >................< * 0000600 00 00 00 00 00 00 00 00 00 00 00 00 00 00 35 cc >..............5.< 0000620 05 63 76 74 2e 6f 00 00 29 ac 08 70 72 6f 6a 65 >.cvt.o..)..proje< 0000640 63 74 73 00 00 00 18 9a 05 63 76 74 2e 63 00 00 >cts......cvt.c..< 0000660 18 d9 03 52 43 53 00 00 18 c0 05 78 2e 64 61 74 >...RCS.....x.dat< [1.22a] * sort -c reports both the number and the contents of the first out-of-order line, in addition to the file name. * `head -c 4096m' is no longer treated just like `head -c 0' now it gets a diagnostic about 4096m being too large. * pr: For compatibility (also more POSIX compliant): Include default separator `TAB' when merging lines of full length. * When POSIXLY_CORRECT is not set, tail -N now accepts more than one file argument, to be consistent with the way head -N works. If POSIXLY_CORRECT is set, using two or more file arguments with the obsolescent form (-N) evokes an error. To avoid the warning or failure, use the POSIX -n N option or the GNU --lines=N option. Changes in release 1.22 [1.21a] * Fix a bug in tail when invoked with an argument like `+NUMBERc' * Add test suite for tail Changes in release 1.21 * Using --program-prefix no longer applies the prefix twice Changes in release 1.20 * fix pr: -l now uses total number of lines per page also with -f * fix pr: use left-hand-side truncation of header string to avoid line overflow * fix pr: it now accepts `form feeds set in input files', also with -m and multiple form feeds at different pages in each file * pr now accepts: -h "", print a blank line header * pr: when skipping pages (+FIRST_PAGE option) line counting (-n option) starts with 1st line of input file (not of 1st page printed) by default * pr accepts new option: -N, start printing with an optional line number * pr -t retains `form feeds set in input files' (`don't destroy page layout') * pr accepts new option: -T, equivalent to -t, but eliminate also form feeds (`clear file') * pr accepts the extension: +FIRST_PAGE[:LAST_PAGE] * pr -w and -s option disentangled (`use a separator' no longer destroys column alignment) * pr accepts new option: -j, merge lines of full length * pr accepts the extension: -s[STRING], use separator string instead of character only * pr -b is no longer an independent option, balancing is always used with -COLUMN (a requirement of unrestricted use of form feeds) * pr accepts new option: --test, to run the pr tests with a constant header string * join passes all of its tests on Alpha OSF 4.0. * sort no longer improperly ignores blanks in determining starting and ending positions for keys with explicit character offsets * fix bug in csplit with regexp and negative offset that led to infinite loop Changes in test release 1.19q * fix bug in sort -c that sometimes resulted in a segfault Changes in test release 1.19p * md5sum's --string option is being deprecated and is no longer documented. It is still accepted, but will be removed altogether in 1.22. * tr '[:lower:]' '[:upper:]' no longer fails when LC_CTYPE is set to iso_8859_1 on Solaris -- or any other character set with differing numbers of uppercase and lowercase characters * split and tail diagnose unrecognized multiplier suffixes, in e.g., `split --bytes=1M' (should be `-b 1m' or `--bytes=1m') * fix bug in md5sum's handling of partial reads * fix bug in treatment by sort -f of bytes with high-bit set * update configuration system to use automake's aclocal program * configure performs sanity check on CC and CFLAGS to avoid a misleading failure that suggested cross-compiling was the cause * distribute test suites for cut, join, sort, and tr * unexpand no longer gets in endless loop * when verifying checksums, md5sum uses the binary mode flag from the input stream rather than the one from the command line Changes in release 1.19 * md5sum can verify digests of files with names containing newline characters * update from gettext-0.10.20. Changes in release 1.18 * when building sort, link with -lm on systems that use the replacement strtod * update from gettext-0.10.17. Changes in release 1.17 * include texinfo.tex in the distribution Changes in release 1.16 * sort is compatible with Unix sort when a key-end spec refers to the N'th character in a field that has fewer than N characters * tail with old-style options like -20k and +31m operates on units of bytes, as the --help usage message says. Before, it used units of lines. Changes in release 1.15 * od gives better diagnostics for invalid format specs * uses automake-generated Makefile templates * configure takes a new option: --enable-maintainer-mode * fix a bug in fmt when prefix has trailing white space * internationalized diagnostic messages * fix a couple bugs in tr involving use of -c and/or -d flags -- see ChangeLog * diagnose some improper or questionable invocations of csplit * properly handle `echo |csplit - 1 1', rather than aborting * fix join: without -t it now ignores leading blanks * sort accepts new option: -z for NUL terminated records * join accepts new option: --ignore-case, -i * uniq accepts new option: --ignore-case, -i User-visible changes in release 1.14 * sort -i and sort -d properly order strings containing ignored characters * nl: rename misleading --first-page=N option to --starting-line-number=N. * sort diagnoses invalid arguments to -k, then fails * sort -n properly orders invalid integers with respect to valid integers * sorting works with character offsets larger than corresponding field width * sort's -b option and `b' modifier work * sort -k2,2 works. * csplit detects integer overflow when converting command line arguments * sort accepts new option/flag, -g, for sorting numbers in scientific notation * join accepts POSIX `-o 0' field specifier. * tr 'a[b*512]' '[a*]' < /dev/null terminates * tr '[:*3][:digit:]' 'a-m' and tr 'a[=*2][=c=]' 'xyyz' no longer fail * special characters in tr's string1 and string2 may be escaped with backslash User-visible changes in release 1.13 * md5sum: with --check, distinguish between open/read failure and bad checksum * md5sum: remove -h, -s, -v short options * md5sum: rename --verbose to --warn, --quiet to --status * md5sum --check fails if it finds no properly formatted checksum lines * sort -c prints `disorder on...' message on standard error, not stdout * sort -k works as described in the texinfo documentation * tail works on NetBSD * md5sum reads and writes (de facto) standard Plumb/Lankester format * sort accepts -.1 +.2 options for compatibility * od works properly when dump limit is specified and is a multiple of bytes_per_block (set by --width, 16 by default). User-visible changes in release 1.12 * sort no longer reports spurious errors on Ultrix systems * new program: md5sum * all --help messages have been improved * join's -a1 and -a2 options work * tr '[:upper:]' '[:lower:]' no longer reads uninitialized memory * sort properly handles command line arguments like `+7.2n' * fmt properly formats paragraphs not terminated by a newline * tail -f flushes stdout before sleeping so that it will output partial lines sooner * sort properly orders fields where one field is a proper prefix of the other * sort properly interprets field offsets specified via the -k option * dd, od, and tail work on systems for which off_t is long long (e.g. BSD4.4) * wc is faster when not counting words * wc now works even when file pointer isn't at beginning of file * expand no longer seg faults with very long tab lists User-visible changes in release 1.11 * fmt is built User-visible changes in release 1.10 * skeletal texinfo documentation (mainly just the `invoking' nodes) * new program: fmt * tail -f on multiple files reports file truncation * tail -q has been fixed so it never prints headers * wc -c is much faster when operating on non-regular files * unexpand gives a diagnostic (rather than a segfault) when given a name of a nonexistent file. * cat, csplit, head, split, sum, tac, tail, tr, and wc no longer fail gratuitously when continued after a suspended read or write system call. * cut interprets -d '' to mean `use the NUL byte as the delimiter' rather than reporting that no delimiter was specified and failing. * `echo a:b:c: | cut -d: -f3,4' prints `c:'. Before it printed just `c'. * cut has been rewritten, is markedly faster for large inputs, and passes a fairly large test suite. * sort properly handles the argument to the -T option. Major changes in release 1.9.1: * cut no longer ignores the last line of input when that line lacks a trailing newline character Major changes in release 1.9: * `echo a:b:c: | cut -d: -f3-' prints `c:' and `echo a:b | cut -d: -f1' prints `a'. * the command `printf '\t\n' |fold -w n' now terminates. Before, it wouldn't stop for n less than 8. * sort accepts and ignores -y[string] options for compatibilty with Solaris. * cat -v /dev/null works on more systems * od's --compatible (-C) flag renamed to --traditional (no short option) * --help and --version exit successfully * --help gives a one-line description of each option and shows the correspondence between short and long-named options. * fix bug in cut. Now `echo 'a:b:c:' | cut -d: -f3-' works. Before it printed `c' instead of `c:' * csplit allows repeat counts to be specified via `{*}'. * csplit accepts a new option, --suffix=format that supercedes the --digits option. The --digits option will continue to work. * csplit accepts a new option, --elide-empty-files. * configure uses config.h, so DEFS won't exceed preprocessor limits of some compilers on the number of symbols defined via -D. * work around problem where $(srcdir)/config.h was used instead of ../config.h -- this happened only when building in a subdirectory and when config.h remained in $(srcdir) from a previous ./configure. Major changes in release 1.8: * added non-ANSIfied version of memchr.c from GNU libc. Major changes in release 1.7: * none Major changes in release 1.6: * with the --version option programs print the version and exit immediately * pr -2a really terminates * pr -n produces multi-column output Major changes in release 1.5: * sort is 8-bit clean * sort's -n and -M options no longer imply -b * several bugs in sort have been fixed * all programs accept --help and --version options * od --compatible accepts pre-POSIX arguments * pr -2a terminates Major changes in release 1.4: * add od and cksum programs * move cmp to GNU diff distribution * tail -f works for multiple files * pr prints the file name in error messages * fix some off by 1 errors in pr and fold * optimize wc -c on regular files * sort handles `-' argument correctly * sort supports -T option * tr ranges like a-a work * tr x '' fails gracefully * default sum output format is BSD compatible * paste -d '' works diff --git a/contrib/gnu-sort/README b/contrib/gnu-sort/README index 48e7bec14904..e2fb18e34228 100644 --- a/contrib/gnu-sort/README +++ b/contrib/gnu-sort/README @@ -1,41 +1,41 @@ These are the GNU text file (actually, file contents) processing utilities. Most of these programs have significant advantages over their Unix counterparts, such as greater speed, additional options, and fewer arbitrary limits. The programs that can be built with this package are: cat, cksum, comm, -csplit, cut, expand, fmt, fold, head, join, md5sum, nl, od, paste, pr, -ptx, sort, split, sum, tac, tail, tr, tsort, unexpand, uniq, and wc. +csplit, cut, expand, fmt, fold, head, join, md5sum, nl, od, paste, pr, ptx +sha1sum, sort, split, sum, tac, tail, tr, tsort, unexpand, uniq, and wc. See the file NEWS for a list of major changes in the current release. See the file INSTALL for compilation and installation instructions. I M P O R T A N T: Some Cray C compilers lack support for a 32-bit arithmetic type. Since gettext requires such a type, you should use the --disable-nls option when invoking configure on such systems. The textutils are intended to be POSIX.2 compliant (with BSD and other extensions), like the rest of the GNU system. They are almost there, but a few incompatibilities remain. Note that the man pages are now automatically generated from templates and from the --help usage message. Patches to the template files (man/*.x) are welcome. However, the authoritative documentation is in texinfo form in the doc directory. These programs all recognize the `--version' option. When reporting bugs, please include in the subject line both the package name/version and the name of the program for which you found a problem. For general documentation on the coding and usage standards this distribution follows, see the GNU standards document http://www.gnu.org/prep/standards.html, and the documentation for automake and autoconf: http://www.gnu.org/software/autoconf/autoconf.html, http://www.gnu.org/software/automake/automake.html. Mail suggestions and bug reports for these programs to bug-textutils@gnu.org. diff --git a/contrib/gnu-sort/THANKS b/contrib/gnu-sort/THANKS index 08e018628205..25dfff890a04 100644 --- a/contrib/gnu-sort/THANKS +++ b/contrib/gnu-sort/THANKS @@ -1,340 +1,355 @@ These people have contributed to the GNU fileutils, textutils, and/or sh-utils packages. Some have reported problems, others have contributed improvements to the documentation, actual code, and even complete programs. Those contributions are described in the ChangeLog files. If your name has been left out, if you'd rather not be listed, or if you'd prefer a different address be used, please let me know. Some old names and addresses are still in the ChangeLog. Achim Blumensath blume@corona.oche.de Adam Klein aklein@debian.org Akim Demaille demaille@inf.enst.fr Alain Magloire alain@qnx.com Alan Iwi iwi@atm.ox.ac.uk +Albert Chin-A-Young china@thewrittenword.com Albert Hopkins ahopkins@dynacare.com Alberto Accomazzi alberto@cfa0.harvard.edu aldomel aldomel@ix.netcom.com Alen Muzinic zveki@fly.cc.fer.hr Alexandre Duret-Lutz duret_g@epita.fr Alexey Solovyov alekso@math.uu.se +Alfred M. Szmidt ams@kemisten.nu +Andi Kleen freitag@alancoxonachip.com Andre Novaes Cunha Andre.Cunha@br.global-one.net Andreas Gruenbacher ag@bestbits.at Andreas Jaeger jaeger@gnu.org Andreas Luik luik@isa.de Andreas Schwab schwab@suse.de Andreas Stolcke stolcke@ICSI.Berkeley.EDU Andres Soolo andres@soolo.matti.ee Andrew Burgess aab@cichlid.com Andrew Dalke dalke@bioreason.com Andrew Pham andpha@us.ibm.com Andrew Tridgell tridge@samba.org Andries Brouwer Andries.Brouwer@cwi.nl Andy Longton alongton@metamark.com Antonio Rendas ajrendas@yahoo.com Ariel Faigon ariel@cthulhu.engr.sgi.com Arne H. Juul arnej@solan.unit.no Arne Henrik Juul arnej@imf.unit.no Arthur Pool pool@commerce.uq.edu.au Austin Donnelly Austin.Donnelly@cl.cam.ac.uk Axel Kittenberger Anshil@gmx.net Bauke Jan Douma bjdouma@xs4all.nl Bengt Martensson bengt@mathematik.uni-Bremen.de Bernd Leibing bernd.leibing@rz.uni-ulm.de Bernhard Baehr bernhard.baehr@gmx.de Bernhard Rosenkraenzer bero@redhat.de Bill Peters peters@gaffel.as.arizona.edu Bjorn Helgaas helgaas@rsn.hp.com Bob McCracken kerouac@ravenet.com Bob Proulx rwp@fc.hp.com Branden Robinson branden@necrotic.deadbeast.net Brendan O'Dea bod@compusol.com.au Brian Kimball bfk@footbag.org Brian Youmans 3diff@gnu.org Bruno Haible haible@clisp.cons.org Carl Johnson carlj@cjlinux.home.org Carl Lowenstein cdl@mpl.UCSD.EDU Carlos Canau Carlos.Canau@relay.puug.pt Charles Karney karney@pppl.gov Charles Randall crandall@matchlogic.com Chip Salzenberg chip@valinux.com Chris Faylor cgf@cygnus.com Chris J. Bednar cjb@AdvancedDataSolutions.com Chris Sylvain csylvain@umm.edu Chris Yeo cyeo@biking.org Christi Alice Scarborough christi@chiark.greenend.org.uk Christian Harkort christian.harkort@web.de Christian Krackowizer ckrackowiz@std.schuler-ag.com Christian Rose menthos@menthos.com Christian von Roques roques@pond.sub.org Chuck Hedrick hedrick@klinzhai.rutgers.edu Clark Morgan cmorgan@aracnet.com Colin Plumb colin@nyx.net Colin Watson cjw44@riva.ucam.org Collin Rogowski collin@rogowski.de Dale Scheetz dwarf@polaris.net Dan Hagerty hag@gnu.ai.it.edu Dan Pascu dan@services.iiruc.ro Daniel Bergstrom noa@melody.se Darren Salt ds@youmustbejoking.demon.co.uk +Dave Beckett dajobe@dajobe.org David Dyck dcd@tc.fluke.COM David Godfrey dave@delta.demon.co.uk David Luyer david_luyer@pacific.net.au Dennis Henriksen opus@flamingo.osrl.dk Derek Clegg dclegg@next.com Dick Streefland dick_streefland@tasking.com Dirk Lattermann dlatt@t-online.de Dirk-Jan Faber djfaber@snow.nl Dan Jacobson http://www.geocities.com/jidani Don Parsons dparsons@synapse.kent.edu Donni Erpel donald@appc11.gsi.de Doug McLaren dougmc@comco.com Dragos Harabor dharabor@us.oracle.com Ed Avis epa98@doc.ic.ac.uk Edzer Pebesma Edzer.Pebesma@rivm.nl Eirik Fuller eirik@hackrat.com Eivind eivindt@multinet.no Eli Zaretskii eliz@is.elta.co.il Emile LeBlanc leblanc@math.toronto.edu Eric Backus ericb@lsid.hp.com Eric G. Miller egm2@jps.net Eric Pemente pemente@northpark.edu Eric S. Raymond esr@snark.thyrsus.com Erik Bennett bennett@cvo.oneworld.com Erik Corry erik@kroete2.freinet.de Felix Lee flee@teleport.com Fletcher Mattox fletcher@cs.utexas.edu Florin Iucha fiucha@hsys.mic.ro Frank Adler fadler@allesklar.de Frank T Lofaro ftlofaro@snooks.Egr.UNLV.EDU François Pinard pinard@iro.umontreal.ca Fred Fish fnf@ninemoons.com Frederik Eaton frederik@caltech.edu Frédéric L. W. Meunier 0@pervalidus.net Gabor Z. Papp gzp@gzp.org.hu Galen Hazelwood galenh@micron.net Gary Anderson ganderson@clark.net +Gaute Hvoslef Kvalnes gaute@verdsveven.com Gaël Quéri gqueri@mail.dotcom.fr Geoff Kuenning geoff@cs.hmc.edu Geoff Odhner geoff@franklin.com Geoff Whale geoffw@cse.unsw.EDU.AU Gerhard Poul gpoul@gnu.org Germano Leichsenring germano@jedi.cs.kobe-u.ac.jp GOTO Masanori gotom@debian.or.jp Greg Louis glouis@dynamicro.on.ca Greg McGary gkm@gnu.org Greg Troxel gdt@bbn.com Greg Wooledge gawooledge@sherwin.com Gregory Leblanc gleblanc@cu-portland.edu Göran Uddeborg goeran@uddeborg.pp.se H. J. Lu hjl@valinux.com Hans Verkuil hans@wyst.hobby.nl Harry Liu rliu@lek.ugcs.caltech.edu Herbert Xu herbert@gondor.apana.org.au Holger Berger hberger@ess.nec.de Hon-Yin Kok hkok@yoda.unl.edu Hugh Daniel hugh@xanadu.com Ian Bruce ian.bruce@myrealbox.com Ian Jackson ijackson@chiark.greenend.org.uk Ian Lance Taylor ian@cygnus.com Ian Turner vectro@pipeline.com +Iida Yosiaki iida@gnu.org James james@albion.glarp.com James Antill jmanti%essex.ac.uk@seralph21.essex.ac.uk James Sneeringer jvs@ocslink.com James Tanis jtt@soscorp.com James Youngman james+usenet@free-lunch.demon.co.uk Jamie Lokier jamie@imbolc.ucc.ie Jan Fedak J.Fedak@sh.cvut.cz Jan Nieuwenhuizen janneke@gnu.org Janos Farkas chexum@shadow.banki.hu Jarkko Hietaniemi jhi@epsilon.hut.fi Jeff Moore jbm@mordor.com Jeff Sheinberg jeffsh@erols.com Jens Schmidt jms@jsds.hamburg.com Jerome Abela abela@hsc.fr Jesse Thilo jgt2@eecs.lehigh.edu Jie Xu xuj@iag.net Jim Blandy jimb@cyclic.com Jim Dennis jimd@starshine.org Joakim Rosqvist dvljrt@cs.umu.se Jochen Hein jochen@jochen.org Joe Orton joe@orton.demon.co.uk Johan Danielsson joda@pdc.kth.se John Bley jbb6@acpub.duke.edu John David Anglin dave.anglin@nrc.ca John Gatewood Ham zappaman@alphabox.compsci.buu.ac.th John Gotts jgotts@umich.edu John Kendall kendall@capps.com John Kodis kodis@acm.org John Murphy jam@philabs.research.philips.com John Roll john@panic.harvard.edu John Salmon johns@mullet.anu.edu.au John Summerfield summer@OS2.ami.com.au Joost van Baal joostvb@xs4all.nl Jorge Stolfi stolfi@ic.unicamp.br Joseph S. Myers jsm28@cam.ac.uk Juan F. Codagnone juam@arnet.com.ar Jungshik Shin jshin@pantheon.yale.edu Jürgen Fluk louis@dachau.marco.de jvogel jvogel@linkny.com Kai-Uwe Rommel rommel@informatik.tu-muenchen.de Kai Henningsen kai@debian.org Kalle Olavi Niemitalo tosi@stekt.oulu.fi Kamal Paul Nigam Kamal_Paul_Nigam@gs35.sp.cs.cmu.edu Karl Eichwalder keichwa@gmx.net Karl Heuer kwzh@gnu.org Karl-Michael Schneider schneide@phil.uni-passau.de Karsten Thygesen karthy@kom.auc.dk Kaveh R. Ghazi ghazi@caip.rutgers.edu Keith Owens kaos@audio.apana.org.au Keith Thompson kst@sdsc.edu Ken Pizzini kenp@halcyon.com Kjetil Torgrim Homme kjetilho@ifi.uio.no Kristoffer Rose kris@diku.dk ??? kytek@cybercomm.net Larry McVoy lm@sgi.com Lars Hecking lhecking@nmrc.ucc.ie Lehti Rami rammer@cs.tut.fi Leonard N. Zubkoff lnz@dandelion.com Lorne Baker lbaker@nitro.avint.net M. P. Suzuki mpsuzuki@hiroshima-u.ac.jp Maciej Kwapulinski pikpok@univ.gda.pl Manas Garg manas@cygsoft.com Manfred Hollstein manfred@s-direktnet.de Marc Boucher marc@mbsi.ca Marc Olzheim marcolz@stack.nl Marco Franzen Marco.Franzen@Thyron.com +Marcus Brinkmann http://www.marcus-brinkmann.de Marcus Daniels marcus@ee.pdx.edu Mark A. Thomas thommark@access.digex.net Mark D. Roth roth@uiuc.edu Mark Harris mark@monitor.designacc.com Mark Hewitt mhewitt@armature.com Mark Hounschell markh@compro.net Mark Kettenis kettenis@phys.uva.nl Mark Nudelman marknu@flash.net Mark W. Eichin eichin@cygnus.com Markus Demleitner msdemlei@auriga.ari.uni-heidelberg.de Martin martin@dresden.nacamar.de Martin Gallant martyg@goodbit.net Martin Hippe martin.hippe@schlund.de +Martin Michlmayr tbm@cyrius.com Martin Mitchell martin@debian.org Martin P.J. Zinser zinser@decus.de Marty Leisner leisner@sdsp.mc.xerox.com Masami Takikawa takikawm@CS.ORST.EDU Mate Wierdl mw@moni.msci.memphis.edu Matej Vela mvela@public.srce.hr Matt Perry matt@primefactor.com Matt Schalit mschalit@pacbell.net +Matthew Arnison maffew@cat.org.au Matthew Braun matthew@ans.net Matthew Clarke Matthew_Clarke@mindlink.bc.ca Matthew S. Levine mslevine@theory.lcs.mit.edu Matthew Smith matts@bluesguitar.org Matthew Swift swift@alum.mit.edu Mattias Wadenstein maswan@acc.umu.se Matthias Urlichs smurf@noris.de Meelis Roos mroos@tartu.cyber.ee Michael ??? michael@roka.net Michael Deutschmann michael@talamasca.ocis.net Michael Gaughen mgaughen@polyserve.com Michael Hasselberg mikelh@zonta.ping.de Michael Hohn hohn@math.utah.edu Michael J. Croghan mcroghan@usatoday.com +Michael Piefel piefel@informatik.hu-berlin.de Michael Steffens michael.steffens@s.netic.de Michael Stone mstone@debian.org Michael Stutz stutz@dsl.org Michael Veksler mveksler@techunix.technion.ac.il Michel Robitaille robitail@IRO.UMontreal.CA Michiel Bacchiani bacchian@raven.bu.edu +Mike Castle dalgoda@ix.netcom.com Miles Bader miles@gnu.ai.mit.edu Minh Tran-Le tranle@intellicorp.com Morten Welinder terra@diku.dk Neal H Walfield neal@cs.uml.edu Nelson H. F. Beebe beebe@math.utah.edu Nick Lawes nlawes@silverplatter.com Niklas Edmundsson nikke@acc.umu.se Noah Friedman friedman@splode.com Noel Cragg noel@red-bean.com Olav Morkrid olav@funcom.com Ole Laursen olau@hardworking.dk Oskar Liljeblad osk@hem.passagen.se Ørn E. Hansen oehansen@daimi.aau.dk Paul Eggert eggert@twinsun.com Paul Nevai nevai@ops.mps.ohio-state.edu Paul Sauer paul@alexa.com Paul Slootman paul@debian.org Pawel Prokop pablo@wizard.ae.krakow.pl Per Cederqvist ceder@lysator.liu.se Per Kristian Hove perhov@math.ntnu.no Peter Eriksson peter@ifm.liu.se Peter Moulder reiter@netspace.net.au Peter Samuelson psamuels@sampo.creighton.edu Peter Seebach seebs@taniemarie.solon.com Petter Reinholdtsen pere@hungry.com Phelippe Neveu pneveu@pcigeomatics.com Phil Richards phil.richards@vf.vodafone.co.uk Philippe De Muyter phdm@macqel.be Philippe Schnoebelen Philippe.Schnoebelen@imag.fr Piergiorgio Sartor sartor@sony.de Piotr Kwapulinski kwap@univ.gda.pl Prashant TR tr@eth.net Rainer Orth ro@TechFak.Uni-Bielefeld.DE Ralf W. Stephan stephan@tmt.de Ralph Loader loader@maths.ox.ac.uk Raul Miller moth@magenta.com Richard Braakman dark@xs4all.nl Richard Dawe richdawe@bigfoot.com Richard J. Rauenzahn rrauenza@hairball.cup.hp.com Richard Sharman rsharman@magmacom.com Rick Sladkey jrs@world.std.com Rik Faith faith@cs.unc.edu Risto Kankkunen kankkune@lingsoft.fi Robert H. de Vries robert@and.nl Rogier Wolff R.E.Wolff@BitWizard.nl Roland Huebner ro-huebner@gmx.de Ronald F. Guilmette rfg@netcom.com Ross Alexander r.alexander@auckland.ac.nz Ross Paterson rap@doc.ic.ac.uk Ross Ridge rridge@calum.csclub.uwaterloo.ca Sami Farin sfarin@ratol.fi Samuli Karkkainen Samuli.Karkkainen@hut.fi Sander van Malssen svm@kozmix.ow.nl Santiago Vila Doncel sanvila@unex.es Savochkin Andrey Vladimirovich saw@msu.ru Scott Lurndal slurn@griffin.engr.sgi.com Soeren Sonnenburg sonnenburg@informatik.hu-berlin.de Stéphane Chazelas Stephane_CHAZELAS@yahoo.fr Stephen Eglen eglen@pcg.wustl.edu Stephen Gildea gildea@stop.mail-abuse.org Stephen Smoogen ?????????? Steve McConnel steve@acadcomp.sil.org Stuart Kemp skemp@peter.bmc.com Tadayoshi Funaba tadf@kt.rim.or.jp +TAKAI Kousuke takai@vlsi.kuee.kyoto-u.ac.jp Theodore Ts'o tytso@rsts-11.mit.edu Thomas Bushnell thomas@gnu.ai.mit.edu Thomas Quinot thomas@Cuivre.FR.EU.ORG +Tim J. Robbins tjr@FreeBSD.org Tim Smithers mouse@dmouse.com.au Tim Waugh twaugh@redhat Todd A. Jacobs tjacobs@codegnome.org Tom Quinn trq@dionysos.thphys.ox.ac.uk Ton Hospel thospel@mail.dma.be +Tony Kocurko akocurko@mun.ca Tony Leneis tony@plaza.ds.adp.com Tony Robinson ajr@eng.cam.ac.uk Torbjorn Granlund tege@nada.kth.se Torbjorn Lindgren tl@funcom.no Torsten Landschoff torsten@pclab.ifg.uni-kiel.de Ulrich Drepper drepper@gnu.org Urs Thuermann urs@isnogud.escape.de Uwe H. Steinfeld usteinfeld@gmx.net Vesselin Atanasov vesselin@bgnet.bg Vin Shelton acs@alumni.princeton.edu Volker Borchert bt@teknon.de Wayne Stewart wstewa@atl.com Wenjun Zheng zwj@yahoo.com Werner Almesberger Werner.Almesberger@epfl.ch Wichert Akkerman wichert@cistron.nl Will Edgington wedgingt@acm.org William Bader william@nscs.fast.net William Dowling will@franklin.com William Lewis wiml@omnigroup.com +Wojciech Purczynski cliph@isec.pl Wolfram Kleff kleff@cs.uni-bonn.de Won-kyu Park wkpark@chem.skku.ac.kr Zvi Har'El rl@math.technion.ac.il diff --git a/contrib/gnu-sort/lib/getopt.c b/contrib/gnu-sort/lib/getopt.c index fa46c85e3cae..08ba35be3637 100644 --- a/contrib/gnu-sort/lib/getopt.c +++ b/contrib/gnu-sort/lib/getopt.c @@ -1,772 +1,1068 @@ /* Getopt for GNU. NOTE: getopt is now part of the C library, so if you don't know what - "Keep this file name-space clean" means, talk to roland@gnu.ai.mit.edu + "Keep this file name-space clean" means, talk to drepper@gnu.org before changing it! - - Copyright (C) 1987, 88, 89, 90, 91, 92, 93, 94, 95 + Copyright (C) 1987,88,89,90,91,92,93,94,95,96,98,99,2000,2001 Free Software Foundation, Inc. + This file is part of the GNU C Library. - This program is free software; you can redistribute it and/or modify it - under the terms of the GNU General Public License as published by the - Free Software Foundation; either version 2, or (at your option) any - later version. + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. - This program is distributed in the hope that it will be useful, + The GNU C Library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of - MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - GNU General Public License for more details. - - You should have received a copy of the GNU General Public License - along with this program; if not, write to the Free Software - Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. -/* $FreeBSD$ */ + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, write to the Free + Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA + 02111-1307 USA. */ /* This tells Alpha OSF/1 not to define a getopt prototype in . Ditto for AIX 3.2 and . */ #ifndef _NO_PROTO -#define _NO_PROTO +# define _NO_PROTO #endif #ifdef HAVE_CONFIG_H -#include +# include #endif -#if !defined (__STDC__) || !__STDC__ +#if !defined __STDC__ || !__STDC__ /* This is a separate conditional since some stdc systems reject `defined (const)'. */ -#ifndef const -#define const -#endif +# ifndef const +# define const +# endif #endif #include /* Comment out all this code if we are using the GNU C Library, and are not actually compiling the library itself. This code is part of the GNU C Library, but also included in many other GNU distributions. Compiling and linking in this code is a waste when using the GNU C library (especially if it is a shared library). Rather than having every GNU program understand `configure --with-gnu-libc' and omit the object files, it is simpler to just do this in the source for each such file. */ -#if defined (_LIBC) || !defined (__GNU_LIBRARY__) +#define GETOPT_INTERFACE_VERSION 2 +#if !defined _LIBC && defined __GLIBC__ && __GLIBC__ >= 2 +# include +# if _GNU_GETOPT_INTERFACE_VERSION == GETOPT_INTERFACE_VERSION +# define ELIDE_CODE +# endif +#endif + +#ifndef ELIDE_CODE /* This needs to come after some library #include to get __GNU_LIBRARY__ defined. */ #ifdef __GNU_LIBRARY__ /* Don't include stdlib.h for non-GNU C libraries because some of them contain conflicting prototypes for getopt. */ -#include +# include +# include #endif /* GNU C library. */ -#ifndef _ -/* This is for other GNU distributions with internationalized messages. - When compiling libc, the _ macro is predefined. */ -#ifdef HAVE_LIBINTL_H -# include -# define _(msgid) gettext (msgid) -#else -# define _(msgid) (msgid) +#ifdef VMS +# include +# if HAVE_STRING_H - 0 +# include +# endif #endif + +#ifndef _ +/* This is for other GNU distributions with internationalized messages. */ +# if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC +# include +# ifndef _ +# define _(msgid) gettext (msgid) +# endif +# else +# define _(msgid) (msgid) +# endif #endif /* This version of `getopt' appears to the caller like standard Unix `getopt' but it behaves differently for the user, since it allows the user to intersperse the options with the other arguments. As `getopt' works, it permutes the elements of ARGV so that, when it is done, all the options precede everything else. Thus all application programs are extended to handle flexible argument order. Setting the environment variable POSIXLY_CORRECT disables permutation. Then the behavior is completely standard. GNU application programs can use a third alternative mode in which they can distinguish the relative order of options and other arguments. */ #include "getopt.h" /* For communication from `getopt' to the caller. When `getopt' finds an option that takes an argument, the argument value is returned here. Also, when `ordering' is RETURN_IN_ORDER, each non-option ARGV-element is returned here. */ -char *optarg = NULL; +char *optarg; /* Index in ARGV of the next element to be scanned. This is used for communication to and from the caller and for communication between successive calls to `getopt'. On entry to `getopt', zero means this is the first call; initialize. - When `getopt' returns EOF, this is the index of the first of the + When `getopt' returns -1, this is the index of the first of the non-option elements that the caller should itself scan. Otherwise, `optind' communicates from one call to the next how much of ARGV has been scanned so far. */ -/* XXX 1003.2 says this must be 1 before any call. */ -int optind = 0; +/* 1003.2 says this must be 1 before any call. */ +int optind = 1; + +/* Formerly, initialization of getopt depended on optind==0, which + causes problems with re-calling getopt as programs generally don't + know that. */ + +int __getopt_initialized; /* The next char to be scanned in the option-element in which the last option character we returned was found. This allows us to pick up the scan where we left off. If this is zero, or a null string, it means resume the scan by advancing to the next ARGV-element. */ static char *nextchar; /* Callers store zero here to inhibit the error message for unrecognized options. */ int opterr = 1; /* Set to an option character which was unrecognized. This must be initialized on some systems to avoid linking in the system's own getopt implementation. */ int optopt = '?'; /* Describe how to deal with options that follow non-option ARGV-elements. If the caller did not specify anything, the default is REQUIRE_ORDER if the environment variable POSIXLY_CORRECT is defined, PERMUTE otherwise. REQUIRE_ORDER means don't recognize them as options; stop option processing when the first non-option is seen. This is what Unix does. This mode of operation is selected by either setting the environment variable POSIXLY_CORRECT, or using `+' as the first character of the list of option characters. PERMUTE is the default. We permute the contents of ARGV as we scan, so that eventually all the non-options are at the end. This allows options to be given in any order, even with programs that were not written to expect this. RETURN_IN_ORDER is an option available to programs that were written to expect options and other ARGV-elements in any order and that care about the ordering of the two. We describe each non-option ARGV-element as if it were the argument of an option with character code 1. Using `-' as the first character of the list of option characters selects this mode of operation. The special argument `--' forces an end of option-scanning regardless of the value of `ordering'. In the case of RETURN_IN_ORDER, only - `--' can cause `getopt' to return EOF with `optind' != ARGC. */ + `--' can cause `getopt' to return -1 with `optind' != ARGC. */ static enum { REQUIRE_ORDER, PERMUTE, RETURN_IN_ORDER } ordering; /* Value of POSIXLY_CORRECT environment variable. */ static char *posixly_correct; #ifdef __GNU_LIBRARY__ /* We want to avoid inclusion of string.h with non-GNU libraries because there are many ways it can cause trouble. On some systems, it contains special magic macros that don't work in GCC. */ -#include -#define my_index strchr +# include +# define my_index strchr #else +# if HAVE_STRING_H +# include +# else +# include +# endif + /* Avoid depending on library functions or files whose names are inconsistent. */ -char *getenv (); +#ifndef getenv +extern char *getenv (); +#endif static char * my_index (str, chr) const char *str; int chr; { while (*str) { if (*str == chr) return (char *) str; str++; } return 0; } /* If using GCC, we can safely declare strlen this way. If not using GCC, it is ok not to declare it. */ #ifdef __GNUC__ /* Note that Motorola Delta 68k R3V7 comes with GCC but not stddef.h. That was relevant to code that was here before. */ -#if !defined (__STDC__) || !__STDC__ +# if (!defined __STDC__ || !__STDC__) && !defined strlen /* gcc with -traditional declares the built-in strlen to return int, and has done so at least since version 2.4.5. -- rms. */ extern int strlen (const char *); -#endif /* not __STDC__ */ +# endif /* not __STDC__ */ #endif /* __GNUC__ */ #endif /* not __GNU_LIBRARY__ */ /* Handle permutation of arguments. */ /* Describe the part of ARGV that contains non-options that have been skipped. `first_nonopt' is the index in ARGV of the first of them; `last_nonopt' is the index after the last of them. */ static int first_nonopt; static int last_nonopt; +#ifdef _LIBC +/* Bash 2.0 gives us an environment variable containing flags + indicating ARGV elements that should not be considered arguments. */ + +#ifdef USE_NONOPTION_FLAGS +/* Defined in getopt_init.c */ +extern char *__getopt_nonoption_flags; + +static int nonoption_flags_max_len; +static int nonoption_flags_len; +#endif + +static int original_argc; +static char *const *original_argv; + +/* Make sure the environment variable bash 2.0 puts in the environment + is valid for the getopt call we must make sure that the ARGV passed + to getopt is that one passed to the process. */ +static void +__attribute__ ((unused)) +store_args_and_env (int argc, char *const *argv) +{ + /* XXX This is no good solution. We should rather copy the args so + that we can compare them later. But we must not use malloc(3). */ + original_argc = argc; + original_argv = argv; +} +# ifdef text_set_element +text_set_element (__libc_subinit, store_args_and_env); +# endif /* text_set_element */ + +# ifdef USE_NONOPTION_FLAGS +# define SWAP_FLAGS(ch1, ch2) \ + if (nonoption_flags_len > 0) \ + { \ + char __tmp = __getopt_nonoption_flags[ch1]; \ + __getopt_nonoption_flags[ch1] = __getopt_nonoption_flags[ch2]; \ + __getopt_nonoption_flags[ch2] = __tmp; \ + } +# else +# define SWAP_FLAGS(ch1, ch2) +# endif +#else /* !_LIBC */ +# define SWAP_FLAGS(ch1, ch2) +#endif /* _LIBC */ + /* Exchange two adjacent subsequences of ARGV. One subsequence is elements [first_nonopt,last_nonopt) which contains all the non-options that have been skipped so far. The other is elements [last_nonopt,optind), which contains all the options processed since those non-options were skipped. `first_nonopt' and `last_nonopt' are relocated so that they describe the new indices of the non-options in ARGV after they are moved. */ +#if defined __STDC__ && __STDC__ +static void exchange (char **); +#endif + static void exchange (argv) char **argv; { int bottom = first_nonopt; int middle = last_nonopt; int top = optind; char *tem; /* Exchange the shorter segment with the far end of the longer segment. That puts the shorter segment into the right place. It leaves the longer segment in the right place overall, but it consists of two parts that need to be swapped next. */ +#if defined _LIBC && defined USE_NONOPTION_FLAGS + /* First make sure the handling of the `__getopt_nonoption_flags' + string can work normally. Our top argument must be in the range + of the string. */ + if (nonoption_flags_len > 0 && top >= nonoption_flags_max_len) + { + /* We must extend the array. The user plays games with us and + presents new arguments. */ + char *new_str = malloc (top + 1); + if (new_str == NULL) + nonoption_flags_len = nonoption_flags_max_len = 0; + else + { + memset (__mempcpy (new_str, __getopt_nonoption_flags, + nonoption_flags_max_len), + '\0', top + 1 - nonoption_flags_max_len); + nonoption_flags_max_len = top + 1; + __getopt_nonoption_flags = new_str; + } + } +#endif + while (top > middle && middle > bottom) { if (top - middle > middle - bottom) { /* Bottom segment is the short one. */ int len = middle - bottom; register int i; /* Swap it with the top part of the top segment. */ for (i = 0; i < len; i++) { tem = argv[bottom + i]; argv[bottom + i] = argv[top - (middle - bottom) + i]; argv[top - (middle - bottom) + i] = tem; + SWAP_FLAGS (bottom + i, top - (middle - bottom) + i); } /* Exclude the moved bottom segment from further swapping. */ top -= len; } else { /* Top segment is the short one. */ int len = top - middle; register int i; /* Swap it with the bottom part of the bottom segment. */ for (i = 0; i < len; i++) { tem = argv[bottom + i]; argv[bottom + i] = argv[middle + i]; argv[middle + i] = tem; + SWAP_FLAGS (bottom + i, middle + i); } /* Exclude the moved top segment from further swapping. */ bottom += len; } } /* Update records for the slots the non-options now occupy. */ first_nonopt += (optind - last_nonopt); last_nonopt = optind; } /* Initialize the internal data when the first call is made. */ +#if defined __STDC__ && __STDC__ +static const char *_getopt_initialize (int, char *const *, const char *); +#endif static const char * -_getopt_initialize (optstring) +_getopt_initialize (argc, argv, optstring) + int argc; + char *const *argv; const char *optstring; { /* Start processing options with ARGV-element 1 (since ARGV-element 0 is the program name); the sequence of previously skipped non-option ARGV-elements is empty. */ - first_nonopt = last_nonopt = optind = 1; + first_nonopt = last_nonopt = optind; nextchar = NULL; posixly_correct = getenv ("POSIXLY_CORRECT"); /* Determine how to handle the ordering of options and nonoptions. */ if (optstring[0] == '-') { ordering = RETURN_IN_ORDER; ++optstring; } else if (optstring[0] == '+') { ordering = REQUIRE_ORDER; ++optstring; } else if (posixly_correct != NULL) ordering = REQUIRE_ORDER; else ordering = PERMUTE; +#if defined _LIBC && defined USE_NONOPTION_FLAGS + if (posixly_correct == NULL + && argc == original_argc && argv == original_argv) + { + if (nonoption_flags_max_len == 0) + { + if (__getopt_nonoption_flags == NULL + || __getopt_nonoption_flags[0] == '\0') + nonoption_flags_max_len = -1; + else + { + const char *orig_str = __getopt_nonoption_flags; + int len = nonoption_flags_max_len = strlen (orig_str); + if (nonoption_flags_max_len < argc) + nonoption_flags_max_len = argc; + __getopt_nonoption_flags = + (char *) malloc (nonoption_flags_max_len); + if (__getopt_nonoption_flags == NULL) + nonoption_flags_max_len = -1; + else + memset (__mempcpy (__getopt_nonoption_flags, orig_str, len), + '\0', nonoption_flags_max_len - len); + } + } + nonoption_flags_len = nonoption_flags_max_len; + } + else + nonoption_flags_len = 0; +#endif + return optstring; } /* Scan elements of ARGV (whose length is ARGC) for option characters given in OPTSTRING. If an element of ARGV starts with '-', and is not exactly "-" or "--", then it is an option element. The characters of this element (aside from the initial '-') are option characters. If `getopt' is called repeatedly, it returns successively each of the option characters from each of the option elements. If `getopt' finds another option character, it returns that character, updating `optind' and `nextchar' so that the next call to `getopt' can resume the scan with the following option character or ARGV-element. - If there are no more option characters, `getopt' returns `EOF'. + If there are no more option characters, `getopt' returns -1. Then `optind' is the index in ARGV of the first ARGV-element that is not an option. (The ARGV-elements have been permuted so that those that are not options now come last.) OPTSTRING is a string containing the legitimate option characters. If an option character is seen that is not listed in OPTSTRING, return '?' after printing an error message. If you set `opterr' to zero, the error message is suppressed but we still return '?'. If a char in OPTSTRING is followed by a colon, that means it wants an arg, so the following text in the same ARGV-element, or the text of the following ARGV-element, is returned in `optarg'. Two colons mean an option that wants an optional arg; if there is text in the current ARGV-element, it is returned in `optarg', otherwise `optarg' is set to zero. If OPTSTRING starts with `-' or `+', it requests different methods of handling the non-option ARGV-elements. See the comments about RETURN_IN_ORDER and REQUIRE_ORDER, above. Long-named options begin with `--' instead of `-'. Their names may be abbreviated as long as the abbreviation is unique or is an exact match for some defined option. If they have an argument, it follows the option name in the same ARGV-element, separated from the option name by a `=', or else the in next ARGV-element. When `getopt' finds a long-named option, it returns 0 if that option's `flag' field is nonzero, the value of the option's `val' field if the `flag' field is zero. The elements of ARGV aren't really const, because we permute them. But we pretend they're const in the prototype to be compatible with other systems. LONGOPTS is a vector of `struct option' terminated by an element containing a name which is zero. LONGIND returns the index in LONGOPT of the long-named option found. It is only valid when a long-named option has been found by the most recent call. If LONG_ONLY is nonzero, '-' as well as '--' can introduce long-named options. */ int _getopt_internal (argc, argv, optstring, longopts, longind, long_only) int argc; char *const *argv; const char *optstring; const struct option *longopts; int *longind; int long_only; { + int print_errors = opterr; + if (optstring[0] == ':') + print_errors = 0; + + if (argc < 1) + return -1; + optarg = NULL; - if (optind == 0) + if (optind == 0 || !__getopt_initialized) { - optstring = _getopt_initialize (optstring); - optind = 1; /* Don't scan ARGV[0], the program name. */ + if (optind == 0) + optind = 1; /* Don't scan ARGV[0], the program name. */ + optstring = _getopt_initialize (argc, argv, optstring); + __getopt_initialized = 1; } + /* Test whether ARGV[optind] points to a non-option argument. + Either it does not have option syntax, or there is an environment flag + from the shell indicating it is not an option. The later information + is only used when the used in the GNU libc. */ +#if defined _LIBC && defined USE_NONOPTION_FLAGS +# define NONOPTION_P (argv[optind][0] != '-' || argv[optind][1] == '\0' \ + || (optind < nonoption_flags_len \ + && __getopt_nonoption_flags[optind] == '1')) +#else +# define NONOPTION_P (argv[optind][0] != '-' || argv[optind][1] == '\0') +#endif + if (nextchar == NULL || *nextchar == '\0') { /* Advance to the next ARGV-element. */ + /* Give FIRST_NONOPT & LAST_NONOPT rational values if OPTIND has been + moved back by the user (who may also have changed the arguments). */ + if (last_nonopt > optind) + last_nonopt = optind; + if (first_nonopt > optind) + first_nonopt = optind; + if (ordering == PERMUTE) { /* If we have just processed some options following some non-options, exchange them so that the options come first. */ if (first_nonopt != last_nonopt && last_nonopt != optind) exchange ((char **) argv); else if (last_nonopt != optind) first_nonopt = optind; /* Skip any additional non-options and extend the range of non-options previously skipped. */ - while (optind < argc - && (argv[optind][0] != '-' || argv[optind][1] == '\0')) + while (optind < argc && NONOPTION_P) optind++; last_nonopt = optind; } /* The special ARGV-element `--' means premature end of options. Skip it like a null option, then exchange with previous non-options as if it were an option, then skip everything else like a non-option. */ if (optind != argc && !strcmp (argv[optind], "--")) { optind++; if (first_nonopt != last_nonopt && last_nonopt != optind) exchange ((char **) argv); else if (first_nonopt == last_nonopt) first_nonopt = optind; last_nonopt = argc; optind = argc; } /* If we have done all the ARGV-elements, stop the scan and back over any non-options that we skipped and permuted. */ if (optind == argc) { /* Set the next-arg-index to point at the non-options that we previously skipped, so the caller will digest them. */ if (first_nonopt != last_nonopt) optind = first_nonopt; - return EOF; + return -1; } /* If we have come to a non-option and did not permute it, either stop the scan or describe it to the caller and pass it by. */ - if ((argv[optind][0] != '-' || argv[optind][1] == '\0')) + if (NONOPTION_P) { if (ordering == REQUIRE_ORDER) - return EOF; + return -1; optarg = argv[optind++]; return 1; } /* We have found another option-ARGV-element. Skip the initial punctuation. */ nextchar = (argv[optind] + 1 + (longopts != NULL && argv[optind][1] == '-')); } /* Decode the current option-ARGV-element. */ /* Check whether the ARGV-element is a long option. If long_only and the ARGV-element has the form "-f", where f is a valid short option, don't consider it an abbreviated form of a long option that starts with f. Otherwise there would be no way to give the -f short option. On the other hand, if there's a long option "fubar" and the ARGV-element is "-fu", do consider that an abbreviation of the long option, just like "--fu", and not "-f" with arg "u". This distinction seems to be the most useful approach. */ if (longopts != NULL && (argv[optind][1] == '-' - || (long_only && (argv[optind][2] - || !my_index (optstring, argv[optind][1]))))) + || (long_only && (argv[optind][2] || !my_index (optstring, argv[optind][1]))))) { char *nameend; const struct option *p; const struct option *pfound = NULL; int exact = 0; int ambig = 0; - int indfound; + int indfound = -1; int option_index; for (nameend = nextchar; *nameend && *nameend != '='; nameend++) /* Do nothing. */ ; -#ifdef lint - indfound = 0; /* Avoid spurious compiler warning. */ -#endif - /* Test all long options for either exact match or abbreviated matches. */ for (p = longopts, option_index = 0; p->name; p++, option_index++) if (!strncmp (p->name, nextchar, nameend - nextchar)) { - if (nameend - nextchar == strlen (p->name)) + if ((unsigned int) (nameend - nextchar) + == (unsigned int) strlen (p->name)) { /* Exact match found. */ pfound = p; indfound = option_index; exact = 1; break; } else if (pfound == NULL) { /* First nonexact match found. */ pfound = p; indfound = option_index; } - else + else if (long_only + || pfound->has_arg != p->has_arg + || pfound->flag != p->flag + || pfound->val != p->val) /* Second or later nonexact match found. */ ambig = 1; } if (ambig && !exact) { - if (opterr) + if (print_errors) fprintf (stderr, _("%s: option `%s' is ambiguous\n"), argv[0], argv[optind]); nextchar += strlen (nextchar); optind++; + optopt = 0; return '?'; } if (pfound != NULL) { option_index = indfound; optind++; if (*nameend) { /* Don't test has_arg with >, because some C compilers don't allow it to be used on enums. */ if (pfound->has_arg) optarg = nameend + 1; else { - if (opterr) - if (argv[optind - 1][1] == '-') - /* --option */ - fprintf (stderr, - _("%s: option `--%s' doesn't allow an argument\n"), - argv[0], pfound->name); - else - /* +option or -option */ - fprintf (stderr, - _("%s: option `%c%s' doesn't allow an argument\n"), - argv[0], argv[optind - 1][0], pfound->name); + if (print_errors) + { + if (argv[optind - 1][1] == '-') + /* --option */ + fprintf (stderr, + _("%s: option `--%s' doesn't allow an argument\n"), + argv[0], pfound->name); + else + /* +option or -option */ + fprintf (stderr, + _("%s: option `%c%s' doesn't allow an argument\n"), + argv[0], argv[optind - 1][0], pfound->name); + } nextchar += strlen (nextchar); + + optopt = pfound->val; return '?'; } } else if (pfound->has_arg == 1) { if (optind < argc) optarg = argv[optind++]; else { - if (opterr) + if (print_errors) fprintf (stderr, _("%s: option `%s' requires an argument\n"), argv[0], argv[optind - 1]); nextchar += strlen (nextchar); + optopt = pfound->val; return optstring[0] == ':' ? ':' : '?'; } } nextchar += strlen (nextchar); if (longind != NULL) *longind = option_index; if (pfound->flag) { *(pfound->flag) = pfound->val; return 0; } return pfound->val; } /* Can't find it as a long option. If this is not getopt_long_only, or the option starts with '--' or is not a valid short option, then it's an error. Otherwise interpret it as a short option. */ if (!long_only || argv[optind][1] == '-' || my_index (optstring, *nextchar) == NULL) { - if (opterr) + if (print_errors) { if (argv[optind][1] == '-') /* --option */ fprintf (stderr, _("%s: unrecognized option `--%s'\n"), argv[0], nextchar); else /* +option or -option */ fprintf (stderr, _("%s: unrecognized option `%c%s'\n"), argv[0], argv[optind][0], nextchar); } nextchar = (char *) ""; optind++; + optopt = 0; return '?'; } } /* Look at and handle the next short option-character. */ { char c = *nextchar++; char *temp = my_index (optstring, c); /* Increment `optind' when we start to process its last character. */ if (*nextchar == '\0') ++optind; if (temp == NULL || c == ':') { - if (opterr) + if (print_errors) { if (posixly_correct) /* 1003.2 specifies the format of this message. */ fprintf (stderr, _("%s: illegal option -- %c\n"), argv[0], c); else fprintf (stderr, _("%s: invalid option -- %c\n"), argv[0], c); } optopt = c; return '?'; } + /* Convenience. Treat POSIX -W foo same as long option --foo */ + if (temp[0] == 'W' && temp[1] == ';') + { + char *nameend; + const struct option *p; + const struct option *pfound = NULL; + int exact = 0; + int ambig = 0; + int indfound = 0; + int option_index; + + /* This is an option that requires an argument. */ + if (*nextchar != '\0') + { + optarg = nextchar; + /* If we end this ARGV-element by taking the rest as an arg, + we must advance to the next element now. */ + optind++; + } + else if (optind == argc) + { + if (print_errors) + { + /* 1003.2 specifies the format of this message. */ + fprintf (stderr, _("%s: option requires an argument -- %c\n"), + argv[0], c); + } + optopt = c; + if (optstring[0] == ':') + c = ':'; + else + c = '?'; + return c; + } + else + /* We already incremented `optind' once; + increment it again when taking next ARGV-elt as argument. */ + optarg = argv[optind++]; + + /* optarg is now the argument, see if it's in the + table of longopts. */ + + for (nextchar = nameend = optarg; *nameend && *nameend != '='; nameend++) + /* Do nothing. */ ; + + /* Test all long options for either exact match + or abbreviated matches. */ + for (p = longopts, option_index = 0; p->name; p++, option_index++) + if (!strncmp (p->name, nextchar, nameend - nextchar)) + { + if ((unsigned int) (nameend - nextchar) == strlen (p->name)) + { + /* Exact match found. */ + pfound = p; + indfound = option_index; + exact = 1; + break; + } + else if (pfound == NULL) + { + /* First nonexact match found. */ + pfound = p; + indfound = option_index; + } + else + /* Second or later nonexact match found. */ + ambig = 1; + } + if (ambig && !exact) + { + if (print_errors) + fprintf (stderr, _("%s: option `-W %s' is ambiguous\n"), + argv[0], argv[optind]); + nextchar += strlen (nextchar); + optind++; + return '?'; + } + if (pfound != NULL) + { + option_index = indfound; + if (*nameend) + { + /* Don't test has_arg with >, because some C compilers don't + allow it to be used on enums. */ + if (pfound->has_arg) + optarg = nameend + 1; + else + { + if (print_errors) + fprintf (stderr, _("\ +%s: option `-W %s' doesn't allow an argument\n"), + argv[0], pfound->name); + + nextchar += strlen (nextchar); + return '?'; + } + } + else if (pfound->has_arg == 1) + { + if (optind < argc) + optarg = argv[optind++]; + else + { + if (print_errors) + fprintf (stderr, + _("%s: option `%s' requires an argument\n"), + argv[0], argv[optind - 1]); + nextchar += strlen (nextchar); + return optstring[0] == ':' ? ':' : '?'; + } + } + nextchar += strlen (nextchar); + if (longind != NULL) + *longind = option_index; + if (pfound->flag) + { + *(pfound->flag) = pfound->val; + return 0; + } + return pfound->val; + } + nextchar = NULL; + return 'W'; /* Let the application handle it. */ + } if (temp[1] == ':') { if (temp[2] == ':') { /* This is an option that accepts an argument optionally. */ if (*nextchar != '\0') { optarg = nextchar; optind++; } else optarg = NULL; nextchar = NULL; } else { /* This is an option that requires an argument. */ if (*nextchar != '\0') { optarg = nextchar; /* If we end this ARGV-element by taking the rest as an arg, we must advance to the next element now. */ optind++; } else if (optind == argc) { - if (opterr) + if (print_errors) { /* 1003.2 specifies the format of this message. */ fprintf (stderr, - _("%s: option requires an argument -- %c\n"), - argv[0], c); + _("%s: option requires an argument -- %c\n"), + argv[0], c); } optopt = c; if (optstring[0] == ':') c = ':'; else c = '?'; } else /* We already incremented `optind' once; increment it again when taking next ARGV-elt as argument. */ optarg = argv[optind++]; nextchar = NULL; } } return c; } } int getopt (argc, argv, optstring) int argc; char *const *argv; const char *optstring; { return _getopt_internal (argc, argv, optstring, (const struct option *) 0, (int *) 0, 0); } -#endif /* _LIBC or not __GNU_LIBRARY__. */ +#endif /* Not ELIDE_CODE. */ #ifdef TEST /* Compile with -DTEST to make an executable for use in testing the above definition of `getopt'. */ int main (argc, argv) int argc; char **argv; { int c; int digit_optind = 0; while (1) { int this_option_optind = optind ? optind : 1; c = getopt (argc, argv, "abc:d:0123456789"); - if (c == EOF) + if (c == -1) break; switch (c) { case '0': case '1': case '2': case '3': case '4': case '5': case '6': case '7': case '8': case '9': if (digit_optind != 0 && digit_optind != this_option_optind) printf ("digits occur in two different argv-elements.\n"); digit_optind = this_option_optind; printf ("option %c\n", c); break; case 'a': printf ("option a\n"); break; case 'b': printf ("option b\n"); break; case 'c': printf ("option c with value `%s'\n", optarg); break; case '?': break; default: printf ("?? getopt returned character code 0%o ??\n", c); } } if (optind < argc) { printf ("non-option ARGV-elements: "); while (optind < argc) printf ("%s ", argv[optind++]); printf ("\n"); } exit (0); } #endif /* TEST */ diff --git a/contrib/gnu-sort/lib/getopt.h b/contrib/gnu-sort/lib/getopt.h index 89bc731dfa29..a1b8dd6658f2 100644 --- a/contrib/gnu-sort/lib/getopt.h +++ b/contrib/gnu-sort/lib/getopt.h @@ -1,131 +1,180 @@ /* Declarations for getopt. - Copyright (C) 1989, 90, 91, 92, 93, 94 Free Software Foundation, Inc. + Copyright (C) 1989-1994, 1996-1999, 2001 Free Software Foundation, Inc. + This file is part of the GNU C Library. - This program is free software; you can redistribute it and/or modify it - under the terms of the GNU General Public License as published by the - Free Software Foundation; either version 2, or (at your option) any - later version. + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. - This program is distributed in the hope that it will be useful, + The GNU C Library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of - MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - GNU General Public License for more details. + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. - You should have received a copy of the GNU General Public License - along with this program; if not, write to the Free Software - Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ - -/* $FreeBSD$ */ + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, write to the Free + Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA + 02111-1307 USA. */ #ifndef _GETOPT_H -#define _GETOPT_H 1 + +#ifndef __need_getopt +# define _GETOPT_H 1 +#endif + +/* If __GNU_LIBRARY__ is not already defined, either we are being used + standalone, or this is the first header included in the source file. + If we are being used with glibc, we need to include , but + that does not exist if we are standalone. So: if __GNU_LIBRARY__ is + not defined, include , which will pull in for us + if it's from glibc. (Why ctype.h? It's guaranteed to exist and it + doesn't flood the namespace with stuff the way some other headers do.) */ +#if !defined __GNU_LIBRARY__ +# include +#endif #ifdef __cplusplus extern "C" { #endif /* For communication from `getopt' to the caller. When `getopt' finds an option that takes an argument, the argument value is returned here. Also, when `ordering' is RETURN_IN_ORDER, each non-option ARGV-element is returned here. */ extern char *optarg; /* Index in ARGV of the next element to be scanned. This is used for communication to and from the caller and for communication between successive calls to `getopt'. On entry to `getopt', zero means this is the first call; initialize. - When `getopt' returns EOF, this is the index of the first of the + When `getopt' returns -1, this is the index of the first of the non-option elements that the caller should itself scan. Otherwise, `optind' communicates from one call to the next how much of ARGV has been scanned so far. */ extern int optind; /* Callers store zero here to inhibit the error message `getopt' prints for unrecognized options. */ extern int opterr; /* Set to an option character which was unrecognized. */ extern int optopt; +#ifndef __need_getopt /* Describe the long-named options requested by the application. The LONG_OPTIONS argument to getopt_long or getopt_long_only is a vector of `struct option' terminated by an element containing a name which is zero. The field `has_arg' is: no_argument (or 0) if the option does not take an argument, required_argument (or 1) if the option requires an argument, optional_argument (or 2) if the option takes an optional argument. If the field `flag' is not NULL, it points to a variable that is set to the value given in the field `val' when the option is found, but left unchanged if the option is not found. To have a long-named option do something other than set an `int' to a compiled-in constant, such as set a value from `optarg', set the option's `flag' field to zero and its `val' field to a nonzero value (the equivalent single-letter option character, if there is one). For long options that have a zero `flag' field, `getopt' returns the contents of the `val' field. */ struct option { -#if defined (__STDC__) && __STDC__ +# if (defined __STDC__ && __STDC__) || defined __cplusplus const char *name; -#else +# else char *name; -#endif +# endif /* has_arg can't be an enum because some compilers complain about type mismatches in all the code that assumes it is an int. */ int has_arg; int *flag; int val; }; /* Names for the values of the `has_arg' field of `struct option'. */ -#define no_argument 0 -#define required_argument 1 -#define optional_argument 2 +# define no_argument 0 +# define required_argument 1 +# define optional_argument 2 +#endif /* need getopt */ + + +/* Get definitions and prototypes for functions to process the + arguments in ARGV (ARGC of them, minus the program name) for + options given in OPTS. -#if defined (__STDC__) && __STDC__ -#ifdef __GNU_LIBRARY__ + Return the option character from OPTS just read. Return -1 when + there are no more options. For unrecognized options, or options + missing arguments, `optopt' is set to the option letter, and '?' is + returned. + + The OPTS string is a list of characters which are recognized option + letters, optionally followed by colons, specifying that that letter + takes an argument, to be placed in `optarg'. + + If a letter in OPTS is followed by two colons, its argument is + optional. This behavior is specific to the GNU `getopt'. + + The argument `--' causes premature termination of argument + scanning, explicitly telling `getopt' that there are no more + options. + + If OPTS begins with `--', then non-option arguments are treated as + arguments to the option '\0'. This behavior is specific to the GNU + `getopt'. */ + +#if (defined __STDC__ && __STDC__) || defined __cplusplus +# ifdef __GNU_LIBRARY__ /* Many other libraries have conflicting prototypes for getopt, with differences in the consts, in stdlib.h. To avoid compilation errors, only prototype getopt for the GNU C library. */ -extern int getopt (int argc, char *const *argv, const char *shortopts); -#else /* not __GNU_LIBRARY__ */ +extern int getopt (int __argc, char *const *__argv, const char *__shortopts); +# else /* not __GNU_LIBRARY__ */ extern int getopt (); -#endif /* __GNU_LIBRARY__ */ -extern int getopt_long (int argc, char *const *argv, const char *shortopts, - const struct option *longopts, int *longind); -extern int getopt_long_only (int argc, char *const *argv, - const char *shortopts, - const struct option *longopts, int *longind); +# endif /* __GNU_LIBRARY__ */ + +# ifndef __need_getopt +extern int getopt_long (int __argc, char *const *__argv, const char *__shortopts, + const struct option *__longopts, int *__longind); +extern int getopt_long_only (int __argc, char *const *__argv, + const char *__shortopts, + const struct option *__longopts, int *__longind); /* Internal only. Users should not call this directly. */ -extern int _getopt_internal (int argc, char *const *argv, - const char *shortopts, - const struct option *longopts, int *longind, - int long_only); +extern int _getopt_internal (int __argc, char *const *__argv, + const char *__shortopts, + const struct option *__longopts, int *__longind, + int __long_only); +# endif #else /* not __STDC__ */ extern int getopt (); +# ifndef __need_getopt extern int getopt_long (); extern int getopt_long_only (); extern int _getopt_internal (); +# endif #endif /* __STDC__ */ #ifdef __cplusplus } #endif -#endif /* _GETOPT_H */ +/* Make sure we later can get all the definitions and declarations. */ +#undef __need_getopt + +#endif /* getopt.h */ diff --git a/contrib/gnu-sort/lib/getopt1.c b/contrib/gnu-sort/lib/getopt1.c index 3d068e998b5b..22a7efbdd19a 100644 --- a/contrib/gnu-sort/lib/getopt1.c +++ b/contrib/gnu-sort/lib/getopt1.c @@ -1,182 +1,188 @@ /* getopt_long and getopt_long_only entry points for GNU getopt. - Copyright (C) 1987, 88, 89, 90, 91, 92, 1993, 1994 - Free Software Foundation, Inc. + Copyright (C) 1987,88,89,90,91,92,93,94,96,97,98 + Free Software Foundation, Inc. + This file is part of the GNU C Library. - This program is free software; you can redistribute it and/or modify it - under the terms of the GNU General Public License as published by the - Free Software Foundation; either version 2, or (at your option) any - later version. + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. - This program is distributed in the hope that it will be useful, + The GNU C Library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of - MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - GNU General Public License for more details. + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. - You should have received a copy of the GNU General Public License - along with this program; if not, write to the Free Software - Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ - -/* $FreeBSD$ */ + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, write to the Free + Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA + 02111-1307 USA. */ #ifdef HAVE_CONFIG_H #include #endif #include "getopt.h" -#if !defined (__STDC__) || !__STDC__ +#if !defined __STDC__ || !__STDC__ /* This is a separate conditional since some stdc systems reject `defined (const)'. */ #ifndef const #define const #endif #endif #include /* Comment out all this code if we are using the GNU C Library, and are not actually compiling the library itself. This code is part of the GNU C Library, but also included in many other GNU distributions. Compiling and linking in this code is a waste when using the GNU C library (especially if it is a shared library). Rather than having every GNU program understand `configure --with-gnu-libc' and omit the object files, it is simpler to just do this in the source for each such file. */ -#if defined (_LIBC) || !defined (__GNU_LIBRARY__) +#define GETOPT_INTERFACE_VERSION 2 +#if !defined _LIBC && defined __GLIBC__ && __GLIBC__ >= 2 +#include +#if _GNU_GETOPT_INTERFACE_VERSION == GETOPT_INTERFACE_VERSION +#define ELIDE_CODE +#endif +#endif + +#ifndef ELIDE_CODE /* This needs to come after some library #include to get __GNU_LIBRARY__ defined. */ #ifdef __GNU_LIBRARY__ #include -#else -char *getenv (); #endif #ifndef NULL #define NULL 0 #endif int getopt_long (argc, argv, options, long_options, opt_index) int argc; char *const *argv; const char *options; const struct option *long_options; int *opt_index; { return _getopt_internal (argc, argv, options, long_options, opt_index, 0); } /* Like getopt_long, but '-' as well as '--' can indicate a long option. If an option that starts with '-' (not '--') doesn't match a long option, but does match a short option, it is parsed as a short option instead. */ int getopt_long_only (argc, argv, options, long_options, opt_index) int argc; char *const *argv; const char *options; const struct option *long_options; int *opt_index; { return _getopt_internal (argc, argv, options, long_options, opt_index, 1); } -#endif /* _LIBC or not __GNU_LIBRARY__. */ +#endif /* Not ELIDE_CODE. */ #ifdef TEST #include int main (argc, argv) int argc; char **argv; { int c; int digit_optind = 0; while (1) { int this_option_optind = optind ? optind : 1; int option_index = 0; static struct option long_options[] = { {"add", 1, 0, 0}, {"append", 0, 0, 0}, {"delete", 1, 0, 0}, {"verbose", 0, 0, 0}, {"create", 0, 0, 0}, {"file", 1, 0, 0}, {0, 0, 0, 0} }; c = getopt_long (argc, argv, "abc:d:0123456789", long_options, &option_index); - if (c == EOF) + if (c == -1) break; switch (c) { case 0: printf ("option %s", long_options[option_index].name); if (optarg) printf (" with arg %s", optarg); printf ("\n"); break; case '0': case '1': case '2': case '3': case '4': case '5': case '6': case '7': case '8': case '9': if (digit_optind != 0 && digit_optind != this_option_optind) printf ("digits occur in two different argv-elements.\n"); digit_optind = this_option_optind; printf ("option %c\n", c); break; case 'a': printf ("option a\n"); break; case 'b': printf ("option b\n"); break; case 'c': printf ("option c with value `%s'\n", optarg); break; case 'd': printf ("option d with value `%s'\n", optarg); break; case '?': break; default: printf ("?? getopt returned character code 0%o ??\n", c); } } if (optind < argc) { printf ("non-option ARGV-elements: "); while (optind < argc) printf ("%s ", argv[optind++]); printf ("\n"); } exit (0); } #endif /* TEST */ diff --git a/contrib/gnu-sort/lib/gettext.h b/contrib/gnu-sort/lib/gettext.h new file mode 100644 index 000000000000..6d2d4f8ea154 --- /dev/null +++ b/contrib/gnu-sort/lib/gettext.h @@ -0,0 +1,69 @@ +/* Convenience header for conditional use of GNU . + Copyright (C) 1995-1998, 2000-2002 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify it + under the terms of the GNU Library General Public License as published + by the Free Software Foundation; either version 2, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Library General Public License for more details. + + You should have received a copy of the GNU Library General Public + License along with this program; if not, write to the Free Software + Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, + USA. */ + +#ifndef _LIBGETTEXT_H +# define _LIBGETTEXT_H 1 + +/* NLS can be disabled through the configure --disable-nls option. */ +# if ENABLE_NLS + +/* Get declarations of GNU message catalog functions. */ +# include + +# else + +/* Solaris /usr/include/locale.h includes /usr/include/libintl.h, which + chokes if dcgettext is defined as a macro. So include it now, to make + later inclusions of a NOP. We don't include + as well because people using "gettext.h" will not include , + and also including would fail on SunOS 4, whereas + is OK. */ +# if defined(__sun) +# include +# endif + +/* Disabled NLS. + The casts to 'const char *' serve the purpose of producing warnings + for invalid uses of the value returned from these functions. + On pre-ANSI systems without 'const', the config.h file is supposed to + contain "#define const". */ +# define gettext(Msgid) ((const char *) (Msgid)) +# define dgettext(Domainname, Msgid) ((const char *) (Msgid)) +# define dcgettext(Domainname, Msgid, Category) ((const char *) (Msgid)) +# define ngettext(Msgid1, Msgid2, N) \ + ((N) == 1 ? (const char *) (Msgid1) : (const char *) (Msgid2)) +# define dngettext(Domainname, Msgid1, Msgid2, N) \ + ((N) == 1 ? (const char *) (Msgid1) : (const char *) (Msgid2)) +# define dcngettext(Domainname, Msgid1, Msgid2, N, Category) \ + ((N) == 1 ? (const char *) (Msgid1) : (const char *) (Msgid2)) +# define textdomain(Domainname) ((const char *) (Domainname)) +# define bindtextdomain(Domainname, Dirname) ((const char *) (Dirname)) +# define bind_textdomain_codeset(Domainname, Codeset) ((const char *) (Codeset)) + +# endif + +/* A pseudo function call that serves as a marker for the automated + extraction of messages, but does not call gettext(). The run-time + translation is done at a different place in the code. + The argument, String, should be a literal string. Concatenated strings + and other string expressions won't work. + The macro's expansion is not parenthesized, so that it is suitable as + initializer for static 'char[]' or 'const char[]' variables. */ +# define gettext_noop(String) String + +#endif /* _LIBGETTEXT_H */ diff --git a/contrib/gnu-sort/lib/xmalloc.c b/contrib/gnu-sort/lib/xmalloc.c index 2f103d604917..bc6c0bd14ad4 100644 --- a/contrib/gnu-sort/lib/xmalloc.c +++ b/contrib/gnu-sort/lib/xmalloc.c @@ -1,116 +1,116 @@ /* xmalloc.c -- malloc with out of memory checking - Copyright (C) 1990-1999, 2000 Free Software Foundation, Inc. + Copyright (C) 1990-1999, 2000, 2002 Free Software Foundation, Inc. This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ #if HAVE_CONFIG_H # include #endif #include #if STDC_HEADERS # include #else void *calloc (); void *malloc (); void *realloc (); void free (); #endif #if ENABLE_NLS # include # define _(Text) gettext (Text) #else # define textdomain(Domain) # define _(Text) Text #endif #define N_(Text) Text #include "error.h" #include "xalloc.h" #ifndef EXIT_FAILURE # define EXIT_FAILURE 1 #endif -#ifndef HAVE_DONE_WORKING_MALLOC_CHECK -"you must run the autoconf test for a properly working malloc -- see malloc.m4" +#ifndef HAVE_MALLOC +"you must run the autoconf test for a properly working malloc" #endif -#ifndef HAVE_DONE_WORKING_REALLOC_CHECK -"you must run the autoconf test for a properly working realloc --see realloc.m4" +#ifndef HAVE_REALLOC +"you must run the autoconf test for a properly working realloc" #endif /* Exit value when the requested amount of memory is not available. The caller may set it to some other value. */ int xalloc_exit_failure = EXIT_FAILURE; /* If non NULL, call this function when memory is exhausted. */ void (*xalloc_fail_func) PARAMS ((void)) = 0; /* If XALLOC_FAIL_FUNC is NULL, or does return, display this message before exiting when memory is exhausted. Goes through gettext. */ char const xalloc_msg_memory_exhausted[] = N_("memory exhausted"); void xalloc_die (void) { if (xalloc_fail_func) (*xalloc_fail_func) (); error (xalloc_exit_failure, 0, "%s", _(xalloc_msg_memory_exhausted)); /* The `noreturn' cannot be given to error, since it may return if its first argument is 0. To help compilers understand the xalloc_die does terminate, call exit. */ exit (EXIT_FAILURE); } /* Allocate N bytes of memory dynamically, with error checking. */ void * xmalloc (size_t n) { void *p; p = malloc (n); if (p == 0) xalloc_die (); return p; } /* Change the size of an allocated block of memory P to N bytes, with error checking. */ void * xrealloc (void *p, size_t n) { p = realloc (p, n); if (p == 0) xalloc_die (); return p; } /* Allocate memory for N elements of S bytes, with error checking. */ void * xcalloc (size_t n, size_t s) { void *p; p = calloc (n, s); if (p == 0) xalloc_die (); return p; } diff --git a/contrib/gnu-sort/man/sort.1 b/contrib/gnu-sort/man/sort.1 index 919d35512348..9dcbace72740 100644 --- a/contrib/gnu-sort/man/sort.1 +++ b/contrib/gnu-sort/man/sort.1 @@ -1,113 +1,113 @@ -.\" DO NOT MODIFY THIS FILE! It was generated by help2man 1.25. -.TH SORT "1" "February 2002" "sort (textutils) 2.0.21" "User Commands" +.\" DO NOT MODIFY THIS FILE! It was generated by help2man 1.28. +.TH SORT "1" "July 2002" "sort (textutils) 2.1" "User Commands" .SH NAME sort \- sort lines of text files .SH SYNOPSIS .B sort [\fIOPTION\fR]... [\fIFILE\fR]... .SH DESCRIPTION .\" Add any additional description here .PP Write sorted concatenation of all FILE(s) to standard output. .PP Ordering options: .PP Mandatory arguments to long options are mandatory for short options too. .HP \fB\-b\fR, \fB\-\-ignore\-leading\-blanks\fR ignore leading blanks .TP \fB\-d\fR, \fB\-\-dictionary\-order\fR consider only blanks and alphanumeric characters .TP \fB\-f\fR, \fB\-\-ignore\-case\fR fold lower case to upper case characters .TP \fB\-g\fR, \fB\-\-general\-numeric\-sort\fR compare according to general numerical value .TP \fB\-i\fR, \fB\-\-ignore\-nonprinting\fR consider only printable characters .TP \fB\-M\fR, \fB\-\-month\-sort\fR compare (unknown) < `JAN' < ... < `DEC' .TP \fB\-n\fR, \fB\-\-numeric\-sort\fR compare according to string numerical value .TP \fB\-r\fR, \fB\-\-reverse\fR reverse the result of comparisons .PP Other options: .TP \fB\-c\fR, \fB\-\-check\fR check whether input is sorted; do not sort .TP \fB\-k\fR, \fB\-\-key\fR=\fIPOS1[\fR,POS2] start a key at POS1, end it at POS 2 (origin 1) .TP \fB\-m\fR, \fB\-\-merge\fR merge already sorted files; do not sort .TP \fB\-o\fR, \fB\-\-output\fR=\fIFILE\fR write result to FILE instead of standard output .TP \fB\-s\fR, \fB\-\-stable\fR stabilize sort by disabling last-resort comparison .TP \fB\-S\fR, \fB\-\-buffer\-size\fR=\fISIZE\fR use SIZE for main memory buffer .HP \fB\-t\fR, \fB\-\-field\-separator\fR=\fISEP\fR use SEP instead of non- to whitespace transition .TP \fB\-T\fR, \fB\-\-temporary\-directory\fR=\fIDIR\fR use DIR for temporaries, not $TMPDIR or /tmp multiple options specify multiple directories .TP \fB\-u\fR, \fB\-\-unique\fR with \fB\-c\fR: check for strict ordering otherwise: output only the first of an equal run .TP \fB\-z\fR, \fB\-\-zero\-terminated\fR end lines with 0 byte, not newline .TP \fB\-\-help\fR display this help and exit .TP \fB\-\-version\fR output version information and exit .PP POS is F[.C][OPTS], where F is the field number and C the character position in the field. OPTS is one or more single-letter ordering options, which override global ordering options for that key. If no key is given, use the entire line as the key. .PP SIZE may be followed by the following multiplicative suffixes: % 1% of memory, b 1, K 1024 (default), and so on for M, G, T, P, E, Z, Y. .PP With no FILE, or when FILE is -, read standard input. .PP *** WARNING *** The locale specified by the environment affects sort order. Set LC_ALL=C to get the traditional sort order that uses native byte values. .SH AUTHOR Written by Mike Haertel and Paul Eggert. .SH "REPORTING BUGS" Report bugs to . .SH COPYRIGHT Copyright \(co 2002 Free Software Foundation, Inc. .br This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. .SH "SEE ALSO" The full documentation for .B sort is maintained as a Texinfo manual. If the .B info and .B sort programs are properly installed at your site, the command .IP .B info sort .PP should give you access to the complete manual. diff --git a/contrib/gnu-sort/src/sort.c b/contrib/gnu-sort/src/sort.c index a2b0ef16d921..5b75ebe45a92 100644 --- a/contrib/gnu-sort/src/sort.c +++ b/contrib/gnu-sort/src/sort.c @@ -1,2502 +1,2491 @@ /* sort - sort lines of text (with all kinds of options). Copyright (C) 88, 1991-2002 Free Software Foundation, Inc. This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. Written December 1988 by Mike Haertel. The author may be reached (Email) at the address mike@gnu.ai.mit.edu, or (US mail) as Mike Haertel c/o Free Software Foundation. Ørn E. Hansen added NLS support in 1997. */ #include #include #include #include #include #include #include "system.h" #include "long-options.h" #include "error.h" #include "hard-locale.h" #include "human.h" #include "physmem.h" #include "posixver.h" #include "stdio-safer.h" #include "xmemcoll.h" #include "xstrtol.h" #if HAVE_SYS_RESOURCE_H # include #endif #ifndef RLIMIT_DATA struct rlimit { size_t rlim_cur; }; # define getrlimit(Resource, Rlp) (-1) #endif /* The official name of this program (e.g., no `g' prefix). */ #define PROGRAM_NAME "sort" #define AUTHORS N_ ("Mike Haertel and Paul Eggert") -#if defined ENABLE_NLS && HAVE_LANGINFO_H +#if HAVE_LANGINFO_H # include #endif #ifndef SA_NOCLDSTOP # define sigprocmask(How, Set, Oset) /* empty */ # define sigset_t int #endif #ifndef STDC_HEADERS double strtod (); #endif /* Undefine, to avoid warning about redefinition on some systems. */ /* FIXME: Remove these: use MIN/MAX from sys2.h. */ #undef min #define min(a, b) ((a) < (b) ? (a) : (b)) #undef max #define max(a, b) ((a) > (b) ? (a) : (b)) #define UCHAR_LIM (UCHAR_MAX + 1) #define UCHAR(c) ((unsigned char) (c)) #ifndef DEFAULT_TMPDIR # define DEFAULT_TMPDIR "/tmp" #endif /* Use this as exit status in case of error, not EXIT_FAILURE. This is necessary because EXIT_FAILURE is usually 1 and POSIX requires that sort exit with status 1 IFF invoked with -c and the input is not properly sorted. Any other irregular exit must exit with a status code greater than 1. */ #define SORT_FAILURE 2 #define SORT_OUT_OF_ORDER 1 #define C_DECIMAL_POINT '.' #define NEGATION_SIGN '-' #define NUMERIC_ZERO '0' -#ifdef ENABLE_NLS +#if HAVE_SETLOCALE static char decimal_point; static int th_sep; /* if CHAR_MAX + 1, then there is no thousands separator */ /* Nonzero if the corresponding locales are hard. */ static int hard_LC_COLLATE; # if HAVE_NL_LANGINFO static int hard_LC_TIME; # endif # define IS_THOUSANDS_SEP(x) ((x) == th_sep) #else # define decimal_point C_DECIMAL_POINT # define IS_THOUSANDS_SEP(x) 0 #endif #define NONZERO(x) (x != 0) /* The kind of blanks for '-b' to skip in various options. */ enum blanktype { bl_start, bl_end, bl_both }; /* The character marking end of line. Default to \n. */ static int eolchar = '\n'; /* Lines are held in core as counted strings. */ struct line { char *text; /* Text of the line. */ size_t length; /* Length including final newline. */ char *keybeg; /* Start of first key. */ char *keylim; /* Limit of first key. */ }; /* Input buffers. */ struct buffer { char *buf; /* Dynamically allocated buffer, partitioned into 3 regions: - input data; - unused area; - an array of lines, in reverse order. */ size_t used; /* Number of bytes used for input data. */ size_t nlines; /* Number of lines in the line array. */ size_t alloc; /* Number of bytes allocated. */ size_t left; /* Number of bytes left from previous reads. */ size_t line_bytes; /* Number of bytes to reserve for each line. */ int eof; /* An EOF has been read. */ }; struct keyfield { size_t sword; /* Zero-origin 'word' to start at. */ size_t schar; /* Additional characters to skip. */ int skipsblanks; /* Skip leading white space at start. */ size_t eword; /* Zero-origin first word after field. */ size_t echar; /* Additional characters in field. */ int skipeblanks; /* Skip trailing white space at finish. */ int *ignore; /* Boolean array of characters to ignore. */ char *translate; /* Translation applied to characters. */ int numeric; /* Flag for numeric comparison. Handle strings of digits with optional decimal point, but no exponential notation. */ int general_numeric; /* Flag for general, numeric comparison. Handle numbers in exponential notation. */ int month; /* Flag for comparison by month name. */ int reverse; /* Reverse the sense of comparison. */ struct keyfield *next; /* Next keyfield to try. */ }; struct month { char *name; int val; }; /* The name this program was run with. */ char *program_name; /* FIXME: None of these tables work with multibyte character sets. Also, there are many other bugs when handling multibyte characters, or even unibyte encodings where line boundaries are not in the initial shift state. One way to fix this is to rewrite `sort' to use wide characters internally, but doing this with good performance is a bit tricky. */ /* Table of white space. */ static int blanks[UCHAR_LIM]; /* Table of non-printing characters. */ static int nonprinting[UCHAR_LIM]; /* Table of non-dictionary characters (not letters, digits, or blanks). */ static int nondictionary[UCHAR_LIM]; /* Translation table folding lower case to upper. */ static char fold_toupper[UCHAR_LIM]; #define MONTHS_PER_YEAR 12 -#if defined ENABLE_NLS && HAVE_NL_LANGINFO -# define MONTHTAB_CONST /* empty */ -#else -# define MONTHTAB_CONST const -#endif - /* Table mapping month names to integers. Alphabetic order allows binary search. */ -static MONTHTAB_CONST struct month monthtab[] = +static struct month monthtab[] = { {"APR", 4}, {"AUG", 8}, {"DEC", 12}, {"FEB", 2}, {"JAN", 1}, {"JUL", 7}, {"JUN", 6}, {"MAR", 3}, {"MAY", 5}, {"NOV", 11}, {"OCT", 10}, {"SEP", 9} }; /* During the merge phase, the number of files to merge at once. */ #define NMERGE 16 /* Minimum size for a merge or check buffer. */ #define MIN_MERGE_BUFFER_SIZE (2 + sizeof (struct line)) /* Minimum sort size; the code might not work with smaller sizes. */ #define MIN_SORT_SIZE (NMERGE * MIN_MERGE_BUFFER_SIZE) /* The number of bytes needed for a merge or check buffer, which can function relatively efficiently even if it holds only one line. If a longer line is seen, this value is increased. */ static size_t merge_buffer_size = MAX (MIN_MERGE_BUFFER_SIZE, 256 * 1024); /* The approximate maximum number of bytes of main memory to use, as specified by the user. Zero if the user has not specified a size. */ static size_t sort_size; /* The guessed size for non-regular files. */ #define INPUT_FILE_SIZE_GUESS (1024 * 1024) /* Array of directory names in which any temporary files are to be created. */ static char const **temp_dirs; /* Number of temporary directory names used. */ static size_t temp_dir_count; /* Number of allocated slots in temp_dirs. */ static size_t temp_dir_alloc; /* Our process ID. */ static pid_t process_id; /* Flag to reverse the order of all comparisons. */ static int reverse; /* Flag for stable sort. This turns off the last ditch bytewise comparison of lines, and instead leaves lines in the same order they were read if all keys compare equal. */ static int stable; /* Tab character separating fields. If NUL, then fields are separated by the empty string between a non-whitespace character and a whitespace character. */ static char tab; /* Flag to remove consecutive duplicate lines from the output. Only the last of a sequence of equal lines will be output. */ static int unique; /* Nonzero if any of the input files are the standard input. */ static int have_read_stdin; /* List of key field comparisons to be tried. */ static struct keyfield *keylist; void usage (int status) { if (status != 0) fprintf (stderr, _("Try `%s --help' for more information.\n"), program_name); else { printf (_("\ Usage: %s [OPTION]... [FILE]...\n\ "), program_name); fputs (_("\ Write sorted concatenation of all FILE(s) to standard output.\n\ \n\ Ordering options:\n\ \n\ "), stdout); fputs (_("\ Mandatory arguments to long options are mandatory for short options too.\n\ "), stdout); fputs (_("\ -b, --ignore-leading-blanks ignore leading blanks\n\ -d, --dictionary-order consider only blanks and alphanumeric characters\n\ -f, --ignore-case fold lower case to upper case characters\n\ "), stdout); fputs (_("\ -g, --general-numeric-sort compare according to general numerical value\n\ -i, --ignore-nonprinting consider only printable characters\n\ -M, --month-sort compare (unknown) < `JAN' < ... < `DEC'\n\ -n, --numeric-sort compare according to string numerical value\n\ -r, --reverse reverse the result of comparisons\n\ \n\ "), stdout); fputs (_("\ Other options:\n\ \n\ -c, --check check whether input is sorted; do not sort\n\ -k, --key=POS1[,POS2] start a key at POS1, end it at POS 2 (origin 1)\n\ -m, --merge merge already sorted files; do not sort\n\ -o, --output=FILE write result to FILE instead of standard output\n\ -s, --stable stabilize sort by disabling last-resort comparison\n\ -S, --buffer-size=SIZE use SIZE for main memory buffer\n\ "), stdout); printf (_("\ -t, --field-separator=SEP use SEP instead of non- to whitespace transition\n\ -T, --temporary-directory=DIR use DIR for temporaries, not $TMPDIR or %s\n\ multiple options specify multiple directories\n\ -u, --unique with -c: check for strict ordering\n\ otherwise: output only the first of an equal run\n\ "), DEFAULT_TMPDIR); fputs (_("\ -z, --zero-terminated end lines with 0 byte, not newline\n\ "), stdout); fputs (HELP_OPTION_DESCRIPTION, stdout); fputs (VERSION_OPTION_DESCRIPTION, stdout); fputs (_("\ \n\ POS is F[.C][OPTS], where F is the field number and C the character position\n\ in the field. OPTS is one or more single-letter ordering options, which\n\ override global ordering options for that key. If no key is given, use the\n\ entire line as the key.\n\ \n\ SIZE may be followed by the following multiplicative suffixes:\n\ "), stdout); fputs (_("\ % 1% of memory, b 1, K 1024 (default), and so on for M, G, T, P, E, Z, Y.\n\ \n\ With no FILE, or when FILE is -, read standard input.\n\ \n\ *** WARNING ***\n\ The locale specified by the environment affects sort order.\n\ Set LC_ALL=C to get the traditional sort order that uses\n\ native byte values.\n\ "), stdout ); - puts (_("\nReport bugs to .")); + printf (_("\nReport bugs to <%s>.\n"), PACKAGE_BUGREPORT); } /* Don't use EXIT_FAILURE here in case it is defined to be 1. POSIX requires that sort return 1 IFF invoked with -c and the input is not properly sorted. */ assert (status == 0 || status == SORT_FAILURE); exit (status); } #define COMMON_SHORT_OPTIONS "-bcdfgik:mMno:rsS:t:T:uz" static struct option const long_options[] = { {"ignore-leading-blanks", no_argument, NULL, 'b'}, {"check", no_argument, NULL, 'c'}, {"dictionary-order", no_argument, NULL, 'd'}, {"ignore-case", no_argument, NULL, 'f'}, {"general-numeric-sort", no_argument, NULL, 'g'}, {"ignore-nonprinting", no_argument, NULL, 'i'}, {"key", required_argument, NULL, 'k'}, {"merge", no_argument, NULL, 'm'}, {"month-sort", no_argument, NULL, 'M'}, {"numeric-sort", no_argument, NULL, 'n'}, {"output", required_argument, NULL, 'o'}, {"reverse", no_argument, NULL, 'r'}, {"stable", no_argument, NULL, 's'}, {"buffer-size", required_argument, NULL, 'S'}, {"field-separator", required_argument, NULL, 't'}, {"temporary-directory", required_argument, NULL, 'T'}, {"unique", no_argument, NULL, 'u'}, {"zero-terminated", no_argument, NULL, 'z'}, {GETOPT_HELP_OPTION_DECL}, {GETOPT_VERSION_OPTION_DECL}, {0, 0, 0, 0}, }; /* The set of signals that are caught. */ static sigset_t caught_signals; /* The list of temporary files. */ struct tempnode { struct tempnode *volatile next; char name[1]; /* Actual size is 1 + file name length. */ }; static struct tempnode *volatile temphead; /* Clean up any remaining temporary files. */ static void cleanup (void) { struct tempnode *node; for (node = temphead; node; node = node->next) unlink (node->name); } /* Report MESSAGE for FILE, then clean up and exit. */ static void die PARAMS ((char const *, char const *)) ATTRIBUTE_NORETURN; static void die (char const *message, char const *file) { error (0, errno, "%s: %s", message, file); exit (SORT_FAILURE); } /* Create a new temporary file, returning its newly allocated name. Store into *PFP a stream open for writing. */ static char * create_temp_file (FILE **pfp) { static char const slashbase[] = "/sortXXXXXX"; static size_t temp_dir_index; sigset_t oldset; int fd; int saved_errno; char const *temp_dir = temp_dirs[temp_dir_index]; size_t len = strlen (temp_dir); struct tempnode *node = (struct tempnode *) xmalloc (sizeof node->next + len + sizeof slashbase); char *file = node->name; memcpy (file, temp_dir, len); memcpy (file + len, slashbase, sizeof slashbase); node->next = temphead; if (++temp_dir_index == temp_dir_count) temp_dir_index = 0; /* Create the temporary file in a critical section, to avoid races. */ sigprocmask (SIG_BLOCK, &caught_signals, &oldset); fd = mkstemp (file); if (0 <= fd) temphead = node; saved_errno = errno; sigprocmask (SIG_SETMASK, &oldset, NULL); errno = saved_errno; if (fd < 0 || (*pfp = fdopen (fd, "w")) == NULL) die (_("cannot create temporary file"), file); return file; } static FILE * xfopen (const char *file, const char *how) { FILE *fp; if (STREQ (file, "-")) { if (*how == 'r') { have_read_stdin = 1; fp = stdin; } else fp = stdout; } else { if ((fp = fopen_safer (file, how)) == NULL) die (_("open failed"), file); } return fp; } /* Close FP, whose name is FILE, and report any errors. */ static void xfclose (FILE *fp, char const *file) { if (fp == stdin) { /* Allow reading stdin from tty more than once. */ if (feof (fp)) clearerr (fp); } else { if (fclose (fp) != 0) die (_("close failed"), file); } } static void write_bytes (const char *buf, size_t n_bytes, FILE *fp, const char *output_file) { if (fwrite (buf, 1, n_bytes, fp) != n_bytes) die (_("write failed"), output_file); } /* Append DIR to the array of temporary directory names. */ static void add_temp_dir (char const *dir) { if (temp_dir_count == temp_dir_alloc) { temp_dir_alloc = temp_dir_alloc ? temp_dir_alloc * 2 : 16; temp_dirs = xrealloc (temp_dirs, sizeof (temp_dirs) * temp_dir_alloc); } temp_dirs[temp_dir_count++] = dir; } /* Search through the list of temporary files for NAME; remove it if it is found on the list. */ static void zaptemp (const char *name) { struct tempnode *volatile *pnode; struct tempnode *node; for (pnode = &temphead; (node = *pnode); pnode = &node->next) if (node->name == name) { unlink (name); *pnode = node->next; free ((char *) node); break; } } -#if defined ENABLE_NLS && HAVE_NL_LANGINFO +#if HAVE_NL_LANGINFO static int struct_month_cmp (const void *m1, const void *m2) { return strcmp (((const struct month *) m1)->name, ((const struct month *) m2)->name); } #endif /* Initialize the character class tables. */ static void inittables (void) { int i; for (i = 0; i < UCHAR_LIM; ++i) { if (ISBLANK (i)) blanks[i] = 1; if (!ISPRINT (i)) nonprinting[i] = 1; if (!ISALNUM (i) && !ISBLANK (i)) nondictionary[i] = 1; if (ISLOWER (i)) fold_toupper[i] = toupper (i); else fold_toupper[i] = i; } -#if defined ENABLE_NLS && HAVE_NL_LANGINFO +#if HAVE_NL_LANGINFO /* If we're not in the "C" locale, read different names for months. */ if (hard_LC_TIME) { for (i = 0; i < MONTHS_PER_YEAR; i++) { char *s; size_t s_len; size_t j; char *name; s = (char *) nl_langinfo (ABMON_1 + i); s_len = strlen (s); monthtab[i].name = name = (char *) xmalloc (s_len + 1); monthtab[i].val = i + 1; for (j = 0; j < s_len; j++) name[j] = fold_toupper[UCHAR (s[j])]; name[j] = '\0'; } qsort ((void *) monthtab, MONTHS_PER_YEAR, sizeof (struct month), struct_month_cmp); } -#endif /* NLS */ +#endif } /* Specify the amount of main memory to use when sorting. */ static void specify_sort_size (char const *s) { uintmax_t n; char *suffix; enum strtol_error e = xstrtoumax (s, &suffix, 10, &n, "EgGkKmMPtTYZ"); /* The default unit is KiB. */ if (e == LONGINT_OK && ISDIGIT (suffix[-1])) { if (n <= UINTMAX_MAX / 1024) n *= 1024; else e = LONGINT_OVERFLOW; } /* A 'b' suffix means bytes; a '%' suffix means percent of memory. */ if (e == LONGINT_INVALID_SUFFIX_CHAR && ISDIGIT (suffix[-1]) && ! suffix[1]) switch (suffix[0]) { case 'b': e = LONGINT_OK; break; case '%': { double mem = physmem_total () * n / 100; /* Use "<", not "<=", to avoid problems with rounding. */ if (mem < UINTMAX_MAX) { n = mem; e = LONGINT_OK; } else e = LONGINT_OVERFLOW; } break; } if (e == LONGINT_OK) { sort_size = n; if (sort_size == n) { sort_size = MAX (sort_size, MIN_SORT_SIZE); return; } e = LONGINT_OVERFLOW; } STRTOL_FATAL_ERROR (s, _("sort size"), e); } /* Return the default sort size. */ static size_t default_sort_size (void) { /* Let MEM be available memory or 1/8 of total memory, whichever is greater. */ double avail = physmem_available (); double total = physmem_total (); double mem = MAX (avail, total / 8); struct rlimit rlimit; /* Let SIZE be MEM, but no more than the maximum object size or system resource limits. Avoid the MIN macro here, as it is not quite right when only one argument is floating point. Don't bother to check for values like RLIM_INFINITY since in practice they are not much less than SIZE_MAX. */ size_t size = SIZE_MAX; if (mem < size) size = mem; if (getrlimit (RLIMIT_DATA, &rlimit) == 0 && rlimit.rlim_cur < size) size = rlimit.rlim_cur; #ifdef RLIMIT_AS if (getrlimit (RLIMIT_AS, &rlimit) == 0 && rlimit.rlim_cur < size) size = rlimit.rlim_cur; #endif /* Leave a large safety margin for the above limits, as failure can occur when they are exceeded. */ size /= 2; #ifdef RLIMIT_RSS /* Leave a 1/16 margin for RSS to leave room for code, stack, etc. Exceeding RSS is not fatal, but can be quite slow. */ if (getrlimit (RLIMIT_RSS, &rlimit) == 0 && rlimit.rlim_cur / 16 * 15 < size) size = rlimit.rlim_cur / 16 * 15; #endif /* Use no less than the minimum. */ return MAX (size, MIN_SORT_SIZE); } /* Return the sort buffer size to use with the input files identified by FPS and FILES, which are alternate paths to the same files. NFILES gives the number of input files; NFPS may be less. Assume that each input line requires LINE_BYTES extra bytes' worth of line information. Return at most SIZE_BOUND. */ static size_t sort_buffer_size (FILE *const *fps, int nfps, char *const *files, int nfiles, size_t line_bytes, size_t size_bound) { /* In the worst case, each input byte is a newline. */ size_t worst_case_per_input_byte = line_bytes + 1; /* Keep enough room for one extra input line and an extra byte. This extra room might be needed when preparing to read EOF. */ size_t size = worst_case_per_input_byte + 1; int i; for (i = 0; i < nfiles; i++) { struct stat st; off_t file_size; size_t worst_case; if ((i < nfps ? fstat (fileno (fps[i]), &st) : strcmp (files[i], "-") == 0 ? fstat (STDIN_FILENO, &st) : stat (files[i], &st)) != 0) die (_("stat failed"), files[i]); file_size = S_ISREG (st.st_mode) ? st.st_size : INPUT_FILE_SIZE_GUESS; /* Add the amount of memory needed to represent the worst case where the input consists entirely of newlines followed by a single non-newline. Check for overflow. */ worst_case = file_size * worst_case_per_input_byte + 1; if (file_size != worst_case / worst_case_per_input_byte || size_bound - size <= worst_case) return size_bound; size += worst_case; } return size; } /* Initialize BUF. Reserve LINE_BYTES bytes for each line; LINE_BYTES must be at least sizeof (struct line). Allocate ALLOC bytes initially. */ static void initbuf (struct buffer *buf, size_t line_bytes, size_t alloc) { /* Ensure that the line array is properly aligned. If the desired size cannot be allocated, repeatedly halve it until allocation succeeds. The smaller allocation may hurt overall performance, but that's better than failing. */ for (;;) { alloc += sizeof (struct line) - alloc % sizeof (struct line); buf->buf = malloc (alloc); if (buf->buf) break; alloc /= 2; if (alloc <= line_bytes + 1) xalloc_die (); } buf->line_bytes = line_bytes; buf->alloc = alloc; buf->used = buf->left = buf->nlines = 0; buf->eof = 0; } /* Return one past the limit of the line array. */ static inline struct line * buffer_linelim (struct buffer const *buf) { return (struct line *) (buf->buf + buf->alloc); } /* Return a pointer to the first character of the field specified by KEY in LINE. */ static char * begfield (const struct line *line, const struct keyfield *key) { register char *ptr = line->text, *lim = ptr + line->length - 1; register size_t sword = key->sword; register size_t schar = key->schar; if (tab) while (ptr < lim && sword--) { while (ptr < lim && *ptr != tab) ++ptr; if (ptr < lim) ++ptr; } else while (ptr < lim && sword--) { while (ptr < lim && blanks[UCHAR (*ptr)]) ++ptr; while (ptr < lim && !blanks[UCHAR (*ptr)]) ++ptr; } if (key->skipsblanks) while (ptr < lim && blanks[UCHAR (*ptr)]) ++ptr; if (schar < lim - ptr) ptr += schar; else ptr = lim; return ptr; } /* Return the limit of (a pointer to the first character after) the field in LINE specified by KEY. */ static char * limfield (const struct line *line, const struct keyfield *key) { register char *ptr = line->text, *lim = ptr + line->length - 1; register size_t eword = key->eword, echar = key->echar; /* Note: from the POSIX spec: The leading field separator itself is included in a field when -t is not used. FIXME: move this comment up... */ /* Move PTR past EWORD fields or to one past the last byte on LINE, whichever comes first. If there are more than EWORD fields, leave PTR pointing at the beginning of the field having zero-based index, EWORD. If a delimiter character was specified (via -t), then that `beginning' is the first character following the delimiting TAB. Otherwise, leave PTR pointing at the first `blank' character after the preceding field. */ if (tab) while (ptr < lim && eword--) { while (ptr < lim && *ptr != tab) ++ptr; if (ptr < lim && (eword | echar)) ++ptr; } else while (ptr < lim && eword--) { while (ptr < lim && blanks[UCHAR (*ptr)]) ++ptr; while (ptr < lim && !blanks[UCHAR (*ptr)]) ++ptr; } #ifdef POSIX_UNSPECIFIED /* The following block of code makes GNU sort incompatible with standard Unix sort, so it's ifdef'd out for now. The POSIX spec isn't clear on how to interpret this. FIXME: request clarification. From: kwzh@gnu.ai.mit.edu (Karl Heuer) Date: Thu, 30 May 96 12:20:41 -0400 + [Translated to POSIX 1003.1-2001 terminology by Paul Eggert.] [...]I believe I've found another bug in `sort'. $ cat /tmp/sort.in a b c 2 d pq rs 1 t - $ textutils-1.15/src/sort +0.6 -0.7 skipsblanks) while (ptr < lim && blanks[UCHAR (*ptr)]) ++ptr; /* Advance PTR by ECHAR (if possible), but no further than LIM. */ if (echar < lim - ptr) ptr += echar; else ptr = lim; return ptr; } /* FIXME */ static void trim_trailing_blanks (const char *a_start, char **a_end) { while (*a_end > a_start && blanks[UCHAR (*(*a_end - 1))]) --(*a_end); } /* Fill BUF reading from FP, moving buf->left bytes from the end of buf->buf to the beginning first. If EOF is reached and the file wasn't terminated by a newline, supply one. Set up BUF's line table too. FILE is the name of the file corresponding to FP. Return nonzero if some input was read. */ static int fillbuf (struct buffer *buf, register FILE *fp, char const *file) { struct keyfield const *key = keylist; char eol = eolchar; size_t line_bytes = buf->line_bytes; size_t mergesize = merge_buffer_size - MIN_MERGE_BUFFER_SIZE; if (buf->eof) return 0; if (buf->used != buf->left) { memmove (buf->buf, buf->buf + buf->used - buf->left, buf->left); buf->used = buf->left; buf->nlines = 0; } for (;;) { char *ptr = buf->buf + buf->used; struct line *linelim = buffer_linelim (buf); struct line *line = linelim - buf->nlines; size_t avail = (char *) linelim - buf->nlines * line_bytes - ptr; char *line_start = buf->nlines ? line->text + line->length : buf->buf; while (line_bytes + 1 < avail) { /* Read as many bytes as possible, but do not read so many bytes that there might not be enough room for the corresponding line array. The worst case is when the rest of the input file consists entirely of newlines, except that the last byte is not a newline. */ size_t readsize = (avail - 1) / (line_bytes + 1); size_t bytes_read = fread (ptr, 1, readsize, fp); char *ptrlim = ptr + bytes_read; char *p; avail -= bytes_read; if (bytes_read != readsize) { if (ferror (fp)) die (_("read failed"), file); if (feof (fp)) { buf->eof = 1; if (buf->buf == ptrlim) return 0; if (ptrlim[-1] != eol) *ptrlim++ = eol; } } /* Find and record each line in the just-read input. */ while ((p = memchr (ptr, eol, ptrlim - ptr))) { ptr = p + 1; line--; line->text = line_start; line->length = ptr - line_start; mergesize = MAX (mergesize, line->length); avail -= line_bytes; if (key) { /* Precompute the position of the first key for efficiency. */ line->keylim = key->eword == -1 ? p : limfield (line, key); if (key->sword != -1) line->keybeg = begfield (line, key); else { if (key->skipsblanks) while (blanks[UCHAR (*line_start)]) line_start++; line->keybeg = line_start; } if (key->skipeblanks) trim_trailing_blanks (line->keybeg, &line->keylim); } line_start = ptr; } ptr = ptrlim; if (buf->eof) break; } buf->used = ptr - buf->buf; buf->nlines = buffer_linelim (buf) - line; if (buf->nlines != 0) { buf->left = ptr - line_start; merge_buffer_size = mergesize + MIN_MERGE_BUFFER_SIZE; return 1; } /* The current input line is too long to fit in the buffer. Double the buffer size and try again. */ if (2 * buf->alloc < buf->alloc) xalloc_die (); buf->alloc *= 2; buf->buf = xrealloc (buf->buf, buf->alloc); } } /* Compare strings A and B containing decimal fractions < 1. Each string should begin with a decimal point followed immediately by the digits of the fraction. Strings not of this form are considered to be zero. */ /* The goal here, is to take two numbers a and b... compare these in parallel. Instead of converting each, and then comparing the outcome. Most likely stopping the comparison before the conversion is complete. The algorithm used, in the old sort: Algorithm: fraccompare Action : compare two decimal fractions accepts : char *a, char *b returns : -1 if ab. implement: if *a == decimal_point AND *b == decimal_point find first character different in a and b. if both are digits, return the difference *a - *b. if *a is a digit skip past zeros if digit return 1, else 0 if *b is a digit skip past zeros if digit return -1, else 0 if *a is a decimal_point skip past decimal_point and zeros if digit return 1, else 0 if *b is a decimal_point skip past decimal_point and zeros if digit return -1, else 0 return 0 */ static int fraccompare (register const char *a, register const char *b) { if (*a == decimal_point && *b == decimal_point) { while (*++a == *++b) if (! ISDIGIT (*a)) return 0; if (ISDIGIT (*a) && ISDIGIT (*b)) return *a - *b; if (ISDIGIT (*a)) goto a_trailing_nonzero; if (ISDIGIT (*b)) goto b_trailing_nonzero; return 0; } else if (*a++ == decimal_point) { a_trailing_nonzero: while (*a == NUMERIC_ZERO) a++; return ISDIGIT (*a); } else if (*b++ == decimal_point) { b_trailing_nonzero: while (*b == NUMERIC_ZERO) b++; return - ISDIGIT (*b); } return 0; } /* Compare strings A and B as numbers without explicitly converting them to machine numbers. Comparatively slow for short strings, but asymptotically hideously fast. */ static int numcompare (register const char *a, register const char *b) { register int tmpa, tmpb, tmp; register size_t loga, logb; tmpa = *a; tmpb = *b; while (blanks[UCHAR (tmpa)]) tmpa = *++a; while (blanks[UCHAR (tmpb)]) tmpb = *++b; if (tmpa == NEGATION_SIGN) { do tmpa = *++a; while (tmpa == NUMERIC_ZERO || IS_THOUSANDS_SEP (tmpa)); if (tmpb != NEGATION_SIGN) { if (tmpa == decimal_point) do tmpa = *++a; while (tmpa == NUMERIC_ZERO); if (ISDIGIT (tmpa)) return -1; while (tmpb == NUMERIC_ZERO || IS_THOUSANDS_SEP (tmpb)) tmpb = *++b; if (tmpb == decimal_point) do tmpb = *++b; while (tmpb == NUMERIC_ZERO); if (ISDIGIT (tmpb)) return -1; return 0; } do tmpb = *++b; while (tmpb == NUMERIC_ZERO || IS_THOUSANDS_SEP (tmpb)); while (tmpa == tmpb && ISDIGIT (tmpa)) { do tmpa = *++a; while (IS_THOUSANDS_SEP (tmpa)); do tmpb = *++b; while (IS_THOUSANDS_SEP (tmpb)); } if ((tmpa == decimal_point && !ISDIGIT (tmpb)) || (tmpb == decimal_point && !ISDIGIT (tmpa))) return -fraccompare (a, b); tmp = tmpb - tmpa; for (loga = 0; ISDIGIT (tmpa); ++loga) do tmpa = *++a; while (IS_THOUSANDS_SEP (tmpa)); for (logb = 0; ISDIGIT (tmpb); ++logb) do tmpb = *++b; while (IS_THOUSANDS_SEP (tmpb)); if (loga != logb) return loga < logb ? 1 : -1; if (!loga) return 0; return tmp; } else if (tmpb == NEGATION_SIGN) { do tmpb = *++b; while (tmpb == NUMERIC_ZERO || IS_THOUSANDS_SEP (tmpb)); if (tmpb == decimal_point) do tmpb = *++b; while (tmpb == NUMERIC_ZERO); if (ISDIGIT (tmpb)) return 1; while (tmpa == NUMERIC_ZERO || IS_THOUSANDS_SEP (tmpa)) tmpa = *++a; if (tmpa == decimal_point) do tmpa = *++a; while (tmpa == NUMERIC_ZERO); if (ISDIGIT (tmpa)) return 1; return 0; } else { while (tmpa == NUMERIC_ZERO || IS_THOUSANDS_SEP (tmpa)) tmpa = *++a; while (tmpb == NUMERIC_ZERO || IS_THOUSANDS_SEP (tmpb)) tmpb = *++b; while (tmpa == tmpb && ISDIGIT (tmpa)) { do tmpa = *++a; while (IS_THOUSANDS_SEP (tmpa)); do tmpb = *++b; while (IS_THOUSANDS_SEP (tmpb)); } if ((tmpa == decimal_point && !ISDIGIT (tmpb)) || (tmpb == decimal_point && !ISDIGIT (tmpa))) return fraccompare (a, b); tmp = tmpa - tmpb; for (loga = 0; ISDIGIT (tmpa); ++loga) do tmpa = *++a; while (IS_THOUSANDS_SEP (tmpa)); for (logb = 0; ISDIGIT (tmpb); ++logb) do tmpb = *++b; while (IS_THOUSANDS_SEP (tmpb)); if (loga != logb) return loga < logb ? -1 : 1; if (!loga) return 0; return tmp; } } static int general_numcompare (const char *sa, const char *sb) { /* FIXME: add option to warn about failed conversions. */ /* FIXME: maybe add option to try expensive FP conversion only if A and B can't be compared more cheaply/accurately. */ char *ea; char *eb; double a = strtod (sa, &ea); double b = strtod (sb, &eb); /* Put conversion errors at the start of the collating sequence. */ if (sa == ea) return sb == eb ? 0 : -1; if (sb == eb) return 1; /* Sort numbers in the usual way, where -0 == +0. Put NaNs after conversion errors but before numbers; sort them by internal bit-pattern, for lack of a more portable alternative. */ return (a < b ? -1 : a > b ? 1 : a == b ? 0 : b == b ? -1 : a == a ? 1 : memcmp ((char *) &a, (char *) &b, sizeof a)); } /* Return an integer in 1..12 of the month name S with length LEN. Return 0 if the name in S is not recognized. */ static int getmonth (const char *s, size_t len) { char *month; register size_t i; register int lo = 0, hi = MONTHS_PER_YEAR, result; while (len > 0 && blanks[UCHAR (*s)]) { ++s; --len; } if (len == 0) return 0; month = (char *) alloca (len + 1); for (i = 0; i < len; ++i) month[i] = fold_toupper[UCHAR (s[i])]; while (blanks[UCHAR (month[i - 1])]) --i; month[i] = '\0'; do { int ix = (lo + hi) / 2; if (strncmp (month, monthtab[ix].name, strlen (monthtab[ix].name)) < 0) hi = ix; else lo = ix; } while (hi - lo > 1); result = (!strncmp (month, monthtab[lo].name, strlen (monthtab[lo].name)) ? monthtab[lo].val : 0); return result; } /* Compare two lines A and B trying every key in sequence until there are no more keys or a difference is found. */ static int keycompare (const struct line *a, const struct line *b) { struct keyfield *key = keylist; /* For the first iteration only, the key positions have been precomputed for us. */ register char *texta = a->keybeg; register char *textb = b->keybeg; register char *lima = a->keylim; register char *limb = b->keylim; int diff; for (;;) { register unsigned char *translate = (unsigned char *) key->translate; register int *ignore = key->ignore; /* Find the lengths. */ size_t lena = lima <= texta ? 0 : lima - texta; size_t lenb = limb <= textb ? 0 : limb - textb; if (key->skipeblanks) { char *a_end = texta + lena; char *b_end = textb + lenb; trim_trailing_blanks (texta, &a_end); trim_trailing_blanks (textb, &b_end); lena = a_end - texta; lenb = b_end - textb; } /* Actually compare the fields. */ if (key->numeric | key->general_numeric) { char savea = *lima, saveb = *limb; *lima = *limb = '\0'; diff = ((key->numeric ? numcompare : general_numcompare) (texta, textb)); *lima = savea, *limb = saveb; } else if (key->month) diff = getmonth (texta, lena) - getmonth (textb, lenb); -#ifdef ENABLE_NLS /* Sorting like this may become slow, so in a simple locale the user can select a faster sort that is similar to ascii sort */ - else if (hard_LC_COLLATE) + else if (HAVE_SETLOCALE && hard_LC_COLLATE) { if (ignore || translate) { char *copy_a = (char *) alloca (lena + 1 + lenb + 1); char *copy_b = copy_a + lena + 1; size_t new_len_a, new_len_b, i; /* Ignore and/or translate chars before comparing. */ for (new_len_a = new_len_b = i = 0; i < max (lena, lenb); i++) { if (i < lena) { copy_a[new_len_a] = (translate ? translate[UCHAR (texta[i])] : texta[i]); if (!ignore || !ignore[UCHAR (texta[i])]) ++new_len_a; } if (i < lenb) { copy_b[new_len_b] = (translate ? translate[UCHAR (textb[i])] : textb [i]); if (!ignore || !ignore[UCHAR (textb[i])]) ++new_len_b; } } diff = xmemcoll (copy_a, new_len_a, copy_b, new_len_b); } else if (lena == 0) diff = - NONZERO (lenb); else if (lenb == 0) goto greater; else diff = xmemcoll (texta, lena, textb, lenb); } -#endif else if (ignore) { #define CMP_WITH_IGNORE(A, B) \ do \ { \ for (;;) \ { \ while (texta < lima && ignore[UCHAR (*texta)]) \ ++texta; \ while (textb < limb && ignore[UCHAR (*textb)]) \ ++textb; \ if (! (texta < lima && textb < limb)) \ break; \ diff = UCHAR (A) - UCHAR (B); \ if (diff) \ goto not_equal; \ ++texta; \ ++textb; \ } \ \ diff = (texta < lima) - (textb < limb); \ } \ while (0) if (translate) CMP_WITH_IGNORE (translate[UCHAR (*texta)], translate[UCHAR (*textb)]); else CMP_WITH_IGNORE (UCHAR (*texta), UCHAR (*textb)); } else if (lena == 0) diff = - NONZERO (lenb); else if (lenb == 0) goto greater; else { if (translate) { while (texta < lima && textb < limb) { diff = (UCHAR (translate[UCHAR (*texta++)]) - UCHAR (translate[UCHAR (*textb++)])); if (diff) goto not_equal; } } else { diff = memcmp (texta, textb, min (lena, lenb)); if (diff) goto not_equal; } diff = lena < lenb ? -1 : lena != lenb; } if (diff) goto not_equal; key = key->next; if (! key) break; /* Find the beginning and limit of the next field. */ if (key->eword != -1) lima = limfield (a, key), limb = limfield (b, key); else lima = a->text + a->length - 1, limb = b->text + b->length - 1; if (key->sword != -1) texta = begfield (a, key), textb = begfield (b, key); else { texta = a->text, textb = b->text; if (key->skipsblanks) { while (texta < lima && blanks[UCHAR (*texta)]) ++texta; while (textb < limb && blanks[UCHAR (*textb)]) ++textb; } } } return 0; greater: diff = 1; not_equal: return key->reverse ? -diff : diff; } /* Compare two lines A and B, returning negative, zero, or positive depending on whether A compares less than, equal to, or greater than B. */ static int compare (register const struct line *a, register const struct line *b) { int diff; size_t alen, blen; /* First try to compare on the specified keys (if any). The only two cases with no key at all are unadorned sort, and unadorned sort -r. */ if (keylist) { diff = keycompare (a, b); alloca (0); if (diff != 0 || unique || stable) return diff; } /* If the keys all compare equal (or no keys were specified) fall through to the default comparison. */ alen = a->length - 1, blen = b->length - 1; if (alen == 0) diff = - NONZERO (blen); else if (blen == 0) diff = NONZERO (alen); -#ifdef ENABLE_NLS - else if (hard_LC_COLLATE) + else if (HAVE_SETLOCALE && hard_LC_COLLATE) diff = xmemcoll (a->text, alen, b->text, blen); -#endif else if (! (diff = memcmp (a->text, b->text, min (alen, blen)))) diff = alen < blen ? -1 : alen != blen; return reverse ? -diff : diff; } /* Check that the lines read from the given FP come in order. Print a diagnostic (FILE_NAME, line number, contents of line) to stderr and return one if they are not in order. Otherwise, print no diagnostic and return zero. */ static int checkfp (FILE *fp, char *file_name) { struct buffer buf; /* Input buffer. */ struct line temp; /* Copy of previous line. */ size_t alloc = 0; uintmax_t line_number = 0; struct keyfield *key = keylist; int nonunique = 1 - unique; int disordered = 0; initbuf (&buf, sizeof (struct line), MAX (merge_buffer_size, sort_size)); temp.text = NULL; while (fillbuf (&buf, fp, file_name)) { struct line const *line = buffer_linelim (&buf); struct line const *linebase = line - buf.nlines; /* Make sure the line saved from the old buffer contents is less than or equal to the first line of the new buffer. */ if (alloc && nonunique <= compare (&temp, line - 1)) { found_disorder: { char hr_buf[LONGEST_HUMAN_READABLE + 1]; struct line const *disorder_line = line - 1; uintmax_t disorder_line_number = buffer_linelim (&buf) - disorder_line + line_number; fprintf (stderr, _("%s: %s:%s: disorder: "), program_name, file_name, human_readable (disorder_line_number, hr_buf, 1, 1)); write_bytes (disorder_line->text, disorder_line->length, stderr, _("standard error")); disordered = 1; break; } } /* Compare each line in the buffer with its successor. */ while (linebase < --line) if (nonunique <= compare (line, line - 1)) goto found_disorder; line_number += buf.nlines; /* Save the last line of the buffer. */ if (alloc < line->length) { do { alloc *= 2; if (! alloc) { alloc = line->length; break; } } while (alloc < line->length); temp.text = xrealloc (temp.text, alloc); } memcpy (temp.text, line->text, line->length); temp.length = line->length; if (key) { temp.keybeg = temp.text + (line->keybeg - line->text); temp.keylim = temp.text + (line->keylim - line->text); } } xfclose (fp, file_name); free (buf.buf); if (temp.text) free (temp.text); return disordered; } /* Merge lines from FILES onto OFP. NFILES cannot be greater than NMERGE. Close input and output files before returning. OUTPUT_FILE gives the name of the output file; if OFP is NULL, the output file has not been opened yet. */ static void mergefps (char **files, register int nfiles, FILE *ofp, const char *output_file) { FILE *fps[NMERGE]; /* Input streams for each file. */ struct buffer buffer[NMERGE]; /* Input buffers for each file. */ struct line saved; /* Saved line storage for unique check. */ struct line const *savedline = NULL; /* &saved if there is a saved line. */ size_t savealloc = 0; /* Size allocated for the saved line. */ struct line const *cur[NMERGE]; /* Current line in each line table. */ struct line const *base[NMERGE]; /* Base of each line table. */ int ord[NMERGE]; /* Table representing a permutation of fps, such that cur[ord[0]] is the smallest line and will be next output. */ register int i, j, t; struct keyfield *key = keylist; saved.text = NULL; /* Read initial lines from each input file. */ for (i = 0; i < nfiles; ) { fps[i] = xfopen (files[i], "r"); initbuf (&buffer[i], sizeof (struct line), MAX (merge_buffer_size, sort_size / nfiles)); if (fillbuf (&buffer[i], fps[i], files[i])) { struct line const *linelim = buffer_linelim (&buffer[i]); cur[i] = linelim - 1; base[i] = linelim - buffer[i].nlines; i++; } else { /* fps[i] is empty; eliminate it from future consideration. */ xfclose (fps[i], files[i]); zaptemp (files[i]); free (buffer[i].buf); --nfiles; for (j = i; j < nfiles; ++j) files[j] = files[j + 1]; } } if (! ofp) ofp = xfopen (output_file, "w"); /* Set up the ord table according to comparisons among input lines. Since this only reorders two items if one is strictly greater than the other, it is stable. */ for (i = 0; i < nfiles; ++i) ord[i] = i; for (i = 1; i < nfiles; ++i) if (0 < compare (cur[ord[i - 1]], cur[ord[i]])) t = ord[i - 1], ord[i - 1] = ord[i], ord[i] = t, i = 0; /* Repeatedly output the smallest line until no input remains. */ while (nfiles) { struct line const *smallest = cur[ord[0]]; /* If uniquified output is turned on, output only the first of an identical series of lines. */ if (unique) { if (savedline && compare (savedline, smallest)) { savedline = 0; write_bytes (saved.text, saved.length, ofp, output_file); } if (!savedline) { savedline = &saved; if (savealloc < smallest->length) { do if (! savealloc) { savealloc = smallest->length; break; } while ((savealloc *= 2) < smallest->length); saved.text = xrealloc (saved.text, savealloc); } saved.length = smallest->length; memcpy (saved.text, smallest->text, saved.length); if (key) { saved.keybeg = saved.text + (smallest->keybeg - smallest->text); saved.keylim = saved.text + (smallest->keylim - smallest->text); } } } else write_bytes (smallest->text, smallest->length, ofp, output_file); /* Check if we need to read more lines into core. */ if (base[ord[0]] < smallest) cur[ord[0]] = smallest - 1; else { if (fillbuf (&buffer[ord[0]], fps[ord[0]], files[ord[0]])) { struct line const *linelim = buffer_linelim (&buffer[ord[0]]); cur[ord[0]] = linelim - 1; base[ord[0]] = linelim - buffer[ord[0]].nlines; } else { /* We reached EOF on fps[ord[0]]. */ for (i = 1; i < nfiles; ++i) if (ord[i] > ord[0]) --ord[i]; --nfiles; xfclose (fps[ord[0]], files[ord[0]]); zaptemp (files[ord[0]]); free (buffer[ord[0]].buf); for (i = ord[0]; i < nfiles; ++i) { fps[i] = fps[i + 1]; files[i] = files[i + 1]; buffer[i] = buffer[i + 1]; cur[i] = cur[i + 1]; base[i] = base[i + 1]; } for (i = 0; i < nfiles; ++i) ord[i] = ord[i + 1]; continue; } } /* The new line just read in may be larger than other lines already in core; push it back in the queue until we encounter a line larger than it. */ for (i = 1; i < nfiles; ++i) { t = compare (cur[ord[0]], cur[ord[i]]); if (!t) t = ord[0] - ord[i]; if (t < 0) break; } t = ord[0]; for (j = 1; j < i; ++j) ord[j - 1] = ord[j]; ord[i - 1] = t; } if (unique && savedline) { write_bytes (saved.text, saved.length, ofp, output_file); free (saved.text); } xfclose (ofp, output_file); } /* Sort the array LINES with NLINES members, using TEMP for temporary space. The input and output arrays are in reverse order, and LINES and TEMP point just past the end of their respective arrays. */ static void sortlines (struct line *lines, size_t nlines, struct line *temp) { register struct line *lo, *hi, *t; register size_t nlo, nhi; if (nlines == 2) { if (0 < compare (&lines[-1], &lines[-2])) { struct line tmp = lines[-1]; lines[-1] = lines[-2]; lines[-2] = tmp; } return; } nlo = nlines / 2; lo = lines; nhi = nlines - nlo; hi = lines - nlo; if (nlo > 1) sortlines (lo, nlo, temp); if (nhi > 1) sortlines (hi, nhi, temp); t = temp; while (nlo && nhi) if (compare (lo - 1, hi - 1) <= 0) *--t = *--lo, --nlo; else *--t = *--hi, --nhi; while (nlo--) *--t = *--lo; for (lo = lines, nlo = nlines - nhi, t = temp; nlo; --nlo) *--lo = *--t; } /* Return the index of the first of NFILES FILES that is the same file as OUTFILE. If none can be the same, return NFILES. Consider an input pipe to be the same as OUTFILE, since the pipe might be the output of a command like "cat OUTFILE". */ static int first_same_file (char **files, int nfiles, char const *outfile) { int i; int got_outstat = 0; struct stat instat, outstat; for (i = 0; i < nfiles; i++) { int standard_input = STREQ (files[i], "-"); if (STREQ (outfile, files[i]) && ! standard_input) return i; if (! got_outstat) { got_outstat = 1; if ((STREQ (outfile, "-") ? fstat (STDOUT_FILENO, &outstat) : stat (outfile, &outstat)) != 0) return nfiles; } if (((standard_input ? fstat (STDIN_FILENO, &instat) : stat (files[i], &instat)) == 0) && (S_ISFIFO (instat.st_mode) || SAME_INODE (instat, outstat))) return i; } return nfiles; } /* Check that each of the NFILES FILES is ordered. Return a count of disordered files. */ static int check (char **files, int nfiles) { int i, disorders = 0; FILE *fp; for (i = 0; i < nfiles; ++i) { fp = xfopen (files[i], "r"); disorders += checkfp (fp, files[i]); } return disorders; } /* Merge NFILES FILES onto OUTPUT_FILE. However, merge at most MAX_MERGE input files directly onto OUTPUT_FILE. MAX_MERGE cannot exceed NMERGE. */ static void merge (char **files, int nfiles, int max_merge, char const *output_file) { while (max_merge < nfiles) { FILE *tfp; int i, t = 0; char *temp; for (i = 0; i < nfiles / NMERGE; ++i) { temp = create_temp_file (&tfp); mergefps (&files[i * NMERGE], NMERGE, tfp, temp); files[t++] = temp; } temp = create_temp_file (&tfp); mergefps (&files[i * NMERGE], nfiles % NMERGE, tfp, temp); files[t++] = temp; nfiles = t; if (nfiles == 1) break; } mergefps (files, nfiles, NULL, output_file); } /* Sort NFILES FILES onto OUTPUT_FILE. */ static void sort (char **files, int nfiles, char const *output_file) { struct buffer buf; int n_temp_files = 0; int output_file_created = 0; static size_t size; if (! size && ! (size = sort_size)) size = default_sort_size (); buf.alloc = 0; while (nfiles) { char const *temp_output; char const *file = *files; FILE *fp = xfopen (file, "r"); FILE *tfp; if (! buf.alloc) initbuf (&buf, 2 * sizeof (struct line), sort_buffer_size (&fp, 1, files, nfiles, 2 * sizeof (struct line), size)); buf.eof = 0; files++; nfiles--; while (fillbuf (&buf, fp, file)) { struct line *line; struct line *linebase; if (buf.eof && nfiles && (2 * sizeof (struct line) + 1 < (buf.alloc - buf.used - 2 * sizeof (struct line) * buf.nlines))) { /* End of file, but there is more input and buffer room. Concatenate the next input file; this is faster in the usual case. */ buf.left = buf.used; break; } line = buffer_linelim (&buf); linebase = line - buf.nlines; sortlines (line, buf.nlines, linebase); if (buf.eof && !nfiles && !n_temp_files && !buf.left) { xfclose (fp, file); tfp = xfopen (output_file, "w"); temp_output = output_file; output_file_created = 1; } else { ++n_temp_files; temp_output = create_temp_file (&tfp); } do { line--; write_bytes (line->text, line->length, tfp, temp_output); if (unique) while (linebase < line && compare (line, line - 1) == 0) line--; } while (linebase < line); xfclose (tfp, temp_output); if (output_file_created) goto finish; } xfclose (fp, file); } finish: free (buf.buf); if (! output_file_created) { int i = n_temp_files; struct tempnode *node; char **tempfiles = (char **) xmalloc (n_temp_files * sizeof (char *)); for (node = temphead; i > 0; node = node->next) tempfiles[--i] = node->name; merge (tempfiles, n_temp_files, NMERGE, output_file); free ((char *) tempfiles); } } /* Insert key KEY at the end of the key list. */ static void insertkey (struct keyfield *key) { struct keyfield **p; for (p = &keylist; *p; p = &(*p)->next) continue; *p = key; key->next = NULL; } /* Report a bad field specification SPEC, with extra info MSGID. */ static void badfieldspec PARAMS ((char const *, char const *)) ATTRIBUTE_NORETURN; static void badfieldspec (char const *spec, char const *msgid) { error (SORT_FAILURE, 0, _("%s: invalid field specification `%s'"), _(msgid), spec); abort (); } /* Parse the leading integer in STRING and store the resulting value (which must fit into size_t) into *VAL. Return the address of the suffix after the integer. If MSGID is NULL, return NULL after failure; otherwise, report MSGID and exit on failure. */ static char const * parse_field_count (char const *string, size_t *val, char const *msgid) { char *suffix; uintmax_t n; switch (xstrtoumax (string, &suffix, 10, &n, "")) { case LONGINT_OK: case LONGINT_INVALID_SUFFIX_CHAR: *val = n; if (*val == n) break; /* Fall through. */ case LONGINT_OVERFLOW: if (msgid) error (SORT_FAILURE, 0, _("%s: count `%.*s' too large"), _(msgid), (int) (suffix - string), string); return NULL; case LONGINT_INVALID: if (msgid) error (SORT_FAILURE, 0, _("%s: invalid count at start of `%s'"), _(msgid), string); return NULL; } return suffix; } /* Handle interrupts and hangups. */ static void sighandler (int sig) { #ifndef SA_NOCLDSTOP signal (sig, SIG_IGN); #endif cleanup (); #ifdef SA_NOCLDSTOP { struct sigaction sigact; sigact.sa_handler = SIG_DFL; sigemptyset (&sigact.sa_mask); sigact.sa_flags = 0; sigaction (sig, &sigact, NULL); } #else signal (sig, SIG_DFL); #endif kill (process_id, sig); } /* Set the ordering options for KEY specified in S. Return the address of the first character in S that is not a valid ordering option. BLANKTYPE is the kind of blanks that 'b' should skip. */ static char * set_ordering (register const char *s, struct keyfield *key, enum blanktype blanktype) { while (*s) { switch (*s) { case 'b': if (blanktype == bl_start || blanktype == bl_both) key->skipsblanks = 1; if (blanktype == bl_end || blanktype == bl_both) key->skipeblanks = 1; break; case 'd': key->ignore = nondictionary; break; case 'f': key->translate = fold_toupper; break; case 'g': key->general_numeric = 1; break; case 'i': key->ignore = nonprinting; break; case 'M': key->month = 1; break; case 'n': key->numeric = 1; break; case 'r': key->reverse = 1; break; default: return (char *) s; } ++s; } return (char *) s; } static struct keyfield * new_key (void) { struct keyfield *key = (struct keyfield *) xcalloc (1, sizeof *key); key->eword = -1; return key; } int main (int argc, char **argv) { struct keyfield *key; struct keyfield gkey; char const *s; int i; int c = 0; int checkonly = 0, mergeonly = 0, nfiles = 0; int posix_pedantic = (getenv ("POSIXLY_CORRECT") != NULL); bool obsolete_usage = (posix2_version () < 200112); char const *short_options = (obsolete_usage ? COMMON_SHORT_OPTIONS "y::" : COMMON_SHORT_OPTIONS "y:"); char *minus = "-", **files; char const *outfile = minus; static int const sigs[] = { SIGHUP, SIGINT, SIGPIPE, SIGTERM }; int nsigs = sizeof sigs / sizeof *sigs; #ifdef SA_NOCLDSTOP struct sigaction oldact, newact; #endif program_name = argv[0]; process_id = getpid (); setlocale (LC_ALL, ""); bindtextdomain (PACKAGE, LOCALEDIR); textdomain (PACKAGE); atexit (cleanup); -#ifdef ENABLE_NLS - hard_LC_COLLATE = hard_locale (LC_COLLATE); -# if HAVE_NL_LANGINFO +#if HAVE_NL_LANGINFO hard_LC_TIME = hard_locale (LC_TIME); -# endif +#endif +#if HAVE_SETLOCALE /* Let's get locale's representation of the decimal point */ { struct lconv *lconvp = localeconv (); /* If the locale doesn't define a decimal point, or if the decimal point is multibyte, use the C decimal point. We don't support multibyte decimal points yet. */ decimal_point = *lconvp->decimal_point; if (! decimal_point || lconvp->decimal_point[1]) decimal_point = C_DECIMAL_POINT; /* We don't support multibyte thousands separators yet. */ th_sep = *lconvp->thousands_sep; if (! th_sep || lconvp->thousands_sep[1]) th_sep = CHAR_MAX + 1; } - -#endif /* NLS */ +#endif have_read_stdin = 0; inittables (); /* Change the way library functions fail. */ xalloc_exit_failure = SORT_FAILURE; xmemcoll_exit_failure = SORT_FAILURE; #ifdef SA_NOCLDSTOP sigemptyset (&caught_signals); for (i = 0; i < nsigs; i++) sigaddset (&caught_signals, sigs[i]); newact.sa_handler = sighandler; newact.sa_mask = caught_signals; newact.sa_flags = 0; #endif for (i = 0; i < nsigs; i++) { int sig = sigs[i]; #ifdef SA_NOCLDSTOP sigaction (sig, NULL, &oldact); if (oldact.sa_handler != SIG_IGN) sigaction (sig, &newact, NULL); #else if (signal (sig, SIG_IGN) != SIG_IGN) signal (sig, sighandler); #endif } gkey.sword = gkey.eword = -1; gkey.ignore = NULL; gkey.translate = NULL; gkey.numeric = gkey.general_numeric = gkey.month = gkey.reverse = 0; gkey.skipsblanks = gkey.skipeblanks = 0; files = (char **) xmalloc (sizeof (char *) * argc); for (;;) { /* Parse an operand as a file after "--" was seen; or if pedantic and a file was seen, unless the POSIX version predates 1003.1-2001 and -c was not seen and the operand is "-o FILE" or "-oFILE". */ if (c == -1 || (posix_pedantic && nfiles != 0 && ! (obsolete_usage && ! checkonly && optind != argc && argv[optind][0] == '-' && argv[optind][1] == 'o' && (argv[optind][2] || optind + 1 != argc))) || ((c = getopt_long (argc, argv, short_options, long_options, NULL)) == -1)) { if (optind == argc) break; files[nfiles++] = argv[optind++]; } else switch (c) { case 1: key = NULL; if (obsolete_usage && optarg[0] == '+') { /* Treat +POS1 [-POS2] as a key if possible; but silently treat an operand as a file if it is not a valid +POS1. */ key = new_key (); s = parse_field_count (optarg + 1, &key->sword, NULL); if (s && *s == '.') s = parse_field_count (s + 1, &key->schar, NULL); if (! (key->sword | key->schar)) key->sword = -1; if (! s || *set_ordering (s, key, bl_start)) { free (key); key = NULL; } else { if (optind != argc && argv[optind][0] == '-' && ISDIGIT (argv[optind][1])) { char const *optarg1 = argv[optind++]; s = parse_field_count (optarg1 + 1, &key->eword, N_("invalid number after `-'")); if (*s == '.') s = parse_field_count (s + 1, &key->echar, N_("invalid number after `.'")); if (*set_ordering (s, key, bl_end)) badfieldspec (optarg1, N_("stray character in field spec")); } insertkey (key); } } if (! key) files[nfiles++] = optarg; break; case 'b': case 'd': case 'f': case 'g': case 'i': case 'M': case 'n': case 'r': { char str[2]; str[0] = c; str[1] = '\0'; set_ordering (str, &gkey, bl_both); } break; case 'c': checkonly = 1; break; case 'k': key = new_key (); /* Get POS1. */ s = parse_field_count (optarg, &key->sword, N_("invalid number at field start")); if (! key->sword--) { /* Provoke with `sort -k0' */ badfieldspec (optarg, N_("field number is zero")); } if (*s == '.') { s = parse_field_count (s + 1, &key->schar, N_("invalid number after `.'")); if (! key->schar--) { /* Provoke with `sort -k1.0' */ badfieldspec (optarg, N_("character offset is zero")); } } if (! (key->sword | key->schar)) key->sword = -1; s = set_ordering (s, key, bl_start); if (*s != ',') { key->eword = -1; key->echar = 0; } else { /* Get POS2. */ s = parse_field_count (s + 1, &key->eword, N_("invalid number after `,'")); if (! key->eword--) { /* Provoke with `sort -k1,0' */ badfieldspec (optarg, N_("field number is zero")); } if (*s == '.') s = parse_field_count (s + 1, &key->echar, N_("invalid number after `.'")); else { /* `-k 2,3' is equivalent to `+1 -3'. */ key->eword++; } s = set_ordering (s, key, bl_end); } if (*s) badfieldspec (optarg, N_("stray character in field spec")); insertkey (key); break; case 'm': mergeonly = 1; break; case 'o': outfile = optarg; break; case 's': stable = 1; break; case 'S': specify_sort_size (optarg); break; case 't': tab = optarg[0]; if (tab && optarg[1]) { /* Provoke with `sort -txx'. Complain about "multi-character tab" instead of "multibyte tab", so that the diagnostic's wording does not need to be changed once multibyte characters are supported. */ error (SORT_FAILURE, 0, _("multi-character tab `%s'"), optarg); } break; case 'T': add_temp_dir (optarg); break; case 'u': unique = 1; break; case 'y': /* Accept and ignore e.g. -y0 for compatibility with Solaris 2.x through Solaris 7. -y is marked as obsolete starting with Solaris 8. */ break; case 'z': eolchar = 0; break; case_GETOPT_HELP_CHAR; case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS); default: usage (SORT_FAILURE); } } /* Inheritance of global options to individual keys. */ for (key = keylist; key; key = key->next) if (!key->ignore && !key->translate && !key->skipsblanks && !key->reverse && !key->skipeblanks && !key->month && !key->numeric && !key->general_numeric) { key->ignore = gkey.ignore; key->translate = gkey.translate; key->skipsblanks = gkey.skipsblanks; key->skipeblanks = gkey.skipeblanks; key->month = gkey.month; key->numeric = gkey.numeric; key->general_numeric = gkey.general_numeric; key->reverse = gkey.reverse; } if (!keylist && (gkey.ignore || gkey.translate || gkey.skipsblanks || gkey.skipeblanks || gkey.month || gkey.numeric || gkey.general_numeric)) insertkey (&gkey); reverse = gkey.reverse; if (temp_dir_count == 0) { char const *tmp_dir = getenv ("TMPDIR"); add_temp_dir (tmp_dir ? tmp_dir : DEFAULT_TMPDIR); } if (nfiles == 0) { nfiles = 1; files = − } if (checkonly) { if (nfiles > 1) error (SORT_FAILURE, 0, _("extra operand `%s' not allowed with -c"), files[1]); /* POSIX requires that sort return 1 IFF invoked with -c and the input is not properly sorted. */ exit (check (files, nfiles) == 0 ? EXIT_SUCCESS : SORT_OUT_OF_ORDER); } if (mergeonly) { int max_merge = first_same_file (files, MIN (nfiles, NMERGE), outfile); merge (files, nfiles, max_merge, outfile); } else sort (files, nfiles, outfile); if (have_read_stdin && fclose (stdin) == EOF) die (_("close failed"), "-"); exit (EXIT_SUCCESS); } diff --git a/contrib/gnu-sort/src/sys2.h b/contrib/gnu-sort/src/sys2.h index 610c1fd0a06a..0ba910eb9e53 100644 --- a/contrib/gnu-sort/src/sys2.h +++ b/contrib/gnu-sort/src/sys2.h @@ -1,565 +1,641 @@ /* WARNING -- this file is temporary. It is shared between the sh-utils, fileutils, and textutils packages. Once I find a little more time, I'll merge the remaining things in system.h and everything in this file will go back there. */ +#ifndef S_IFMT +# define S_IFMT 0170000 +#endif + #if STAT_MACROS_BROKEN # undef S_ISBLK # undef S_ISCHR # undef S_ISDIR # undef S_ISDOOR # undef S_ISFIFO # undef S_ISLNK +# undef S_ISNAM # undef S_ISMPB # undef S_ISMPC # undef S_ISNWK # undef S_ISREG # undef S_ISSOCK -#endif /* STAT_MACROS_BROKEN. */ +#endif -#ifndef S_IFMT -# define S_IFMT 0170000 + +#ifndef S_ISBLK +# ifdef S_IFBLK +# define S_ISBLK(m) (((m) & S_IFMT) == S_IFBLK) +# else +# define S_ISBLK(m) 0 +# endif #endif -#if !defined(S_ISBLK) && defined(S_IFBLK) -# define S_ISBLK(m) (((m) & S_IFMT) == S_IFBLK) + +#ifndef S_ISCHR +# ifdef S_IFCHR +# define S_ISCHR(m) (((m) & S_IFMT) == S_IFCHR) +# else +# define S_ISCHR(m) 0 +# endif #endif -#if !defined(S_ISCHR) && defined(S_IFCHR) -# define S_ISCHR(m) (((m) & S_IFMT) == S_IFCHR) + +#ifndef S_ISDIR +# ifdef S_IFDIR +# define S_ISDIR(m) (((m) & S_IFMT) == S_IFDIR) +# else +# define S_ISDIR(m) 0 +# endif #endif -#if !defined(S_ISDIR) && defined(S_IFDIR) -# define S_ISDIR(m) (((m) & S_IFMT) == S_IFDIR) + +#ifndef S_ISDOOR /* Solaris 2.5 and up */ +# ifdef S_IFDOOR +# define S_ISDOOR(m) (((m) & S_IFMT) == S_IFDOOR) +# else +# define S_ISDOOR(m) 0 +# endif #endif -#if !defined(S_ISREG) && defined(S_IFREG) -# define S_ISREG(m) (((m) & S_IFMT) == S_IFREG) + +#ifndef S_ISFIFO +# ifdef S_IFIFO +# define S_ISFIFO(m) (((m) & S_IFMT) == S_IFIFO) +# else +# define S_ISFIFO(m) 0 +# endif #endif -#if !defined(S_ISFIFO) && defined(S_IFIFO) -# define S_ISFIFO(m) (((m) & S_IFMT) == S_IFIFO) + +#ifndef S_ISLNK +# ifdef S_IFLNK +# define S_ISLNK(m) (((m) & S_IFMT) == S_IFLNK) +# else +# define S_ISLNK(m) 0 +# endif #endif -#if !defined(S_ISLNK) && defined(S_IFLNK) -# define S_ISLNK(m) (((m) & S_IFMT) == S_IFLNK) + +#ifndef S_ISMPB /* V7 */ +# ifdef S_IFMPB +# define S_ISMPB(m) (((m) & S_IFMT) == S_IFMPB) +# define S_ISMPC(m) (((m) & S_IFMT) == S_IFMPC) +# else +# define S_ISMPB(m) 0 +# define S_ISMPC(m) 0 +# endif #endif -#if !defined(S_ISSOCK) && defined(S_IFSOCK) -# define S_ISSOCK(m) (((m) & S_IFMT) == S_IFSOCK) + +#ifndef S_ISNAM /* Xenix */ +# ifdef S_IFNAM +# define S_ISNAM(m) (((m) & S_IFMT) == S_IFNAM) +# else +# define S_ISNAM(m) 0 +# endif #endif -#if !defined(S_ISMPB) && defined(S_IFMPB) /* V7 */ -# define S_ISMPB(m) (((m) & S_IFMT) == S_IFMPB) -# define S_ISMPC(m) (((m) & S_IFMT) == S_IFMPC) + +#ifndef S_ISNWK /* HP/UX */ +# ifdef S_IFNWK +# define S_ISNWK(m) (((m) & S_IFMT) == S_IFNWK) +# else +# define S_ISNWK(m) 0 +# endif #endif -#if !defined(S_ISNWK) && defined(S_IFNWK) /* HP/UX */ -# define S_ISNWK(m) (((m) & S_IFMT) == S_IFNWK) + +#ifndef S_ISREG +# ifdef S_IFREG +# define S_ISREG(m) (((m) & S_IFMT) == S_IFREG) +# else +# define S_ISREG(m) 0 +# endif #endif -#if !defined(S_ISDOOR) && defined(S_IFDOOR) /* Solaris 2.5 and up */ -# define S_ISDOOR(m) (((m) & S_IFMT) == S_IFDOOR) + +#ifndef S_ISSOCK +# ifdef S_IFSOCK +# define S_ISSOCK(m) (((m) & S_IFMT) == S_IFSOCK) +# else +# define S_ISSOCK(m) 0 +# endif +#endif + + +#ifndef S_TYPEISSEM +# ifdef S_INSEM +# define S_TYPEISSEM(p) (S_ISNAM ((p)->st_mode) && (p)->st_rdev == S_INSEM) +# else +# define S_TYPEISSEM(p) 0 +# endif #endif +#ifndef S_TYPEISSHM +# ifdef S_INSHD +# define S_TYPEISSHM(p) (S_ISNAM ((p)->st_mode) && (p)->st_rdev == S_INSHD) +# else +# define S_TYPEISSHM(p) 0 +# endif +#endif + +#ifndef S_TYPEISMQ +# define S_TYPEISMQ(p) 0 +#endif + + +/* If any of the following are undefined, + define them to their de facto standard values. */ #if !S_ISUID # define S_ISUID 04000 #endif #if !S_ISGID # define S_ISGID 02000 #endif /* S_ISVTX is a common extension to POSIX. */ #ifndef S_ISVTX # define S_ISVTX 01000 #endif #if !S_IRUSR && S_IREAD # define S_IRUSR S_IREAD #endif #if !S_IRUSR # define S_IRUSR 00400 #endif #if !S_IRGRP # define S_IRGRP (S_IRUSR >> 3) #endif #if !S_IROTH # define S_IROTH (S_IRUSR >> 6) #endif #if !S_IWUSR && S_IWRITE # define S_IWUSR S_IWRITE #endif #if !S_IWUSR # define S_IWUSR 00200 #endif #if !S_IWGRP # define S_IWGRP (S_IWUSR >> 3) #endif #if !S_IWOTH # define S_IWOTH (S_IWUSR >> 6) #endif #if !S_IXUSR && S_IEXEC # define S_IXUSR S_IEXEC #endif #if !S_IXUSR # define S_IXUSR 00100 #endif #if !S_IXGRP # define S_IXGRP (S_IXUSR >> 3) #endif #if !S_IXOTH # define S_IXOTH (S_IXUSR >> 6) #endif #if !S_IRWXU # define S_IRWXU (S_IRUSR | S_IWUSR | S_IXUSR) #endif #if !S_IRWXG # define S_IRWXG (S_IRGRP | S_IWGRP | S_IXGRP) #endif #if !S_IRWXO # define S_IRWXO (S_IROTH | S_IWOTH | S_IXOTH) #endif /* S_IXUGO is a common extension to POSIX. */ #if !S_IXUGO # define S_IXUGO (S_IXUSR | S_IXGRP | S_IXOTH) #endif #ifndef S_IRWXUGO # define S_IRWXUGO (S_IRWXU | S_IRWXG | S_IRWXO) #endif /* All the mode bits that can be affected by chmod. */ #define CHMOD_MODE_BITS \ (S_ISUID | S_ISGID | S_ISVTX | S_IRWXU | S_IRWXG | S_IRWXO) #ifdef ST_MTIM_NSEC # define ST_TIME_CMP_NS(a, b, ns) ((a).ns < (b).ns ? -1 : (a).ns > (b).ns) #else # define ST_TIME_CMP_NS(a, b, ns) 0 #endif #define ST_TIME_CMP(a, b, s, ns) \ ((a).s < (b).s ? -1 : (a).s > (b).s ? 1 : ST_TIME_CMP_NS(a, b, ns)) #define ATIME_CMP(a, b) ST_TIME_CMP (a, b, st_atime, st_atim.ST_MTIM_NSEC) #define CTIME_CMP(a, b) ST_TIME_CMP (a, b, st_ctime, st_ctim.ST_MTIM_NSEC) #define MTIME_CMP(a, b) ST_TIME_CMP (a, b, st_mtime, st_mtim.ST_MTIM_NSEC) #ifndef RETSIGTYPE # define RETSIGTYPE void #endif #if __GNUC__ # ifndef alloca # define alloca __builtin_alloca # endif #else # if HAVE_ALLOCA_H # include # else # ifdef _AIX # pragma alloca # else # ifdef _WIN32 # include # include # else # ifndef alloca char *alloca (); # endif # endif # endif # endif #endif #ifdef __DJGPP__ /* We need the declaration of setmode. */ # include /* We need the declaration of __djgpp_set_ctrl_c. */ # include #endif #if HAVE_STDINT_H # include #endif #if HAVE_INTTYPES_H # include /* for the definition of UINTMAX_MAX */ #endif #include /* Jim Meyering writes: "... Some ctype macros are valid only for character codes that isascii says are ASCII (SGI's IRIX-4.0.5 is one such system --when using /bin/cc or gcc but without giving an ansi option). So, all ctype uses should be through macros like ISPRINT... If STDC_HEADERS is defined, then autoconf has verified that the ctype macros don't need to be guarded with references to isascii. ... Defining isascii to 1 should let any compiler worth its salt eliminate the && through constant folding." Bruno Haible adds: "... Furthermore, isupper(c) etc. have an undefined result if c is outside the range -1 <= c <= 255. One is tempted to write isupper(c) with c being of type `char', but this is wrong if c is an 8-bit character >= 128 which gets sign-extended to a negative value. The macro ISUPPER protects against this as well." */ #if STDC_HEADERS || (!defined (isascii) && !HAVE_ISASCII) # define IN_CTYPE_DOMAIN(c) 1 #else # define IN_CTYPE_DOMAIN(c) isascii(c) #endif #ifdef isblank # define ISBLANK(c) (IN_CTYPE_DOMAIN (c) && isblank (c)) #else # define ISBLANK(c) ((c) == ' ' || (c) == '\t') #endif #ifdef isgraph # define ISGRAPH(c) (IN_CTYPE_DOMAIN (c) && isgraph (c)) #else # define ISGRAPH(c) (IN_CTYPE_DOMAIN (c) && isprint (c) && !isspace (c)) #endif /* This is defined in on at least Solaris2.6 systems. */ #undef ISPRINT #define ISPRINT(c) (IN_CTYPE_DOMAIN (c) && isprint (c)) #define ISALNUM(c) (IN_CTYPE_DOMAIN (c) && isalnum (c)) #define ISALPHA(c) (IN_CTYPE_DOMAIN (c) && isalpha (c)) #define ISCNTRL(c) (IN_CTYPE_DOMAIN (c) && iscntrl (c)) #define ISLOWER(c) (IN_CTYPE_DOMAIN (c) && islower (c)) #define ISPUNCT(c) (IN_CTYPE_DOMAIN (c) && ispunct (c)) #define ISSPACE(c) (IN_CTYPE_DOMAIN (c) && isspace (c)) #define ISUPPER(c) (IN_CTYPE_DOMAIN (c) && isupper (c)) #define ISXDIGIT(c) (IN_CTYPE_DOMAIN (c) && isxdigit (c)) #define ISDIGIT_LOCALE(c) (IN_CTYPE_DOMAIN (c) && isdigit (c)) #if STDC_HEADERS # define TOLOWER(Ch) tolower (Ch) # define TOUPPER(Ch) toupper (Ch) #else # define TOLOWER(Ch) (ISUPPER (Ch) ? tolower (Ch) : (Ch)) # define TOUPPER(Ch) (ISLOWER (Ch) ? toupper (Ch) : (Ch)) #endif /* ISDIGIT differs from ISDIGIT_LOCALE, as follows: - Its arg may be any int or unsigned int; it need not be an unsigned char. - It's guaranteed to evaluate its argument exactly once. - It's typically faster. POSIX says that only '0' through '9' are digits. Prefer ISDIGIT to ISDIGIT_LOCALE unless it's important to use the locale's definition of `digit' even when the host does not conform to POSIX. */ #define ISDIGIT(c) ((unsigned) (c) - '0' <= 9) #ifndef PARAMS # if PROTOTYPES # define PARAMS(Args) Args # else # define PARAMS(Args) () # endif #endif /* Take care of NLS matters. */ #if HAVE_LOCALE_H # include -#endif -#if !HAVE_SETLOCALE +#else # define setlocale(Category, Locale) /* empty */ #endif -#if ENABLE_NLS -# include -# if HAVE_GETTEXT && !HAVE_DCGETTEXT && !defined dcgettext -# define dcgettext(Domain, Text, Category) Text -# endif -# define _(Text) gettext (Text) -#else -# undef bindtextdomain -# define bindtextdomain(Domain, Directory) /* empty */ -# undef textdomain -# define textdomain(Domain) /* empty */ -# undef dcgettext -# define dcgettext(Domainname, Text, Category) Text -# define _(Text) Text -#endif -#define N_(Text) Text +#include "gettext.h" + +#define _(msgid) gettext (msgid) +#define N_(msgid) msgid + +#ifndef HAVE_SETLOCALE +# define HAVE_SETLOCALE 0 +#endif #define STREQ(a, b) (strcmp ((a), (b)) == 0) #if !HAVE_DECL_FREE void free (); #endif #if !HAVE_DECL_MALLOC char *malloc (); #endif #if !HAVE_DECL_MEMCHR char *memchr (); #endif #if !HAVE_DECL_REALLOC char *realloc (); #endif #if !HAVE_DECL_STPCPY # ifndef stpcpy char *stpcpy (); # endif #endif #if !HAVE_DECL_STRNDUP char *strndup (); #endif #if !HAVE_DECL_STRSTR char *strstr (); #endif #if !HAVE_DECL_GETENV char *getenv (); #endif #if !HAVE_DECL_LSEEK off_t lseek (); #endif /* This is needed on some AIX systems. */ #if !HAVE_DECL_STRTOUL unsigned long strtoul (); #endif /* This is needed on some AIX systems. */ #if !HAVE_DECL_STRTOULL && HAVE_UNSIGNED_LONG_LONG unsigned long long strtoull (); #endif #if !HAVE_DECL_GETLOGIN char *getlogin (); #endif #if !HAVE_DECL_TTYNAME char *ttyname (); #endif #if !HAVE_DECL_GETEUID uid_t geteuid (); #endif #if !HAVE_DECL_GETPWUID struct passwd *getpwuid (); #endif #if !HAVE_DECL_GETGRGID struct group *getgrgid (); #endif #if !HAVE_DECL_GETUID uid_t getuid (); #endif #include "xalloc.h" #if ! defined HAVE_MEMPCPY && ! defined mempcpy /* Be CAREFUL that there are no side effects in N. */ # define mempcpy(D, S, N) ((void *) ((char *) memcpy (D, S, N) + (N))) #endif /* Include automatically-generated macros for unlocked I/O. */ #include "unlocked-io.h" #define SAME_INODE(Stat_buf_1, Stat_buf_2) \ ((Stat_buf_1).st_ino == (Stat_buf_2).st_ino \ && (Stat_buf_1).st_dev == (Stat_buf_2).st_dev) #define DOT_OR_DOTDOT(Basename) \ (Basename[0] == '.' && (Basename[1] == '\0' \ || (Basename[1] == '.' && Basename[2] == '\0'))) #if SETVBUF_REVERSED # define SETVBUF(Stream, Buffer, Type, Size) \ setvbuf (Stream, Type, Buffer, Size) #else # define SETVBUF(Stream, Buffer, Type, Size) \ setvbuf (Stream, Buffer, Type, Size) #endif /* Factor out some of the common --help and --version processing code. */ /* These enum values cannot possibly conflict with the option values ordinarily used by commands, including CHAR_MAX + 1, etc. Avoid CHAR_MIN - 1, as it may equal -1, the getopt end-of-options value. */ enum { GETOPT_HELP_CHAR = (CHAR_MIN - 2), GETOPT_VERSION_CHAR = (CHAR_MIN - 3) }; #define GETOPT_HELP_OPTION_DECL \ "help", no_argument, 0, GETOPT_HELP_CHAR #define GETOPT_VERSION_OPTION_DECL \ "version", no_argument, 0, GETOPT_VERSION_CHAR #define case_GETOPT_HELP_CHAR \ case GETOPT_HELP_CHAR: \ usage (EXIT_SUCCESS); \ break; #define HELP_OPTION_DESCRIPTION \ _(" --help display this help and exit\n") #define VERSION_OPTION_DESCRIPTION \ _(" --version output version information and exit\n") #include "closeout.h" #include "version-etc.h" #define case_GETOPT_VERSION_CHAR(Program_name, Authors) \ case GETOPT_VERSION_CHAR: \ version_etc (stdout, Program_name, PACKAGE, VERSION, Authors); \ exit (EXIT_SUCCESS); \ break; #ifndef MAX # define MAX(a, b) ((a) > (b) ? (a) : (b)) #endif #ifndef MIN # define MIN(a,b) (((a) < (b)) ? (a) : (b)) #endif #ifndef CHAR_BIT # define CHAR_BIT 8 #endif /* The extra casts work around common compiler bugs. */ #define TYPE_SIGNED(t) (! ((t) 0 < (t) -1)) /* The outer cast is needed to work around a bug in Cray C 5.0.3.0. It is necessary at least when t == time_t. */ #define TYPE_MINIMUM(t) ((t) (TYPE_SIGNED (t) \ ? ~ (t) 0 << (sizeof (t) * CHAR_BIT - 1) : (t) 0)) #define TYPE_MAXIMUM(t) ((t) (~ (t) 0 - TYPE_MINIMUM (t))) /* Upper bound on the string length of an integer converted to string. 302 / 1000 is ceil (log10 (2.0)). Subtract 1 for the sign bit; add 1 for integer division truncation; add 1 more for a minus sign. */ #define INT_STRLEN_BOUND(t) ((sizeof (t) * CHAR_BIT - 1) * 302 / 1000 + 2) #ifndef CHAR_MIN # define CHAR_MIN TYPE_MINIMUM (char) #endif #ifndef CHAR_MAX # define CHAR_MAX TYPE_MAXIMUM (char) #endif #ifndef SCHAR_MIN # define SCHAR_MIN (-1 - SCHAR_MAX) #endif #ifndef SCHAR_MAX # define SCHAR_MAX (CHAR_MAX == UCHAR_MAX ? CHAR_MAX / 2 : CHAR_MAX) #endif #ifndef UCHAR_MAX # define UCHAR_MAX TYPE_MAXIMUM (unsigned char) #endif #ifndef SHRT_MIN # define SHRT_MIN TYPE_MINIMUM (short int) #endif #ifndef SHRT_MAX # define SHRT_MAX TYPE_MAXIMUM (short int) #endif #ifndef INT_MAX # define INT_MAX TYPE_MAXIMUM (int) #endif #ifndef UINT_MAX # define UINT_MAX TYPE_MAXIMUM (unsigned int) #endif #ifndef LONG_MAX # define LONG_MAX TYPE_MAXIMUM (long) #endif #ifndef ULONG_MAX # define ULONG_MAX TYPE_MAXIMUM (unsigned long) #endif #ifndef SIZE_MAX # define SIZE_MAX TYPE_MAXIMUM (size_t) #endif #ifndef UINTMAX_MAX # define UINTMAX_MAX TYPE_MAXIMUM (uintmax_t) #endif #ifndef OFF_T_MIN # define OFF_T_MIN TYPE_MINIMUM (off_t) #endif #ifndef OFF_T_MAX # define OFF_T_MAX TYPE_MAXIMUM (off_t) #endif #ifndef UID_T_MAX # define UID_T_MAX TYPE_MAXIMUM (uid_t) #endif #ifndef GID_T_MAX # define GID_T_MAX TYPE_MAXIMUM (gid_t) #endif #ifndef PID_T_MAX # define PID_T_MAX TYPE_MAXIMUM (pid_t) #endif #ifndef CHAR_BIT # define CHAR_BIT 8 #endif /* Use this to suppress gcc's `...may be used before initialized' warnings. */ #ifdef lint # define IF_LINT(Code) Code #else # define IF_LINT(Code) /* empty */ #endif #ifndef __attribute__ # if __GNUC__ < 2 || (__GNUC__ == 2 && __GNUC_MINOR__ < 8) || __STRICT_ANSI__ # define __attribute__(x) # endif #endif #ifndef ATTRIBUTE_NORETURN # define ATTRIBUTE_NORETURN __attribute__ ((__noreturn__)) #endif #ifndef ATTRIBUTE_UNUSED # define ATTRIBUTE_UNUSED __attribute__ ((__unused__)) #endif #if defined strdupa # define ASSIGN_STRDUPA(DEST, S) \ do { DEST = strdupa(S); } while (0) #else # define ASSIGN_STRDUPA(DEST, S) \ do \ { \ const char *s_ = (S); \ size_t len_ = strlen (s_) + 1; \ char *tmp_dest_ = (char *) alloca (len_); \ DEST = memcpy (tmp_dest_, (s_), len_); \ } \ while (0) #endif #ifndef EOVERFLOW # define EOVERFLOW EINVAL #endif #if ! HAVE_FSEEKO && ! defined fseeko # define fseeko(s, o, w) ((o) == (long) (o) \ ? fseek (s, o, w) \ : (errno = EOVERFLOW, -1)) #endif diff --git a/contrib/gnu-sort/src/system.h b/contrib/gnu-sort/src/system.h index 6e11e44bdce0..a2c08d2f6b82 100644 --- a/contrib/gnu-sort/src/system.h +++ b/contrib/gnu-sort/src/system.h @@ -1,286 +1,292 @@ /* system-dependent definitions for fileutils, textutils, and sh-utils packages. - Copyright (C) 1989, 1991-2001 Free Software Foundation, Inc. + Copyright (C) 1989, 1991-2002 Free Software Foundation, Inc. This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ /* Include sys/types.h before this file. */ +#if 2 <= __GLIBC__ && 2 <= __GLIBC_MINOR__ +# if ! defined _SYS_TYPES_H +you must include before including this file +# endif +#endif + #include -#if !defined(HAVE_MKFIFO) +#if !defined HAVE_MKFIFO # define mkfifo(path, mode) (mknod ((path), (mode) | S_IFIFO, 0)) #endif #if HAVE_SYS_PARAM_H # include #endif /* should be included before any preprocessor test of _POSIX_VERSION. */ #if HAVE_UNISTD_H # include #endif #ifndef STDIN_FILENO # define STDIN_FILENO 0 #endif #ifndef STDOUT_FILENO # define STDOUT_FILENO 1 #endif #ifndef STDERR_FILENO # define STDERR_FILENO 2 #endif #if HAVE_LIMITS_H /* limits.h must come before pathmax.h because limits.h on some systems undefs PATH_MAX, whereas pathmax.h sets PATH_MAX. */ # include #endif #include "pathmax.h" #if TIME_WITH_SYS_TIME # include # include #else # if HAVE_SYS_TIME_H # include # else # include # endif #endif /* Since major is a function on SVR4, we can't use `ifndef major'. */ #if MAJOR_IN_MKDEV # include # define HAVE_MAJOR #endif #if MAJOR_IN_SYSMACROS # include # define HAVE_MAJOR #endif #ifdef major /* Might be defined in sys/types.h. */ # define HAVE_MAJOR #endif #ifndef HAVE_MAJOR # define major(dev) (((dev) >> 8) & 0xff) # define minor(dev) ((dev) & 0xff) # define makedev(maj, min) (((maj) << 8) | (min)) #endif #undef HAVE_MAJOR #if HAVE_UTIME_H # include #endif /* Some systems (even some that do have ) don't declare this structure anywhere. */ #ifndef HAVE_STRUCT_UTIMBUF struct utimbuf { long actime; long modtime; }; #endif /* Don't use bcopy! Use memmove if source and destination may overlap, memcpy otherwise. */ #if HAVE_STRING_H # if !STDC_HEADERS && HAVE_MEMORY_H # include # endif # include #else # include #endif #include #ifndef errno extern int errno; #endif #if HAVE_STDBOOL_H # include #else typedef enum {false = 0, true = 1} bool; #endif #if HAVE_STDLIB_H # define getopt system_getopt # include # undef getopt #endif /* The following test is to work around the gross typo in systems like Sony NEWS-OS Release 4.0C, whereby EXIT_FAILURE is defined to 0, not 1. */ #if !EXIT_FAILURE # undef EXIT_FAILURE # define EXIT_FAILURE 1 #endif #ifndef EXIT_SUCCESS # define EXIT_SUCCESS 0 #endif #if HAVE_FCNTL_H # include #else # include #endif -#if !defined (SEEK_SET) +#if !defined SEEK_SET # define SEEK_SET 0 # define SEEK_CUR 1 # define SEEK_END 2 #endif #ifndef F_OK # define F_OK 0 # define X_OK 1 # define W_OK 2 # define R_OK 4 #endif /* For systems that distinguish between text and binary I/O. O_BINARY is usually declared in fcntl.h */ #if !defined O_BINARY && defined _O_BINARY /* For MSC-compatible compilers. */ # define O_BINARY _O_BINARY # define O_TEXT _O_TEXT #endif #ifdef __BEOS__ /* BeOS 5 has O_BINARY and O_TEXT, but they have no effect. */ # undef O_BINARY # undef O_TEXT #endif #if O_BINARY # ifndef __DJGPP__ # define setmode _setmode # define fileno(_fp) _fileno (_fp) # endif /* not DJGPP */ # define SET_MODE(_f, _m) setmode (_f, _m) # define SET_BINARY(_f) do {if (!isatty(_f)) setmode (_f, O_BINARY);} while (0) # define SET_BINARY2(_f1, _f2) \ do { \ if (!isatty (_f1)) \ { \ setmode (_f1, O_BINARY); \ if (!isatty (_f2)) \ setmode (_f2, O_BINARY); \ } \ } while(0) #else # define SET_MODE(_f, _m) (void)0 # define SET_BINARY(f) (void)0 # define SET_BINARY2(f1,f2) (void)0 # define O_BINARY 0 # define O_TEXT 0 #endif /* O_BINARY */ #if HAVE_DIRENT_H # include # define NLENGTH(direct) (strlen((direct)->d_name)) #else /* not HAVE_DIRENT_H */ # define dirent direct # define NLENGTH(direct) ((direct)->d_namlen) # if HAVE_SYS_NDIR_H # include # endif /* HAVE_SYS_NDIR_H */ # if HAVE_SYS_DIR_H # include # endif /* HAVE_SYS_DIR_H */ # if HAVE_NDIR_H # include # endif /* HAVE_NDIR_H */ #endif /* HAVE_DIRENT_H */ #if CLOSEDIR_VOID /* Fake a return value. */ # define CLOSEDIR(d) (closedir (d), 0) #else # define CLOSEDIR(d) closedir (d) #endif /* Get or fake the disk device blocksize. Usually defined by sys/param.h (if at all). */ #if !defined DEV_BSIZE && defined BSIZE # define DEV_BSIZE BSIZE #endif #if !defined DEV_BSIZE && defined BBSIZE /* SGI */ # define DEV_BSIZE BBSIZE #endif #ifndef DEV_BSIZE # define DEV_BSIZE 4096 #endif /* Extract or fake data from a `struct stat'. ST_BLKSIZE: Preferred I/O blocksize for the file, in bytes. ST_NBLOCKS: Number of blocks in the file, including indirect blocks. ST_NBLOCKSIZE: Size of blocks used when calculating ST_NBLOCKS. */ #ifndef HAVE_STRUCT_STAT_ST_BLOCKS # define ST_BLKSIZE(statbuf) DEV_BSIZE -# if defined(_POSIX_SOURCE) || !defined(BSIZE) /* fileblocks.c uses BSIZE. */ +# if defined _POSIX_SOURCE || !defined BSIZE /* fileblocks.c uses BSIZE. */ # define ST_NBLOCKS(statbuf) \ (S_ISREG ((statbuf).st_mode) \ || S_ISDIR ((statbuf).st_mode) \ ? (statbuf).st_size / ST_NBLOCKSIZE + ((statbuf).st_size % ST_NBLOCKSIZE != 0) : 0) # else /* !_POSIX_SOURCE && BSIZE */ # define ST_NBLOCKS(statbuf) \ (S_ISREG ((statbuf).st_mode) \ || S_ISDIR ((statbuf).st_mode) \ ? st_blocks ((statbuf).st_size) : 0) # endif /* !_POSIX_SOURCE && BSIZE */ #else /* HAVE_STRUCT_STAT_ST_BLOCKS */ /* Some systems, like Sequents, return st_blksize of 0 on pipes. */ # define ST_BLKSIZE(statbuf) ((statbuf).st_blksize > 0 \ ? (statbuf).st_blksize : DEV_BSIZE) -# if defined(hpux) || defined(__hpux__) || defined(__hpux) +# if defined hpux || defined __hpux__ || defined __hpux /* HP-UX counts st_blocks in 1024-byte units. This loses when mixing HP-UX and BSD filesystems with NFS. */ # define ST_NBLOCKSIZE 1024 # else /* !hpux */ -# if defined(_AIX) && defined(_I386) +# if defined _AIX && defined _I386 /* AIX PS/2 counts st_blocks in 4K units. */ # define ST_NBLOCKSIZE (4 * 1024) # else /* not AIX PS/2 */ -# if defined(_CRAY) +# if defined _CRAY # define ST_NBLOCKS(statbuf) \ (S_ISREG ((statbuf).st_mode) \ || S_ISDIR ((statbuf).st_mode) \ ? (statbuf).st_blocks * ST_BLKSIZE(statbuf)/ST_NBLOCKSIZE : 0) # endif /* _CRAY */ # endif /* not AIX PS/2 */ # endif /* !hpux */ #endif /* HAVE_STRUCT_STAT_ST_BLOCKS */ #ifndef ST_NBLOCKS # define ST_NBLOCKS(statbuf) \ (S_ISREG ((statbuf).st_mode) \ || S_ISDIR ((statbuf).st_mode) \ ? (statbuf).st_blocks : 0) #endif #ifndef ST_NBLOCKSIZE # define ST_NBLOCKSIZE 512 #endif #include "sys2.h"