gnulib.git
11 years agoupdate-copyright: avoid copyright notice date corruption
Jim Meyering [Thu, 3 Jan 2013 09:23:30 +0000 (10:23 +0100)]
update-copyright: avoid copyright notice date corruption

Given a sequence of copyright year numbers in which the final
one was a two-digit number that happened to be a substring of
a preceding four-digit year number, we would mistakenly update
the substring (from two- to four-digit) rather than the two-digit
number at the end, which, combined with the addition of the current
4-digit year number would yield two 5-digit year numbers, e.g.,
here, it would convert the first "99" to "1999, 2013" rather than
the final one:
  1991, 99
  11999, 20131, 1999
* build-aux/update-copyright: Tighten a regexp.
* tests/test-update-copyright.sh: Add a test case to trigger the bug.
Reported by Joseph Myers in
http://thread.gmane.org/gmane.comp.lib.gnulib.bugs/32281

11 years agoregex: omit needless signed-pointer casts
Paul Eggert [Wed, 2 Jan 2013 00:27:46 +0000 (16:27 -0800)]
regex: omit needless signed-pointer casts

* lib/regcomp.c (build_charclass, build_charclass_op):
Use char *, not unsigned char *, for class name and extra.
The char values are always nonnegative so there's no need to
insist on unsigned char * here, and using char * removes the need
for casts.  Reported by Aharon Robbins in
<http://sourceware.org/ml/libc-alpha/2012-12/msg00456.html>.

11 years agoregex: support Gawk, which never uses alloca
Paul Eggert [Wed, 2 Jan 2013 00:05:43 +0000 (16:05 -0800)]
regex: support Gawk, which never uses alloca

* lib/regex_internal.h [!_LIBC && !HAVE_ALLOCA]:
Do not include in this case.  Gawk doesn't supply a substitute
alloca.h and doesn't need one.

11 years agoautoupdate
Karl Berry [Wed, 2 Jan 2013 00:01:39 +0000 (16:01 -0800)]
autoupdate

11 years agoupdate from texinfo
Karl Berry [Tue, 1 Jan 2013 23:51:49 +0000 (15:51 -0800)]
update from texinfo

11 years agoupdate from texinfo
Karl Berry [Tue, 1 Jan 2013 18:15:10 +0000 (10:15 -0800)]
update from texinfo

11 years agoautoupdate
Karl Berry [Tue, 1 Jan 2013 17:25:50 +0000 (09:25 -0800)]
autoupdate

11 years agomissing @item
Karl Berry [Tue, 1 Jan 2013 17:17:08 +0000 (09:17 -0800)]
missing @item

11 years agoautoupdate
Karl Berry [Tue, 1 Jan 2013 17:13:02 +0000 (09:13 -0800)]
autoupdate

11 years agoregex: port __libc_lock_define usage to C89
Paul Eggert [Tue, 1 Jan 2013 15:59:12 +0000 (07:59 -0800)]
regex: port __libc_lock_define usage to C89

* lib/regex_internal.h (__libc_lock_define) [!_LIBC]: Remove.
(struct re_dfa_t): Use #ifdef instead.  '__libc_lock_define (, lock)'
does not conform to C89, as it has an empty macro argument.
Reported by Aharon Robbins in
<http://sourceware.org/ml/libc-alpha/2012-12/msg00456.html>.

11 years agomaint: update all copyright year number ranges
Eric Blake [Tue, 1 Jan 2013 00:50:58 +0000 (00:50 +0000)]
maint: update all copyright year number ranges

Run "make update-copyright".  Compare to commit 1602f0a from last year.

Signed-off-by: Eric Blake <eblake@redhat.com>
11 years agoversion-etc: bump copyright year reported in --version
Eric Blake [Tue, 1 Jan 2013 00:50:12 +0000 (00:50 +0000)]
version-etc: bump copyright year reported in --version

Compare to commit fc9846f from last year.

* lib/version-etc.c (COPYRIGHT_YEAR): Update to 2013.

Signed-off-by: Eric Blake <eblake@redhat.com>
11 years agosigprocmask-tests: skip test if pid is unexpectedly large
Eric Blake [Mon, 31 Dec 2012 23:51:29 +0000 (16:51 -0700)]
sigprocmask-tests: skip test if pid is unexpectedly large

At least mingw64 has 8-byte pid_t but only 4-byte long.  Silent
truncation to int in printing a pid value with %d risks killing the
wrong process.  But rather than try to futz with determining
the maximum pid_t, it is simpler to just cap things by realizing
that this test is already skipped on mingw64, so adding a sanity
check bounds comparison (and hard-coding the result rather than
dragging in headers for INT_MAX) is just as effective at avoiding
theoretical problems with no real loss in test coverage.

* tests/test-sigprocmask.c (main): Add range check.

Signed-off-by: Eric Blake <eblake@redhat.com>
11 years agogit-version-gen: avoid test -z portability glitch
Eric Blake [Mon, 31 Dec 2012 23:01:13 +0000 (16:01 -0700)]
git-version-gen: avoid test -z portability glitch

Autoconf warns that there are some broken shells where 'test -z "$x"'
gives 0 exit status if $x is ')'.  Since some of our strings come
from command-line arguments, and since git-version-gen is run on
end-user machines where sh might be broken, we should be robust to
abuse.  Some of the instances replaced here are provably safe,
and could not confuse even broken 'test', but it is easier to
replace all instances of test -[nz].

* build-aux/git-version-gen: Prefer portable test spelling, since
git-version-gen is run on more than just developer machines.

Signed-off-by: Eric Blake <eblake@redhat.com>
11 years agogit-version-gen: add --fallback option to use if git is not present
Peter Rosin [Fri, 28 Dec 2012 23:45:07 +0000 (00:45 +0100)]
git-version-gen: add --fallback option to use if git is not present

When building in a git checkout, but from a system lacking git, it
is useful to fall back to the version determined when the git
checkout was last used from a system sporting git.

* build-aux/git-version-gen: Add support for the new option --fallback,
which comes into play when there is no $tarball_version_file and
git is not working.
(scriptversion): Update.

Copyright-paperwork-exempt: yes
Signed-off-by: Peter Rosin <peda@lysator.liu.se>
11 years agomaint.mk: handle missing git with more grace
Peter Rosin [Fri, 28 Dec 2012 22:13:17 +0000 (23:13 +0100)]
maint.mk: handle missing git with more grace

* top/maint.mk (no-submodule-changes, public-submodule-commit): Quietly
proceed if git is not present.

Copyright-paperwork-exempt: yes
Signed-off-by: Peter Rosin <peda@lysator.liu.se>
11 years agodup2: work around cygwin bug
Eric Blake [Mon, 31 Dec 2012 18:15:53 +0000 (11:15 -0700)]
dup2: work around cygwin bug

Detected by './gnulib-tool --test dup2 cloexec'.
Reported upstream: http://cygwin.com/ml/cygwin/2012-12/msg00377.html
and fixed already: http://cygwin.com/ml/cygwin-cvs/2012-q4/msg00202.html
but as we want to work with older cygwin, we'll have to carry this
in gnulib for a while.

* m4/dup2.m4 (gl_FUNC_DUP2): Flush out cygwin core dump.
* lib/dup2.c (rpl_dup2): Work around it.
* doc/posix-functions/dup2.texi (dup2): Document it.

Signed-off-by: Eric Blake <eblake@redhat.com>
11 years agoregex: remove unnecessary dependency on localcharset.h
Paul Eggert [Mon, 31 Dec 2012 03:25:19 +0000 (19:25 -0800)]
regex: remove unnecessary dependency on localcharset.h

* lib/regex_internal.h [!_LIBC]: Don't include localcharset.h;
hasn't been needed for years.
* modules/regex (Depends-on): Remove localcharset.

11 years agoregex: revert single-byte change
Paul Eggert [Sun, 30 Dec 2012 22:49:05 +0000 (14:49 -0800)]
regex: revert single-byte change

* lib/regexec.c (check_node_accept_bytes): Revert previous change
to this function.  This was alredy fixed in a different way, at
bdb56bacd57070eced9998569ffe3f3c37ef5964 in the glibc git; see
<http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=510219> and
<http://sourceware.org/bugzilla/show_bug.cgi?id=9697>.

11 years agoregex: simplify based on Gawk version
Paul Eggert [Sun, 30 Dec 2012 22:28:55 +0000 (14:28 -0800)]
regex: simplify based on Gawk version

* lib/regex_internal.c (re_dfa_add_node): Simplify.
Reported by Aharon Robbins in
<http://sourceware.org/ml/libc-alpha/2012-12/msg00456.html>.

11 years agoautoupdate
Karl Berry [Sun, 30 Dec 2012 17:12:05 +0000 (09:12 -0800)]
autoupdate

11 years agoregex: check that pattern char is single-byte
Paul Eggert [Sun, 30 Dec 2012 08:10:51 +0000 (00:10 -0800)]
regex: check that pattern char is single-byte

Reported by Aharon Robbins in
<http://sourceware.org/ml/libc-alpha/2012-12/msg00456.html>.
* lib/regexec.c (check_node_accept_bytes):
Return 0 if the pattern string has a multibyte character here.

11 years agoregex: implement rational ranges
Paul Eggert [Sun, 30 Dec 2012 07:31:08 +0000 (23:31 -0800)]
regex: implement rational ranges

Reported by Aharon Robbins in
<http://sourceware.org/ml/libc-alpha/2012-12/msg00456.html>.
* lib/regcomp.c (build_range_exp) [!_LIBC]:
* lib/regexec.c (check_node_accept_bytes) [!_LIBC]:
Implement rational ranges.

11 years agoregex: avoid redefining __wctype
Paul Eggert [Sun, 30 Dec 2012 07:00:48 +0000 (23:00 -0800)]
regex: avoid redefining __wctype

Reported by Aharon Robbins in
<http://sourceware.org/ml/libc-alpha/2012-12/msg00456.html>.
* lib/regex_internal.h (__wctype, __iswctype) [!_LIBC]:
#undef before defining.

11 years agoregex: port to hosts where malloc (0) == NULL
Paul Eggert [Sun, 30 Dec 2012 06:56:31 +0000 (22:56 -0800)]
regex: port to hosts where malloc (0) == NULL

Reported by Aharon Robbins in
<http://sourceware.org/ml/libc-alpha/2012-12/msg00456.html>.
* lib/regex_internal.c (re_node_set_alloc):
Don't assume that malloc (0) yields nonnull.
* lib/regex_internal.h (MALLOC_0_IS_NONNULL): New macro.
* m4/regex.m4 (gl_PREREQ_REGEX): Require gl_EEMALLOC.
* modules/regex (Files): Add m4/eealloc.m4.

11 years agoregex: port to C89
Paul Eggert [Sun, 30 Dec 2012 06:52:17 +0000 (22:52 -0800)]
regex: port to C89

Reported by Aharon Robbins in
<http://sourceware.org/ml/libc-alpha/2012-12/msg00456.html>.
* lib/regcomp.c (init_word_char): Declaration before statement.

11 years agoregex: merge glibc changes
Paul Eggert [Sun, 30 Dec 2012 05:10:29 +0000 (21:10 -0800)]
regex: merge glibc changes

Also, copy the license wording from glibc.  This simplifies
merging changes.  gnulib-tool will change the wording to GPL as
appropriate, when importing it to other packages.  The only
glibc change made since the last merge, which needs merging, is:
2012-05-24 Andreas Schwab <schwab@linux-m68k.org>
* lib/regex_internal.h (gettext): Remove use of INTUSE.

11 years ago* users.txt: Add Emacs.
Paul Eggert [Sun, 30 Dec 2012 01:38:20 +0000 (17:38 -0800)]
* users.txt: Add Emacs.

11 years agodoc: omit mention of version when not needed
Paul Eggert [Sun, 30 Dec 2012 01:38:00 +0000 (17:38 -0800)]
doc: omit mention of version when not needed

* doc/gnulib-intro.texi (Portability and Application Code):
* doc/gnulib.texi (Brief Overview, Legacy Function Substitutes):
Don't mention particular dates or versions when not necessary, so
that the documentation won't go out of date so quickly.

11 years ago* doc/intprops.texi (Integer Properties): Fix Texinfo typo.
Paul Eggert [Sun, 30 Dec 2012 01:36:15 +0000 (17:36 -0800)]
* doc/intprops.texi (Integer Properties): Fix Texinfo typo.

11 years agoautoupdate
Karl Berry [Sat, 29 Dec 2012 14:24:09 +0000 (06:24 -0800)]
autoupdate

11 years agoautoupdate
Karl Berry [Sat, 29 Dec 2012 14:22:27 +0000 (06:22 -0800)]
autoupdate

11 years agobootstrap: pass --force to autoreconf.
Akim Demaille [Fri, 28 Dec 2012 10:48:55 +0000 (11:48 +0100)]
bootstrap: pass --force to autoreconf.

* build-aux/bootstrap (no_recursive): Remove, replaced by...
(AUTORECONFFLAGS): this new variable.
Add "--force" so that Automake's ylwrap and other such tools
be updated at each bootstrap invocation.
Use it.

11 years agoargp: fix port of port new 'inline' approach to Sun C 5.12 + Solaris 10
Paul Eggert [Fri, 28 Dec 2012 02:56:26 +0000 (18:56 -0800)]
argp: fix port of port new 'inline' approach to Sun C 5.12 + Solaris 10

The earlier patch forgot to update one of the #if conditions, causing
a problem on Debian testing i386 reported by Mats Erik Andersson
<http://lists.gnu.org/archive/html/bug-gnulib/2012-12/msg00124.html>.
* lib/argp-fmtstream.h (__argp_fmtstream_putc, argp_fmtstream_putc)
(__argp_fmtstream_puts, argp_fmtstream_puts)
(__argp_fmtstream_write, argp_fmtstream_write)
[!_LIBC && !__OPTIMIZE__]: Declare as ARGP_FS_EI, not as extern.

11 years ago* doc/gnulib-readme.texi: Minor fixups.
Paul Eggert [Fri, 28 Dec 2012 01:39:58 +0000 (17:39 -0800)]
* doc/gnulib-readme.texi: Minor fixups.

(Portability guidelines): Modernize URLs.  Remove some repetition.
(Indent with spaces not TABs): Reword to avoid too-long lines.
Remove some '@ifset standalone' stuff that isn't used.

11 years ago* doc/gnulib-readme.texi (Portability guidelines): ctype.h, not ctime.h.
Paul Eggert [Thu, 27 Dec 2012 09:20:14 +0000 (01:20 -0800)]
* doc/gnulib-readme.texi (Portability guidelines): ctype.h, not ctime.h.

11 years agoHyphen and dash fixes.
Paul Eggert [Thu, 27 Dec 2012 09:16:25 +0000 (01:16 -0800)]
Hyphen and dash fixes.

11 years agoCorrect name of POSIX.1-2001.
Paul Eggert [Wed, 26 Dec 2012 22:00:23 +0000 (14:00 -0800)]
Correct name of POSIX.1-2001.

* doc/posix-functions/fgetc.texi (fgetc):
* doc/posix-functions/fgets.texi (fgets):
* doc/posix-functions/fread.texi (fread):
* doc/posix-functions/fscanf.texi (fscanf):
* doc/posix-functions/getc.texi (getc):
* doc/posix-functions/getchar.texi (getchar):
* doc/posix-functions/scanf.texi (scanf):
POSIX.1-2001, not POSIX-2001.

11 years agodoc: move README into manual
Paul Eggert [Wed, 26 Dec 2012 21:38:22 +0000 (13:38 -0800)]
doc: move README into manual

* README: Move contents to new file doc/gnulib-readme.texi.
Replace with a one-line summary.
* doc/gnulib.texi (Brief Overview): New section,
with old intro preface.  Include gnulib-readme.texi for contents.
(Philosophy): Rename from "Introduction", since this
section no longer introduces the rest.  Write a new preface.
* doc/gnulib-readme.texi: New file, with the old contents of
README texinfo-ized.  This way, the README info appears
in the online and printed manual.

11 years agoc-xvasprintf: Fix "implicit declaration of function" GCC warning.
Ben Pfaff [Wed, 26 Dec 2012 05:18:14 +0000 (21:18 -0800)]
c-xvasprintf: Fix "implicit declaration of function" GCC warning.

* lib/c-xvasprintf.c: Add missing #include "c-vasprintf.h", for
c_vasprintf() prototype.

11 years agoc-vasprintf: Fix "empty declaration" warning reported by GCC.
Ben Pfaff [Tue, 25 Dec 2012 00:50:37 +0000 (16:50 -0800)]
c-vasprintf: Fix "empty declaration" warning reported by GCC.

* lib/c-vasprintf.h: Remove stray semicolon.

11 years agogettext: avoid obsolete macro AM_PROG_MKDIR_P
Paul Eggert [Mon, 24 Dec 2012 01:55:30 +0000 (17:55 -0800)]
gettext: avoid obsolete macro AM_PROG_MKDIR_P

It is obsolete and is planned to be removed from Automake 1.14; see
<http://lists.gnu.org/archive/html/automake/2012-12/msg00029.html>.
* build-aux/po/Makefile.in.in (install-data, install-data-yes)
(installdirs-data, installdirs-data-yes):
Use $(MKDIR_P), not $(mkdir_p).
* m4/intl.m4 (AM_INTL_SUBDIR):
* m4/po.m4 (AM_PO_SUBDIRS):
Require AC_PROG_MKDIR_P, not AM_PROG_MKDIR_P.

11 years agoautoupdate
Karl Berry [Sun, 23 Dec 2012 13:46:25 +0000 (05:46 -0800)]
autoupdate

11 years agoargp: port new 'inline' approach to Sun C 5.12 + Solaris 10
Paul Eggert [Sat, 22 Dec 2012 23:34:16 +0000 (15:34 -0800)]
argp: port new 'inline' approach to Sun C 5.12 + Solaris 10

On this platform, we are not optimizing but we are using
the substitute for extern inlines, so compile as if
C99-style extern inline, or a substitute, is available.
* lib/argp-fmtstream.h (argp_fmtstream_set_lmargin)
(__argp_fmtstream_set_lmargin, argp_fmtstream_set_rmargin)
(__argp_fmtstream_set_rmargin, argp_fmtstream_set_wmargin)
(__argp_fmtstream_set_wmargin, argp_fmtstream_point)
(__argp_fmtstream_point) [!_LIBC && !__OPTIMIZE__]:
Declare as ARGP_FS_EI, not as extern.
* lib/argp.h (argp_usage, __argp_usage, _option_is_short)
(__option_is_short, _option_is_end, __option_is_end)
[!_LIBC && __USE_EXTERN_INLINES]:
Declare as ARGP_EI, not as extern.

11 years agoFix ChangeLog typo.
Paul Eggert [Fri, 21 Dec 2012 16:15:45 +0000 (08:15 -0800)]
Fix ChangeLog typo.

11 years agoAC_PROG_MKDIR_P: port workaround to pre-2.62 Autoconf
Paul Eggert [Fri, 21 Dec 2012 16:12:48 +0000 (08:12 -0800)]
AC_PROG_MKDIR_P: port workaround to pre-2.62 Autoconf

* m4/gnulib-common.m4 (AC_PROG_MKDIR_P, AC_C_RESTRICT):
Use m4_ifndef([AC_AUTOCONF_VERSION], ...), not
m4_if(m4_version_compare(m4_defn([AC_AUTOCONF_VERSION]),[2.62]),[-1],
...), as the latter is fatal with older Autoconfs.
Problewm reported by Eric Blake in
<http://lists.gnu.org/archive/html/bug-gnulib/2012-12/msg00097.html>.

11 years agoAC_PROG_MKDIR_P: don't workaround if not buggy
Paul Eggert [Fri, 21 Dec 2012 06:10:08 +0000 (22:10 -0800)]
AC_PROG_MKDIR_P: don't workaround if not buggy

* m4/gnulib-common.m4 (AC_PROG_MKDIR_P):
Define only for Autoconf versions before 2.62.
(AC_C_RESTRICT): Use documented AC_AUTOCONF_VERSION, not
undocumented m4_PACKAGE_VERSION, for consistency with the
abovementioned change to AC_PROG_MKDIR_P.  This should suffice
since we're checking for 2.62 or later, and AC_AUTOCONF_VERSION
was introduced in 2.62.

11 years agoNew 'c-*printf' modules for formatted output in C locale.
Ben Pfaff [Wed, 19 Dec 2012 05:06:17 +0000 (21:06 -0800)]
New 'c-*printf' modules for formatted output in C locale.

New module 'c-vasnprintf'.
* modules/c-vasnprintf: New file.
* lib/c-vasnprintf.c: New file.
* lib/c-vasnprintf.h: New file.

New module 'c-snprintf'.
* modules/c-snprintf: New file.
* modules/c-snprintf-tests: New file.
* lib/c-snprintf.c: New file.
* lib/c-snprintf.h: New file.
* tests/test-c-snprintf.c: New file.
* tests/test-c-snprintf.sh: New file.

New module 'c-vsnprintf'.
* modules/c-vsnprintf: New file.
* modules/c-vsnprintf-tests: New file.
* lib/c-vsnprintf.c: New file.
* lib/c-vsnprintf.h: New file.
* tests/test-c-vsnprintf.c: New file.
* tests/test-c-vsnprintf.sh: New file.

New module 'c-vasprintf'.
* modules/c-vasprintf: New file.
* modules/c-vasprintf-tests: New file.
* lib/c-asprintf.c: New file.
* lib/c-vasprintf.c: New file.
* lib/c-vasprintf.h: New file.
* tests/test-c-vasprintf.c  +: New file.
* tests/test-c-vasprintf.sh: New file.

New module 'c-xvasprintf'.
* modules/c-xvasprintf: New file.
* modules/c-xvasprintf-tests: New file.
* lib/c-xasprintf.c: New file.
* lib/c-xvasprintf.c: New file.
* lib/c-xvasprintf.h: New file.
* tests/test-c-xvasprintf.c: New file.
* tests/test-c-xvasprintf.sh: New file.

11 years agoargp: better 'inline'
Paul Eggert [Wed, 19 Dec 2012 01:59:33 +0000 (17:59 -0800)]
argp: better 'inline'

Use extern-inline module to declare extern inline functions.
This avoids some bogus warning diagnostics.  Problem discovered
when modifying GNU tar to use the manywarnings module.
* lib/argp.h, lib/argp-xinl.c (ARGP_EI) [!_LIBC]:
* lib/argp-fmtstream.h, lib/argp-fs-xinl.c (ARGP_FS_EI) [!_LIBC]:
Define based on extern-inline.
* modules/argp (Depends-on): Add extern-inline.

11 years agofilemode, sys_stat: Handle MPX files a la AIX.
Paul Eggert [Tue, 18 Dec 2012 06:15:20 +0000 (22:15 -0800)]
filemode, sys_stat: Handle MPX files a la AIX.

* lib/filemode.c (ftypelet): Report 'm' for MPX files.
* lib/sys_stat.in.h (S_ISMPX): New macro.
* tests/test-sys_stat.c: Add tests for MPX files.

11 years agox-to-1: honor $PERL
Alexandre Duret-Lutz [Wed, 12 Dec 2012 10:05:45 +0000 (11:05 +0100)]
x-to-1: honor $PERL

* build-aux/x-to-1.in: Run $HELP2MAN via $PERL so that the user gets
a chance to use his preferred version of Perl.  This is typically
required by Darwin users whose default /usr/bin/perl does not have all
the libraries required by help2man, and who need to use their MacPorts
installation of Perl instead.

11 years agognu-web-doc-update: add all the new files, even in new directories
Akim Demaille [Fri, 14 Dec 2012 11:58:35 +0000 (12:58 +0100)]
gnu-web-doc-update: add all the new files, even in new directories

See http://lists.gnu.org/archive/html/bug-gnulib/2012-12/msg00057.html

* build-aux/gnu-web-doc-update (--dry-run, $dryrun): New.
Use it.
(main): Don't use cvsutils to get the list of unknown files,
just add all the existing files and directories.

11 years agognu-web-doc-update: improve --help
Akim Demaille [Fri, 14 Dec 2012 11:53:34 +0000 (12:53 +0100)]
gnu-web-doc-update: improve --help

* build-aux/gnu-web-doc-update: Move comments into --help.

11 years agomountlist: recognize more "dummy" file systems
Eric Wong [Fri, 7 Dec 2012 23:14:34 +0000 (23:14 +0000)]
mountlist: recognize more "dummy" file systems

* lib/mountlist.c (ME_DUMMY_0):
  Add these dummy FS names to the list:
  - "debugfs" virtual filesystem for kernel debugging
  - "devpts" PTY slave filesystem
  - "devtmpfs" device filesystem on top of tmpfs/ramfs
  - "fusectl" control filesystem for FUSE
  - "mqueue" enumerates POSIX message queues
  - "rpc_pipefs" kernel <-> userspace bridge for NFS
  - "sysfs" is for exporting kernel objects
  - "devfs" device filesystem for Linux 2.4 and FreeBSD

11 years agoextern-inline: avoid incompatibility with Darwin Libc
Paul Eggert [Tue, 11 Dec 2012 16:41:16 +0000 (08:41 -0800)]
extern-inline: avoid incompatibility with Darwin Libc

* m4/extern-inline.m4 (_GL_INLINE, _GL_EXTERN_INLINE): Do not use
extern inline if __APPLE__.  Use _GL_UNUSED in the non-inline branch.
Problem reported by Akim Demaille in
<http://lists.gnu.org/archive/html/bug-gnulib/2012-12/msg00023.html>.

11 years agognupload: Work with GnuPG using gpg-agent (for smartcards).
Simon Josefsson [Tue, 11 Dec 2012 16:14:34 +0000 (17:14 +0100)]
gnupload: Work with GnuPG using gpg-agent (for smartcards).

* build-aux/gnupload: If GnuPG is configured to use gpg-agent,
let it handle password prompting.

11 years agocanonicalize, canonicalize-lgpl: Microsoft Windows prefix fixes
Eli Zaretskii [Mon, 10 Dec 2012 22:52:22 +0000 (14:52 -0800)]
canonicalize, canonicalize-lgpl: Microsoft Windows prefix fixes

* lib/canonicalize.c (canonicalize_filename_mode):
* lib/canonicalize-lgpl.c (__realpath): Recompute prefix_len after
fetching the current directory.  Don't overrun the beginning of
rpath if there's no slashes after the MS-Windows drive letter.

11 years agomaint.mk: typofix in comments
Stefano Lattarini [Sat, 8 Dec 2012 17:09:18 +0000 (18:09 +0100)]
maint.mk: typofix in comments

11 years agomaint.mk: avoid extra forks
Stefano Lattarini [Sat, 8 Dec 2012 17:06:12 +0000 (18:06 +0100)]
maint.mk: avoid extra forks

* top/maint.mk (_cfg_mk): The GNU make manual documents that
"$(wildcard FILE)" expands to empty if FILE doesn't exist.
So use that instead of "$(shell test -f FILE && echo FILE)".

11 years agovasnprintf: fix ASCII_ONLY typo
Paul Eggert [Sat, 8 Dec 2012 03:31:34 +0000 (19:31 -0800)]
vasnprintf: fix ASCII_ONLY typo

* lib/unistdio/u8-vasnprintf.c (FCHAR_T_ONLY_ASCII):
* lib/unistdio/u16-vasnprintf.c (FCHAR_T_ONLY_ASCII):
* lib/unistdio/u32-vasnprintf.c (FCHAR_T_ONLY_ASCII):
New macro, replacing ASCII_ONLY.  This fixes a typo.  See thread at
<http://lists.gnu.org/archive/html/bug-gnulib/2012-12/msg00021.html>.

11 years agoautoupdate
Karl Berry [Fri, 7 Dec 2012 19:26:51 +0000 (11:26 -0800)]
autoupdate

11 years agolist, oset, xlist, xoset: fix extern inline issue with C99
Paul Eggert [Wed, 5 Dec 2012 21:20:25 +0000 (13:20 -0800)]
list, oset, xlist, xoset: fix extern inline issue with C99

This was introduced by my recent changes for 'inline'.
Problem reported for gettext by Daiki Ueno in
<http://lists.gnu.org/archive/html/bug-gnulib/2012-12/msg00000.html>.
* lib/gl_list.h (gl_list_nx_create_empty, gl_list_create)
(gl_list_nx_create, gl_list_size, gl_list_node_value)
(gl_list_node_set_value, gl_list_node_nx_set_value, gl_list_next_node)
(gl_list_previous_node, gl_list_get_at)
(gl_list_nx_set_at, gl_list_search, gl_list_search_from)
(gl_list_search_from_to, gl_list_indexof, gl_list_indexof_from)
(gl_list_indexof_from_to, gl_list_nx_add_first, gl_list_nx_add_last)
(gl_list_nx_add_before, gl_list_nx_add_after, gl_list_nx_add_at)
(gl_list_remove_node, gl_list_remove_at, gl_list_remove, gl_list_free)
(gl_list_iterator, gl_list_iterator_from_to, gl_list_iterator_next)
(gl_list_iterator_free, gl_sortedlist_search)
(gl_sortedlist_search_from_to, gl_sortedlist_indexof)
(gl_sortedlist_indexof_from_to, gl_sortedlist_add, gl_sortedlist_nx_add)
(gl_sortedlist_remove):
* lib/gl_oset.h (go_oset_nx_create_empty, gl_oset_size, gl_oset_search)
(gl_oset_search_atleast, gl_oset_nx_add, gl_oset_remove, gl_oset_free)
(gl_oset_iterator, gl_oset_iterator_next, gl_oset_iterator_free):
* lib/gl_xlist.h (gl_list_create_empty, gl_list_create)
(gl_list_node_set_value, gl_list_set_at, gl_list_add_first)
(gl_list_add_last, gl_list_add_before, gl_list_add_after)
(gl_list_add_at, gl_sortedlist_add):
* lib/gl_xoset.h (gl_oset_create_empty, gl_oset_add):
Wrap these extern decls inside "#if 0", because they are implemented
as inline functions, and extern inline is not what's wanted here.
It would simplify these .h files to remove the extern decls entirely,
although a downside would be less-clear separation between
specification and implementation.

11 years agoautoupdate
Karl Berry [Tue, 4 Dec 2012 19:04:36 +0000 (11:04 -0800)]
autoupdate

11 years agosys_stat: no 'static inline'
Paul Eggert [Wed, 21 Nov 2012 07:13:14 +0000 (23:13 -0800)]
sys_stat: no 'static inline'

* lib/sys_stat.in.h (rpl_mkdir): Now static, not static inline.
* m4/sys_stat_h.m4 (gl_HEADER_SYS_STAT_H): Do not require AC_C_INLINE.

11 years agoextern-inline: no 'static inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:10 +0000 (22:25 -0800)]
extern-inline: no 'static inline'

* m4/extern-inline.m4 (gl_EXTERN_INLINE):
Do not require AC_C_INLINE.
(_GL_INLINE, _GL_EXTERN_INLINE): Define as 'static', not as
'static inline', for older compilers.

11 years agosnippet/warn-on-use: no 'static inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:10 +0000 (22:25 -0800)]
snippet/warn-on-use: no 'static inline'

* build-aux/snippet/warn-on-use.h:
Remove unnecessary 'inline' in comment.

11 years agorbtree-list, rbtreehash-list: no 'static inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:10 +0000 (22:25 -0800)]
rbtree-list, rbtreehash-list: no 'static inline'

* lib/gl_anyrbtree_list2.h (rotate_left, rotate_right):
* lib/gl_anytree_list2.h (node_at):
* lib/gl_anytreehash_list1.h (hash_resize_after_add)
(gl_oset_first, add_nodes_to_buckets):
Now static, not static inline.

11 years agoregex: no 'static inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:10 +0000 (22:25 -0800)]
regex: no 'static inline'

* lib/regex_internal.c (calc_state_hash):
* lib/regex_internal.h (bitset_set, bitset_clear, bitset_contain)
(bitset_empty, bitset_set_all, bitset_copy, bitset_not, bitset_merge)
(bitset_mask, re_string_char_size_at, re_string_wchar_at):
Now static, not static inline.
(inline) [__GNUC__ < 3 && _LIBC]:
Remove macro; no longer needed.

11 years agoxvasprintf: no 'static inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:10 +0000 (22:25 -0800)]
xvasprintf: no 'static inline'

* lib/xvasprintf.c (xstrcat):
Now static, not static inline.
* m4/xvasprintf.m4 (gl_XVASPRINTF):
Do not require AC_C_INLINE.

11 years agoparse-datetime, parse-duration: no 'static inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:09 +0000 (22:25 -0800)]
parse-datetime, parse-duration: no 'static inline'

* lib/parse-datetime.y (to_uchar):
* lib/parse-duration.c (str_const_to_ul, str_const_to_l)
(scale_n_add):
Now static, not static inline.
* m4/parse-datetime.m4 (gl_PARSE_DATETIME):
* modules/parse-duration (configure.ac):
Do not require AC_C_INLINE.

11 years agogetaddrinfo: no 'static inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:09 +0000 (22:25 -0800)]
getaddrinfo: no 'static inline'

* lib/getaddrinfo.c (validate_family):
Now static, not static inline.
* m4/getaddrinfo.m4 (gl_PREREQ_GETADDRINFO):
Do not require AC_C_INLINE.

11 years agoftruncate, fts, lstat, openat, raise: no 'static inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:09 +0000 (22:25 -0800)]
ftruncate, fts, lstat, openat, raise: no 'static inline'

* lib/ftruncate.c (chsize_nothrow):
* lib/fts.c (opendirat, diropen):
* lib/lstat.c (orig_lstat):
* lib/openat.c (orig_openat):
* lib/raise.c (raise_nothrow):
Now static, not static inline.
* m4/ftruncate.m4 (gl_PREREQ_FTRUNCATE):
* m4/fts.m4 (gl_FUNC_FTS_CORE):
* m4/lstat.m4 (gl_PREREQ_LSTAT):
* m4/openat.m4 (gl_PREREQ_OPENAT):
* m4/raise.m4 (gl_PREREQ_RAISE):
Do not require AC_C_INLINE.

11 years agofflush, stat: no 'static inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:09 +0000 (22:25 -0800)]
fflush, stat: no 'static inline'

* lib/fflush.c (clear_ungetc_buffer_preserving_position)
(clear_ungetc_buffer, disable_seek_optimization)
(restore_seek_optimization, update_fpos_cache):
* lib/stat.c (orig_stat):
Now static, not static inline.
* lib/fflush.c (disable_seek_optimization, restore_seek_optimization)
(update_fpos_cache):
Define only if ! (defined _IO_ftrylockfile || __GNU_LIBRARY__ == 1).
* m4/fflush.m4 (gl_PREREQ_FFLUSH):
* m4/stat.m4 (gl_PREREQ_STAT):
Do not require AC_C_INLINE.

11 years agoerror, filevercmp: no 'static inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:09 +0000 (22:25 -0800)]
error, filevercmp: no 'static inline'

* lib/error.c (is_open, flush_stdout):
* lib/filevercmp.c (order):
Now static, not static inline.
* m4/error.m4 (gl_PREREQ_ERROR):
* modules/filevercmp (configure.ac):
Do not require AC_C_INLINE.

11 years agodup, execute, fatal-signal, etc.: no 'static inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:09 +0000 (22:25 -0800)]
dup, execute, fatal-signal, etc.: no 'static inline'

* lib/dup.c (dup_nothrow):
* lib/execute.c (nonintr_close, nonintr_open):
* lib/fatal-signal.c (uninstall_handlers, install_handlers):
* lib/fopen.c (orig_fopen):
* lib/freadseek.c (freadptrinc):
* lib/freopen.c (orig_freopen):
* lib/fstat.c (orig_fstat, fstat_nothrow):
* lib/get-rusage-as.c (get_rusage_as_via_setrlimit)
(get_rusage_as_via_iterator):
* lib/get-rusage-data.c (get_rusage_data_via_setrlimit):
* lib/getdtablesize.c (_setmaxstdio_nothrow):
* lib/isatty.c (_isatty_nothrow):
* lib/open.c (orig_open):
* lib/read.c (read_nothrow):
* lib/sigprocmask.c (signal_nothrow):
* lib/spawn-pipe.c (nonintr_close, nonintr_open):
* lib/vasnprintf.c (MAX_ROOM_NEEDED):
* lib/wait-process.c (unregister_slave_subprocess):
* lib/write.c (write_nothrow):
Now static, not static inline.
* lib/spawn-pipe.c (nonintr_open): Define only if
(defined _WIN32 || defined __WIN32__) && ! defined __CYGWIN__.
* m4/dup.m4 (gl_PREREQ_DUP):
* m4/execute.m4 (gl_EXECUTE):
* m4/fatal-signal.m4 (gl_FATAL_SIGNAL):
* m4/fopen.m4 (gl_PREREQ_FOPEN):
* m4/freadseek.m4 (gl_FUNC_FREADSEEK):
* m4/freopen.m4 (gl_PREREQ_FREOPEN):
* m4/fstat.m4 (gl_PREREQ_FSTAT):
* m4/getdtablesize.m4 (gl_PREREQ_GETDTABLESIZE):
* m4/isatty.m4 (gl_PREREQ_ISATTY):
* m4/open.m4 (gl_PREREQ_OPEN):
* m4/read.m4 (gl_PREREQ_READ):
* m4/signalblocking.m4 (gl_PREREQ_SIGPROCMASK):
* m4/spawn-pipe.m4 (gl_SPAWN_PIPE):
* m4/vasnprintf.m4 (gl_PREREQ_VASNPRINTF):
* m4/wait-process.m4 (gl_WAIT_PROCESS):
* m4/write.m4 (gl_PREREQ_WRITE):
* modules/get-rusage-as, modules/get-rusage-data (configure.ac):
Do not require AC_C_INLINE.

11 years agoc-strtod, memcoll, readutmp: no 'static inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:08 +0000 (22:25 -0800)]
c-strtod, memcoll, readutmp: no 'static inline'

* lib/c-strtod.c (c_locale):
* lib/memcoll.c (strcoll_loop):
* lib/readutmp.c (desirable_utmp_entry):
Now static, not static inline.
* m4/c-strtod.m4 (gl_C_STRTOD, gl_C_STRTOLD):
* m4/memcoll.m4 (gl_MEMCOLL):
* m4/readutmp.m4 (gl_READUTMP):
Do not require AC_C_INLINE.

11 years agoarctwo, md4, md5, sha1, sha256, sha512: no 'static inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:08 +0000 (22:25 -0800)]
arctwo, md4, md5, sha1, sha256, sha512: no 'static inline'

* lib/arctwo.c (to_uchar):
* lib/md4.c (set_uint32):
* lib/md5.c (set_uint32):
* lib/sha1.c (set_uint32):
* lib/sha256.c (set_uint32):
* lib/sha512.c (set_uint64):
Now static, not static inline.  This is a bit simpler, and doesn't
affect performance with GCC and default optimization.
* m4/arctwo.m4 (gl_ARCTWO):
* m4/md4.m4 (gl_MD4):
* m4/md5.m4 (gl_MD5):
* m4/sha1.m4 (gl_SHA1):
* m4/sha256.m4 (gl_SHA256):
* m4/sha512.m4 (gl_SHA512):
Do not require AC_C_INLINE.

11 years agocond, lock, thread: better 'inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:08 +0000 (22:25 -0800)]
cond, lock, thread: better 'inline'

* lib/glthread/cond.c, lib/glthread/cond.h (_GLTHREAD_COND_INLINE):
* lib/glthread/thread.c, lib/glthread/thread.h (_GLTHREAD_THREAD_INLINE):
New macros.  Use them instead of static inline, for header functions.
* lib/glthread/cond.c (gl_waitqueue_init, gl_waitqueue_remove)
(gl_waitqueue_notify_first, gl_waitqueue_notify_all):
* lib/glthread/lock.c (gl_waitqueue_init)
(gl_waitqueue_notify_first, gl_waitqueue_notify_all):
* lib/glthread/thread.c (get_current_thread_handle):
Change 'static inline' to 'inline'.
* lib/glthread/cond.h, lib/glthread/thread.h:
Use _GL_INLINE_HEADER_BEGIN, _GL_INLINE_HEADER_END.
* m4/cond.m4 (gl_COND):
* m4/lock.m4 (gl_PREREQ_LOCK):
* m4/thread.m4 (gl_THREAD):
Do not require AC_C_INLINE.
* modules/cond, modules/thread (Depends-on): Add extern-inline.

11 years agochdir-long, cycle-check, savewd: better 'inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:08 +0000 (22:25 -0800)]
chdir-long, cycle-check, savewd: better 'inline'

* lib/chdir-long.c (cdb_init, cdb_fchdir, cdb_free)
(find_non_slash):
* lib/cycle-check.c (is_zero_or_power_of_two):
* lib/savewd.c (savewd_delegating):
Change 'static inline' to 'inline'.
* lib/savewd.c, lib/savewd.h (SAVEWD_INLINE): New macro.
Replace all remaining uses of 'static inline' with it.
* lib/savewd.h:
Use _GL_INLINE_HEADER_BEGIN, _GL_INLINE_HEADER_END.
* m4/chdir-long.m4 (gl_PREREQ_CHDIR_LONG):
* m4/cycle-check.m4 (gl_CYCLE_CHECK):
* m4/savewd.m4 (gl_SAVEWD):
Do not require AC_C_INLINE.
* modules/savewd (Depends-on): Add extern-inline.

11 years agobase32, base64: no need for 'inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:08 +0000 (22:25 -0800)]
base32, base64: no need for 'inline'

* lib/base32.c (to_uchar, get_8, decode_8):
* lib/base64.c (to_uchar, get_4, decode_4):
Change 'static inline' to 'inline'.
* m4/base32.m4 (gl_PREREQ_BASE32):
* m4/base64.m4 (gl_PREREQ_BASE64):
Do not require AC_C_INLINE.

11 years agoarray-oset, linkedhash-list, rbtree-oset: no need for 'inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:08 +0000 (22:25 -0800)]
array-oset, linkedhash-list, rbtree-oset: no need for 'inline'

* lib/gl_array_oset.c (gl_array_nx_add_at):
(gl_array_remove_at):
* lib/gl_linkedhash_list.c (hash_resize_after_add)
(add_to_bucket, remove_from_bucket):
* lib/gl_rbtree_oset.c (rotate_left, rotate_right):
Change 'static inline' to 'static', as it's simpler to omit
'inline' unless there's a significant performance advantage.

11 years agolist, oset, xlist, xoset, xsublist: simplify via extern inline
Paul Eggert [Wed, 21 Nov 2012 06:25:07 +0000 (22:25 -0800)]
list, oset, xlist, xoset, xsublist: simplify via extern inline

* lib/gl_list.h, lib/gl_list.c (GL_LIST_INLINE):
* lib/gl_oset.c, lib/gl_oset.h (GL_OSET_INLINE):
* lib/gl_xlist.c, lib/gl_xlist.h (GL_XLIST_INLINE):
* lib/gl_xoset.c, lib/gl_xoset.h (GL_XOSET_INLINE):
* lib/gl_xsublist.c, lib/gl_xsublist.h (GL_XSUBLIST_INLINE):
New macro.  Replace all uses of 'static inline' with it.
[HAVE_INLINE]: Implement functions as *_INLINE functions,
instead of as macros FOO that are defined to static inline
functions FOO_inline.
* lib/gl_list.c, lib/gl_oset.c, lib/gl_xlist.c, lib/gl_xoset.c:
* lib/gl_xsublist.c:
Reimplement from scratch, by defining the corresponding *_INLINE
macro and including the corresponding .h file.  This is simpler.
* m4/gl_list.m4: Remove.
* modules/list, modules/oset, modules/xlist, modules/xoset:
(Files): Remove m4/gl_list.m4.
(configure.ac): Remove gl_LIST.
* modules/list, modules/oset, modules/xlist, modules/xoset:
* modules/xsublist:
(Depends-on): Depend on extern-inline, not inline.

11 years agoxalloc: better 'inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:07 +0000 (22:25 -0800)]
xalloc: better 'inline'

* lib/xmalloc.c, lib/xalloc.h (XALLOC_INLINE):
New macro.  Replace all uses of 'static inline' with it.
(static_inline): Remove.
* lib/xalloc.h (xnmalloc, xnrealloc, x2nrealloc, xcharalloc):
Let 'extern inline' do the work automatically, instead of doing
it by hand.
* m4/xalloc.m4 (gl_PREREQ_XALLOC, gl_PREREQ_XMALLOC):
Remove.  All uses removed.
* modules/xalloc (Depends-on): Remove 'inline'.  Add 'extern-inline'.

11 years agogethrxtime: better 'inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:07 +0000 (22:25 -0800)]
gethrxtime: better 'inline'

* lib/xtime.c: New file.
* lib/gethrxtime.c, lib/gethrxtime.h (GETHRXTIME_INLINE):
* lib/xtime.h (XTIME_INCLUDE):
New macros.  Replace all uses of 'static inline' with them.
* lib/gethrxtime.c (gethrxtime): Define only if
! (HAVE_ARITHMETIC_HRTIME_T && HAVE_DECL_GETHRTIME), since
this source file is now always compiled, because of the extern inline.
* lib/gethrxtime.h, lib/xtime.h:
Use _GL_INLINE_HEADER_BEGIN, _GL_INLINE_HEADER_END.
* m4/gethrxtime.m4 (gl_GETHRXTIME): Do not check for clock macros
if gethrtime works, as they're not needed in that case.
(gl_XTIME): Do not require AC_C_INLINE.
(gl_PREREQ_GETHRXTIME): Remove; all uses removed, as it's always
compiled now.  Move the check into gl_GETHRXTIME.
* modules/gethrxtime (Files, lib_SOURCES): Add lib/xtime.c.
(Depends-on): Add extern-inline.
(configure.ac): gethrxtime is always compiled now.
(lib_SOURCES): Add gethrxtime.c.

11 years agowctype-h: better 'inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:07 +0000 (22:25 -0800)]
wctype-h: better 'inline'

* lib/wctype-h.c: New file.
* lib/wctype.in.h (_GL_WCTYPE_INLINE):
New macro.  Replace all uses of 'static inline' with it.
Use _GL_INLINE_HEADER_BEGIN, _GL_INLINE_HEADER_END.
* m4/wctype_h.m4 (gl_WCTYPE_H): Do not require AC_C_INLINE.
* modules/wctype-h (Files, lib_SOURCES): Add lib/wctype-h.c.
(Depends-on): Add extern-inline.

11 years agounistd: better 'inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:07 +0000 (22:25 -0800)]
unistd: better 'inline'

* lib/unistd.c: New file.
* lib/unistd.in.h (_GL_UNISTD_INLINE):
New macro.  Replace all uses of 'static inline' with it.
Use _GL_INLINE_HEADER_BEGIN, _GL_INLINE_HEADER_END.
* m4/unistd_h.m4 (gl_UNISTD_H): Do not require AC_C_INLINE.
* modules/unistd (Files, lib_SOURCES): Add lib/unistd.c.
(Depends-on): Add extern-inline.

11 years agosys_socket: better 'inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:06 +0000 (22:25 -0800)]
sys_socket: better 'inline'

* lib/sys_socket.c: New file.
* lib/sys_socket.in.h (_GL_SYS_SOCKET_INLINE):
New macro.  Replace all uses of 'static inline' with it.
Use _GL_INLINE_HEADER_BEGIN, _GL_INLINE_HEADER_END.
* m4/sys_socket_h.m4 (gl_HEADER_SYS_SOCKET): Do not require AC_C_INLINE.
* modules/sys_socket (Files, lib_SOURCES): Add lib/sys_socket.c.
(Depends-on): Add extern-inline.

11 years agostdio: better 'inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:06 +0000 (22:25 -0800)]
stdio: better 'inline'

* lib/stdio.c: New file.
* lib/stdio.in.h (_GL_STDIO_INLINE):
New macro.  Replace all uses of 'static inline' with it.
Use _GL_INLINE_HEADER_BEGIN, _GL_INLINE_HEADER_END.
* m4/stdio_h.m4 (gl_STDIO_H): Do not require AC_C_INLINE.
* modules/stdio (Files, lib_SOURCES): Add lib/stdio.c.
(Depends-on): Add extern-inline.

11 years agosigaction: better 'inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:06 +0000 (22:25 -0800)]
sigaction: better 'inline'

* lib/sig-handler.c: New file.
* lib/sig-handler.h (SIG_HANDLER_INLINE):
New macro.  Replace all uses of 'static inline' with it.
Use _GL_INLINE_HEADER_BEGIN, _GL_INLINE_HEADER_END.
* m4/sigaction.m4 (gl_PREREQ_SIG_HANDLER_H): Do not require AC_C_INLINE.
* modules/sigaction (Files, lib_SOURCES): Add lib/sig-handler.c.
(Depends-on): Add extern-inline.

11 years agoselinux-h: better 'inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:06 +0000 (22:25 -0800)]
selinux-h: better 'inline'

* lib/se-context.c, lib/se-selinux.c: New files.
* lib/getfilecon.c (map_to_failure): Omit 'inline' for static function.
* lib/se-context.in.h (SE_CONTEXT_INLINE):
New macro.  Replace all uses of 'static inline' with it.
Use _GL_INLINE_HEADER_BEGIN, _GL_INLINE_HEADER_END.
* lib/se-selinux.in.h (SE_SELINUX_INLINE):
New macro.  Replace all uses of 'static inline' with it.
Use _GL_INLINE_HEADER_BEGIN, _GL_INLINE_HEADER_END.
* modules/selinux-h (Files, lib_SOURCES):
Add lib/se-context.c, lib/se-selinux.c.
(Depends-on): Add extern-inline.
(configure.ac): Do not require AC_C_INLINE.

11 years agopthread: better 'inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:06 +0000 (22:25 -0800)]
pthread: better 'inline'

* lib/pthread.c: New file.
* lib/pthread.in.h (_GL_PTHREAD_INLINE):
New macro.  Replace all uses of 'static inline' with it.
Use _GL_INLINE_HEADER_BEGIN, _GL_INLINE_HEADER_END.
* m4/pthread.m4 (gl_PTHREAD_CHECK):
Add AC_LIBOBJ([pthread]).  Do not require AC_C_INLINE.
* modules/pthread (Files): Add lib/pthread.c.
(Depends-on): Add extern-inline.

11 years agomath: better 'inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:06 +0000 (22:25 -0800)]
math: better 'inline'

* lib/math.c: New file.
* lib/math.in.h (_GL_MATH_INLINE):
New macro.  Replace all uses of 'static inline' with it.
Use _GL_INLINE_HEADER_BEGIN, _GL_INLINE_HEADER_END.
* m4/math_h.m4 (gl_MATH_H):
Do not require AC_C_INLINE.
* modules/math (Files, lib_SOURCES):
Add lib/math.c.
(Depends-on): Add extern-inline.

11 years agocount-one-bits: better 'inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:05 +0000 (22:25 -0800)]
count-one-bits: better 'inline'

* lib/count-one-bits.c: New file.
* lib/count-one-bits.h (COUNT_ONE_BITS_INLINE):
New macro.  Replace all uses of 'static inline' with it.
Use _GL_INLINE_HEADER_BEGIN, _GL_INLINE_HEADER_END.
* m4/count-one-bits.m4 (gl_COUNT_ONE_BITS):
Do not require AC_C_INLINE.
* modules/count-one-bits (Files, lib_SOURCES):
Add lib/count-one-bits.c.
(Depends-on): Add extern-inline.

11 years agocount-leading-zeros: better 'inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:05 +0000 (22:25 -0800)]
count-leading-zeros: better 'inline'

* lib/count-leading-zeros.c: New file.
* lib/count-leading-zeros.h (COUNT_LEADING_ZEROS_INLINE):
New macro.  Replace all uses of 'static inline' with it.
Use _GL_INLINE_HEADER_BEGIN, _GL_INLINE_HEADER_END.
* m4/count-leading-zeros.m4 (gl_COUNT_LEADING_ZEROS):
Do not require AC_C_INLINE.
* modules/count-leading-zeros (Files, lib_SOURCES):
Add lib/count-leading-zeros.c.
(Depends-on): Add extern-inline.

11 years agobitrotate: better 'inline'
Paul Eggert [Wed, 21 Nov 2012 06:25:05 +0000 (22:25 -0800)]
bitrotate: better 'inline'

* lib/bitrotate.c: New file.
* lib/bitrotate.h (BITROTATE_INLINE):
New macros.
Replace all uses of 'static inline' with them.
Use _GL_INLINE_HEADER_BEGIN, _GL_INLINE_HEADER_END.
* modules/bitrotate (Files, lib_SOURCES): Add lib/bitrotate.c.
(Depends-on): Add extern-inline.
(configure.ac): Do not require AC_C_INLINE.

11 years agoautoupdate
Karl Berry [Wed, 28 Nov 2012 16:30:11 +0000 (08:30 -0800)]
autoupdate

11 years agoautoupdate
Karl Berry [Tue, 27 Nov 2012 17:59:19 +0000 (09:59 -0800)]
autoupdate

11 years agomaint.mk: avoid gratuitous failure
Theophile Ranquet [Tue, 20 Nov 2012 10:38:45 +0000 (10:38 +0000)]
maint.mk: avoid gratuitous failure

Reported by Stefano Lattarini in
<http://lists.gnu.org/archive/html/bug-bison/2012-11/msg00022.html>
* top/maint.mk (public-submodule-commit): Quote more safely.

11 years agocanonicalize, canonicalize-lgpl: support MS-Windows file names
Eli Zaretskii [Tue, 20 Nov 2012 21:47:22 +0000 (13:47 -0800)]
canonicalize, canonicalize-lgpl: support MS-Windows file names

See <http://lists.gnu.org/archive/html/bug-gnulib/2012-11/msg00074.html>
for test cases, which it'd be nice to add at some point.
* lib/canonicalize.c, lib/canonicalize-lgpl.c: Include dosname.h.
* lib/canonicalize.c (canonicalize_filename_mode):
* lib/canonicalize-lgpl.c (__realpath):
Use FILE_SYSTEM_PREFIX_LEN instead of assuming that the first
slash is at the beginning of the file name.  Use ISSLASH, instead
of a literal '/'.  Use IS_ABSOLUTE_FILE_NAME instead of comparing
the first character with '/'.  Test for
DOUBLE_SLASH_IS_DISTINCT_ROOT only if the file name does not begin
with a drive letter.
* lib/canonicalize.c (SLASHES): New macro.
(canonicalize_filename_mode): Use SLASHES instead of a literal "/".

11 years agofts: introduce FTS_VERBATIM
Dmitry V. Levin [Sun, 18 Nov 2012 00:40:18 +0000 (04:40 +0400)]
fts: introduce FTS_VERBATIM

This gives clients the option to disable stripping of trailing slashes
from input path names during fts_open initialization.

The recent change v0.0-7611-g3a9002d that made fts_open strip trailing
slashes from input path names had a negative impact on findutils that
relies on the old fts_open behavior to implement POSIX requirement that
each path operand of the find utility shall be evaluated unaltered as it
was provided, including all trailing slash characters.

* lib/fts_.h (FTS_VERBATIM): New bit flag.
(FTS_OPTIONMASK, FTS_NAMEONLY, FTS_STOP): Adjust.
* lib/fts.c (fts_open): Honor it.