add a new port: textproc/tttcmds

[daichi]

Summary

add a new port: textproc/tttcmds

I have not been able to commit for a while, so I am currently in a reactivate state.
It is necessary to check with the active ports committer. My current mentor is gnn,
but he wishes to have other committers review the ports. Thank you.

Test Plan

% cd textproc/tttcmds
% make install clean
...snip...
% rehash
% which retusel
/usr/local/bin/retusel
% env LANG=C retusel -h
NAME
retu_select - print the specified columns in the specified order

SYNOPSIS
retu_select [-1] [-e] [-hvD] [--] N|N/M|N.k.s|N/M.k.s [N|N/M|
N.k.s|N/M.k.s ...] [file ...]

DESCRIPTION
Print the specified columns in the specified order.
...snip...
% make deinstall

> Deinstalling for tttcmds

> Deinstalling tttcmds-1.0.20180701

Updating database digests format: 100%
Checking integrity... done (0 conflicting)
Deinstallation has been requested for the following 1 packages (of 0 packages in the universe):

Installed packages to be REMOVED:
tttcmds-1.0.20180701

Number of packages to be removed: 1

The operation will free 2 MiB.
[1/1] Deinstalling tttcmds-1.0.20180701...
[1/1] Deleting files for tttcmds-1.0.20180701: 100%
%

[gnn]

Please find a ports person (I don''t commit enough ports) to review this as well please.

[eadler]

misc/tttcmds/Makefile

1 ↗

(On Diff #44749)

we don't add this anymore.

10 ↗

(On Diff #44749)

no need to repeat the command name here. I'm also not quite sure what a "development commands set" is

13 ↗

(On Diff #44749)

no need for this: its implicit in being BSD2CLAUSE (I think - its been a while)

[danfe]

misc/tttcmds/Makefile

7 ↗

(On Diff #44749)

misc is a last resort category; is there really no better, more specific category for it?

10 ↗

(On Diff #44749)

Bad COMMENT: it starts with the software name (which is specifically advised against in the PHB) and does not describe it very well actually. Something along Unix-style data processing commands and library would be better.

17 ↗

(On Diff #44749)

This is the default value for GH_PROJECT and thus should be dropped.

19 ↗

(On Diff #44749)

Why is it -jX unsafe? Can this be fixed? At the very least it should be explained in the commit log or accompanied by the explanatory comment.

misc/tttcmds/distinfo

1 ↗

(On Diff #44749)

How did you generate the distinfo file, by hand? Because it's missing TIMESTAMP line and you won't be able to commit it (would be blocked by the hook).

misc/tttcmds/pkg-descr

6 ↗

(On Diff #44749)

Typo in the word favorite. Also, I think operating systems are more common than operation ones.

8 ↗

(On Diff #44749)

This is redundant: you already define LICENSE in the Makefile, don't repeat yourself.

misc/tttcmds/pkg-plist

112 ↗

(On Diff #44749)

Oh, not even a README file? It's OK, but good software should come with documentation.

[danfe]

misc/tttcmds/Makefile

1 ↗

(On Diff #44749)

We don't, but it's fine.

[daichi]

I changed according to advice of eadler and danfe. Thank you.

[mat]

textproc/tttcmds/Makefile

17

I do not understand that sentence, what does not correspond with what exactly?

[lwhsu]

BTW, it doesn't build on i386: https://gist.github.com/lwhsu/5f98e31230b909ed8019d8cbdd66adb7

[daichi]

I changed according to lwhs' advice. Thank you.

(tttcmds is developed only for amd64. It can not be built on i386.)

[danfe]

textproc/tttcmds/Makefile

14

It's good manners to include ONLY_FOR_ARCHS_REASON, but frankly, I find it hard to believe that a genetic data processing would be inherently amd64-only.

I'll try to build it on i386 and maybe send some patches.

17

I wouldn't be so hard here on @daichi: while the message is not perfect, it gives enough context for those interested in fixing it.

textproc/tttcmds/pkg-plist

114

For a single docfile, you might want to use PORTDOCS variable instead.

[danfe]

In D16095#343714, @lwhsu wrote:

BTW, it doesn't build on i386: https://gist.github.com/lwhsu/5f98e31230b909ed8019d8cbdd66adb7

...
util_millisecond.c:40:23: error: format specifies type 'long' but the argument has type 'time_t' (aka 'int') [-Werror,-Wformat]
        printf("%ld%03ld\n", tv.tv_sec, tv.tv_usec / 1000);
                ~~~          ^~~~~~~~~
                %d

It looks like pretty minor issue; most certainly it can be made arch-agnostic, and ONLY_FOR_ARCH lifted.

[daichi]

I changed according to danfe's advice. Thank you.

[daichi]

Changed to correspond to make - jX build

[mat]

textproc/tttcmds/Makefile

18

You cannot use PORTDOCS without a DOCS option. Please add OPTIONS_DEFINE=DOCS.

[daichi]

Changed Makefile according to mat's advice. Thank you.

[jbeich]

• GNU objcopy (e.g., via USE_BINUTILS) cannot be called after chmod when building as non-root, see error log
• CFLAGS are hardcoded in vendor Makefile, so debugging via ASan or profiling optimized builds isn't possible
• WITH_DEBUG builds are unconditionally stripped as vendor Makefile uses a different variable
• -fstack-protector-strong and -Wno-unused-local-typedef aren't supported by base Clang 3.4.1 on FreeBSD 10.*, see error log
• -Qunused-arguments, -Wmissing-variable-declarations and some -Wno-* are not supported by GCC, see error log and GCC warnings
• -Werror is fragile on compiler upgrades and should probably be hidden behind MAINTAINER_MODE or something similar

textproc/tttcmds/Makefile

12

Any reason you've dropped LICENSE_FILE? BSD2CLAUSE isn't available under /usr/ports/Templates/Licenses/, so when building in LICENSES_ASK=1 mode the dialog window would show a stub instead of license text.

textproc/tttcmds/pkg-plist

38

devel/open-usp-tukubai installs the same file. Maybe define CONFLICTS_INSTALL in both ports.

58

devel/open-usp-tukubai installs the same file. Maybe define CONFLICTS_INSTALL in both ports.

113

USE_LDCONFIG in Makefile maybe required e.g., for LIB_DEPENDS to work. Even portlint -C complains about this.

ld(1) cannot find dynamic library due to missing *.so symlink.

$ echo 'int main() {}' >a.c
$ cc a.c -lttt -L/usr/local/lib -Wl,--verbose | fgrep ttt
attempt to open /usr/local/lib/libttt.so failed
attempt to open /usr/local/lib/libttt.a succeeded

ABI tracking:

$ make stage-qa
[...]
Warning: /usr/ports/textproc/tttcmds/work/stage/usr/local/lib/libttt.so.0 doesn't have a SONAME.
Warning: pkg(8) will not register it as being provided by the port.
Warning: If another port depend on it, pkg will not be able to know where it comes from.
Warning: It is directly in /usr/local/lib, it is probably used by other ports.

114

%%PORTDOCS%% in pkg-plist acts like a conditional i.e., expands to either empty string (DOCS=on) or @comment (DOCS=off). PORTDOCS defined in Makefile automatically registers the files (as if they were in pkg-plist). So, either drop this line or use %%PORTDOCS%%%%DOCSDIR%%/README.md then drop PORTDOCS in Makefile.

$ make -V '${PLIST_SUB:MPORTDOCS*}' WITH=DOCS
PORTDOCS=""
$ make -V '${PLIST_SUB:MPORTDOCS*}' WITHOUT=DOCS
PORTDOCS="@comment "
$ fgrep README $(make -V TMPPLIST)
/usr/local/share/doc/tttcmds/README.md

[daichi]

Thank you for your detailed advice, jbeich. It has been rewritten according to jbeich's advice.

However, the SONAME problem has not been solved. I need your help.

[daichi]

Fixed SONAME problem.

I fixed all issues that were pointed out. I will commit unless there is a problem in particular.
Thank you.

[jbeich]

Looks good but some D16095#344641 issues are not fixed:

• GNU objcopy pulled via USE_BINUTILS or due to non-default PATH (e.g., /usr/local/bin before /usr/bin) still breaks build, see error log
• -Qunused-arguments as used by bundled one-true-awk still breaks USE_GCC builds, see error log

[daichi]

Fixed the problems pointed out by jbeich:

• build error about objcopy and ongs_awk
• build error for building by gcc

[mat]

head/textproc/tttcmds/Makefile

24 ↗

(On Diff #48263)

post-install-DOCS-on:

[daichi]

Mat advice reflected. Thank you.