file-has-acl: use acl_extended_file_nofollow if available
* lib/acl-internal.h (HAVE_ACL_EXTENDED_FILE): New macro.
(acl_extended_file): New macro.
* lib/file-has-acl.c (file_has_acl): Use acl_extended_file_nofollow.
* m4/acl.m4 (gl_FUNC_ACL): Check for acl_extended_file_nofollow.
This addresses http://bugzilla.redhat.com/692823.
Bruno Haible [Thu, 21 Jul 2011 22:16:46 +0000 (00:16 +0200)]
Declare system functions in a way that works with C++.
* m4/fdopendir.m4 (gl_FUNC_FDOPENDIR) [C++]: In the test program,
declare fdopendir as extern "C".
* m4/frexpl.m4 (gl_FUNC_FREXPL_WORKS) [C++]: In the test program,
declare frexpl as extern "C".
* m4/getaddrinfo.m4 (gl_GETADDRINFO) [C++]: In the test program,
declare gai_strerror as extern "C".
* m4/getdomainname.m4 (gl_FUNC_GETDOMAINNAME) [C++]: In the test
programs, declare gai_strerror as extern "C".
* m4/getlogin_r.m4 (gl_FUNC_GETLOGIN_R) [C++]: In the test program,
declare getlogin_r as extern "C".
* m4/ioctl.m4 (gl_FUNC_IOCTL) [C++]: In the test program, declare ioctl
as extern "C".
* m4/ldexpl.m4 (gl_FUNC_LDEXPL_WORKS) [C++]: In the test program,
declare ldexpl as extern "C".
* m4/logb.m4 (gl_FUNC_LOGB) [C++]: In the test programs, declare logb
as extern "C".
* m4/ls-mntd-fs.m4 (gl_LIST_MOUNTED_FILE_SYSTEMS) [C++]: In the test
program, declare getmntinfo as extern "C".
* m4/stpncpy.m4 (gl_FUNC_STPNCPY) [C++]: In the test program, declare
stpncpy as extern "C".
* m4/strerror_r.m4 (gl_FUNC_STRERROR_R_WORKS) [C++]: In the test
program, declare __xpg_strerror_r as extern "C".
* m4/strndup.m4 (gl_FUNC_STRNDUP) [C++]: In the test program, declare
strndup as extern "C".
* m4/sys_select_h.m4 (gl_HEADER_SYS_SELECT) [C++]: In the test program,
declare memset and bzero as extern "C".
Reported by Sam Steingold <sds@gnu.org>.
timer-time: A new module to check for timer_settime()
* m4/timer_time.m4: Check for the POSIX function.
* modules/timer-time: Add the new module.
* MODULES.html.sh (Compat checks for POSIX:2008 functions):
Mention it.
* doc/posix-functions/timer_create.texi: Add timer-time as
the module, and update the platforms where unavailable.
* doc/posix-functions/timer_delete.texi: Likewise.
* doc/posix-functions/timer_gettime.texi: Likewise.
* doc/posix-functions/timer_settime.texi: Likewise.
Paul Eggert [Wed, 20 Jul 2011 00:57:20 +0000 (02:57 +0200)]
pthread_sigmask: assume POSIX threads if --avoid=threadlib
* m4/pthread_sigmask.m4 (gl_FUNC_PTHREAD_SIGMASK): If gl_THREADLIB is
not defined, assume POSIX threads and look for pthread_sigmask in
$LIBS, without changing $CPPFLAGS.
Bruno Haible [Tue, 19 Jul 2011 21:43:05 +0000 (23:43 +0200)]
strstr: Update cross-compilation guess.
* m4/strstr.m4 (gl_FUNC_STRSTR): On glibc > 2.12 with x86 or x86_64
CPUs, guess no, in view of glibc
BZ #12100 <http://sourceware.org/bugzilla/show_bug.cgi?id=12100>.
Suggested by Eric Blake. Reported by Reuben Thomas.
getopt-gnu: suppress core dumps from detection code
* m4/getopt.m4 (gl_GETOPT_CHECK_HEADERS): Use the nocrash module
to suppress core dumps that may well occur on glibc systems.
These core dumps might not be cleaned up automatically, or could
trigger some system core dump handling logic.
* modules/getopt-gnu: Depend on nocrash.
Paul Eggert [Sat, 16 Jul 2011 01:27:51 +0000 (18:27 -0700)]
doc: Document NonStop portability issues.
* doc/posix-functions/sigaction.texi (sigaction):
* doc/posix-headers/signal.texi (signal.h):
Document NonStop. See Joachim Schmitz in
http://lists.gnu.org/archive/html/bug-coreutils/2011-07/msg00062.html
Bruno Haible [Sat, 16 Jul 2011 00:36:07 +0000 (02:36 +0200)]
ffs: More tests.
* tests/test-ffs.c (NBITS): New macro.
(main): Add more tests.
* tests/test-ffsl.c (NBITS): New macro.
(main): Add more tests.
* tests/test-ffsll.c (NBITS): New macro.
(main): Add more tests.
Bruno Haible [Tue, 12 Jul 2011 09:41:24 +0000 (11:41 +0200)]
Rename module 'link-warning' to 'snippet/link-warning'.
* modules/snippet/link-warning: Renamed from modules/link-warning.
(Files, Makefile.am): Update.
* build-aux/snippet/link-warning.h: Renamed from
build-aux/link-warning.h.
* MODULES.html.sh (Support for building libraries and executables):
Update.
Paul Eggert [Tue, 12 Jul 2011 08:16:45 +0000 (01:16 -0700)]
modules/_Exit-tests: test _Noreturn too
* tests/test-_Exit.c (MTA, Charlie): New functions, taken from the
old tests/test-stdnoreturn.c. This tests the _Noreturn keyword a bit.
(main): Use them.
Paul Eggert [Tue, 12 Jul 2011 08:16:08 +0000 (01:16 -0700)]
stdnoreturn, stdnoreturn-tests: remove modules
They're not needed here and a bit premature for use elsewhere. See
<http://lists.gnu.org/archive/html/bug-gnulib/2011-07/msg00209.html>.
* m4/stdnoreturn.m4, modules/stdnoreturn, modules/stdnoreturn-tests:
* tests/test-stdnoreturn.c: Remove files.
* MODULES.html.sh (_Noreturn <stdnoreturn.h>): Remove section.
* lib/openat.h, lib/sigpipe-die.h, lib/xalloc.h, lib/xmemdup0.h:
* lib/xstrtol.h: Use _Noreturn rather than including <stdnoreturn.h>
and using noreturn.
* modules/openat, modules/sigpipe-die, modules/xalloc:
* modules/xmemdup0, modules/xstrtol:
Remove dependency on stdnoreturn.
Paul Eggert [Tue, 12 Jul 2011 08:14:47 +0000 (01:14 -0700)]
_Noreturn-tests: remove module
* modules/_Noreturn-tests: Remove.
* modules/stdnoreturn-tests (Files): Remove tests/test-_Noreturn.c.
* tests/test-_Noreturn.c: Remove.
* tests/test-stdnoreturn.c: Merge from the old
tests/test-_Noreturn.c, testing both noreturn and _Noreturn.
Paul Eggert [Tue, 12 Jul 2011 08:02:54 +0000 (01:02 -0700)]
stdlib: use _Noreturn
* lib/stdlib.in.h (_GL_ATTRIBUTE_NORETURN): Remove.
(_Exit): Use _Noreturn rather than _GL_ATTRIBUTE_NORETURN.
* modules/stdlib (Depends-on): Add _Noreturn.
(stdlib.h): Depend on $(_NORETURN_H), and copy it in.
Paul Eggert [Sat, 9 Jul 2011 21:56:56 +0000 (14:56 -0700)]
stdint: respect system's intmax_t if INTMAX_MAX
* lib/stdint.in.h (intmax_t, uintmax_t): If the system defines
INTMAX_MAX, assume its intmax_t is OK. Similarly for and
uintmax_t. This is for some Mac OS X builds, where intmax_t is
long but int64_t is long long, and where we will clash with the
system intmax_t if we override it. See
<http://lists.gnu.org/archive/html/bug-gnulib/2011-07/msg00160.html>.
(INTMAX_C, UINTMAX_C): For consistency, respect the system's
INTMAX_C if INTMAX_MAX and INTMAX_C are both defined, and
similarly for UINTMAX_C.
Bruno Haible [Fri, 8 Jul 2011 23:37:58 +0000 (01:37 +0200)]
pthread_sigmask: Work around IRIX bug.
* m4/pthread_sigmask.m4 (gl_FUNC_PTHREAD_SIGMASK): Test for the IRIX
bug.
* lib/pthread_sigmask.c (pthread_sigmask): usleep for some time when
there may be unblocked pending signals.
* doc/posix-functions/pthread_sigmask.texi: Mention the IRIX bug.
Bruno Haible [Fri, 8 Jul 2011 23:01:36 +0000 (01:01 +0200)]
pthread_sigmask: Work around Cygwin bug.
* m4/pthread_sigmask.m4 (gl_FUNC_PTHREAD_SIGMASK): Test for the Cygwin
bug.
* lib/pthread_sigmask.c (pthread_sigmask): Fix the return value from
the system's pthread_sigmask function.
* doc/posix-functions/pthread_sigmask.texi: Mention the Cygwin bug.
Bruno Haible [Fri, 8 Jul 2011 22:27:00 +0000 (00:27 +0200)]
pthread_sigmask: Work around bug in single-threaded implementation.
* m4/pthread_sigmask.m4 (gl_FUNC_PTHREAD_SIGMASK): Test for the
FreeBSD, HP-UX, Solaris bug.
(gl_PREREQ_PTHREAD_SIGMASK): New macro.
* lib/pthread_sigmask.c: Include <stddef.h>.
(pthread_sigmask): If HAVE_PTHREAD_SIGMASK, define as a wrapper around
the system's pthread_sigmask function.
* modules/pthread_sigmask (configure.ac): Invoke
gl_PREREQ_PTHREAD_SIGMASK.
* doc/posix-functions/pthread_sigmask.texi: Mention bug on FreeBSD,
HP-UX, Solaris.
Bruno Haible [Fri, 8 Jul 2011 17:12:14 +0000 (19:12 +0200)]
Tests for module 'pthread_sigmask'.
* modules/pthread_sigmask-tests: New file.
* tests/test-pthread_sigmask1.c: New file, based on
tests/test-sigprocmask.c.
* tests/test-pthread_sigmask2.c: New file.
Paul Eggert [Fri, 8 Jul 2011 17:49:07 +0000 (10:49 -0700)]
pthread_sigmask: Assume POSIX when not gl_THREADLIB.
* m4/pthread_sigmask.m4 (gl_FUNC_PTHREAD_SIGMASK):
When gl_THREADLIB is not in use, assume that the POSIX sematics
are desired. This is better for Emacs, which uses POSIX semantics
on GNUish and/or POSIXish platforms, and does not use threads at
all otherwise.
Eric Blake [Thu, 7 Jul 2011 17:37:37 +0000 (11:37 -0600)]
fts: introduce FTS_NOATIME
This gives clients the option to try a non-invasive traversal,
where merely visiting a directory does not update its timestamp,
where such is supported by the kernel.
Note that this is best-effort. FTS_NOATIME is silently ignored
on systems that lack O_NOATIME support - it is up to the caller
to decide whether to issue an error when the gnulib replacement
<fcntl.h> defined O_NOATIME to 0, and/or detect older Linux
kernels where O_NOATIME is defined to non-zero but has no effect.
Note that whiteout support and O_NOATIME support are currently
orthogonal: there is no way to get O_NOATIME behavior when using
__opendir2 to visit whiteouts on BSD systems. So far, I don't
know of any system with both __opendir2 and O_NOATIME; if such
a system exists, then fts_build() needs a tweak (then again,
such a system would probably add DTF_NOATIME for __opendir2).
* lib/fts_.h (FTS_NOATIME): New bit flag.
(FTS_OPTIONMASK): Adjust.
* lib/fts.c (diropen, fts_open, fts_build): Honor it.
(fd_ring_check): Debug code unconditionally uses O_NOATIME.
Needed for findutils bug http://savannah.gnu.org/bugs/?33724
Bruno Haible [Fri, 8 Jul 2011 02:13:53 +0000 (04:13 +0200)]
signal-c++-tests: Check declaration of pthread_sigmask.
* tests/test-signal-c++.cc: Check declaration of pthread_sigmask.
* modules/signal-c++-tests (Makefile.am): Link test-signal-c++ against
$(LIB_PTHREAD_SIGMASK).
Bruno Haible [Fri, 8 Jul 2011 02:05:45 +0000 (04:05 +0200)]
pthread_sigmask: Fix link requirements on OSF/1 5.1 and with pth.
* lib/signal.in.h (pthread_sigmask): Override if
REPLACE_PTHREAD_SIGMASK is 1.
* m4/signal_h.m4 (gl_SIGNAL_H_DEFAULTS): Initialize
REPLACE_PTHREAD_SIGMASK.
* m4/pthread_sigmask.m4 (gl_FUNC_PTHREAD_SIGMASK): Set also
REPLACE_PTHREAD_SIGMASK. Set and substitute LIB_PTHREAD_SIGMASK.
* modules/signal (Makefile.am): Substitute REPLACE_PTHREAD_SIGMASK.
* modules/pthread_sigmask (Depends-on, configure.ac): Update condition.
(Link): Set to $(LIB_PTHREAD_SIGMASK), not $(LIBMULTITHREAD).
Bruno Haible [Fri, 8 Jul 2011 00:08:15 +0000 (02:08 +0200)]
pthread_sigmask: Make declaration C++ safe.
* lib/signal.in.h: In two special conditions, just do an #include_next.
(pthread_sigmask): Test HAVE_PTHREAD_SIGMASK, not REPLACE_PTHREAD_MASK.
Invoke _GL_CXXALIAS_SYS and _GL_CXXALIASWARN.
* m4/signal_h.m4 (gl_SIGNAL_H_DEFAULTS): Initialize
HAVE_PTHREAD_SIGMASK, not REPLACE_PTHREAD_MASK.
* m4/pthread_sigmask.m4 (gl_PTHREAD_SIGMASK): Set HAVE_PTHREAD_SIGMASK,
not REPLACE_PTHREAD_MASK.
* modules/signal (Makefile.am): Substitute HAVE_PTHREAD_SIGMASK,
not REPLACE_PTHREAD_MASK.
* modules/pthread_sigmask (Depends-on, configure.ac): Update condition.
Eric Blake [Thu, 7 Jul 2011 17:02:23 +0000 (11:02 -0600)]
getopt: avoid compiler warning during configure
Some compilers (or warning settings within a compiler) rightfully
complain about assigning a string literal to a char *, since this
can lead to undefined behavior if those characters are modified.
Use the same casts as in the test file to avoid a spurious rejection
of a system getopt merely because of warnings.
* m4/getopt.m4 (gl_GETOPT_CHECK_HEADRS): Avoid problems with
assigning string literals to non-const pointer.
Paul Eggert [Wed, 6 Jul 2011 07:42:24 +0000 (00:42 -0700)]
sys_select: define sigset_t more portably
* lib/sys_select.in.h: Always include <sys/types.h>, since
we now need sigset_t and mingw defines it there.
Include <signal.h> before split inclusion guard, to avoid
mishaps on Solaris, whose <signal.h> eventually includes us.
* m4/signal_h.m4 (gl_SIGNAL_H): Require gl_CHECK_TYPE_SIGSET_T.
(gl_CHECK_TYPE_SIGSET_T): New macro, most of the contents of
which come from ...
* m4/signalblocking.m4 (gl_SIGNALBLOCKING): Require
gl_CHECK_TYPE_SIGSET_T.
(gl_PREREQ_SIGPROCMASK): Now a no-op, since gl_CHECK_TYPE_SIGSET_T
does the real work.
* modules/sys_select (Depends-on): Add 'signal'.
Paul Eggert [Wed, 6 Jul 2011 07:05:40 +0000 (00:05 -0700)]
pselect: Use pthread_sigmask, not sigprocmask.
* lib/pselect.c (pselect): Use pthread_sigmask, as it supports
multithreaded apps better than sigprocmask does.
* modules/pselect (Depends-on): Depend on pthread_sigmask, not
sigprocmask directly.
Paul Eggert [Wed, 6 Jul 2011 06:58:08 +0000 (23:58 -0700)]
* lib/pselect.c (pselect): Use plain name, without "rpl_".
Don't #undef, since we don't need any underlying pselect.
* modules/pselect (Depends-on): Use our pselect.o if !HAVE_PSELECT.
Both changes suggested by Bruno Haible.
Paul Eggert [Wed, 6 Jul 2011 06:44:24 +0000 (23:44 -0700)]
pthread_sigmask: new module
* MODULES.html.sh (Support for systems lacking POSIX:2008): Add it.
* doc/posix-functions/pthread_sigmask.texi: Document new module.
* lib/signal.in.h (pthread_sigmask): Arrange for replacement.
This is done only as a macro; I don't know how well that'll
work for C++. Move <sys/types.h> include before the include_next,
to avoid mishap on Solaris.
* m4/signal_h.m4 (gl_SIGNAL_H, gl_SIGNAL_H_DEFAULTS): Check for it.
* modules/signal (Makefile.am): Substitute the check's results.
* modules/pthread_sigmask, m4/pthread_sigmask.m4: New files.
Paul Eggert [Tue, 5 Jul 2011 21:46:08 +0000 (14:46 -0700)]
test-pselect: new module
* modules/pselect-tests, tests/test-pselect.c: New files.
* tests/test-select.c, tests/test-sys_select-c++.cc:
If TEST_PSELECT is defined, test pselect instead of testing select.
Eric Blake [Fri, 1 Jul 2011 14:20:06 +0000 (08:20 -0600)]
snprintf: guarantee %1$d, for libintl
Newer mingw (but not yet mingw64) provides two flavors of
snprintf: _snprintf defers straight to msvcrt, which has broken
return value and does not understand %llu or %zu; and snprintf,
which fixes these two bugs but does not understand %1$s.
Libintl specifically favors _snprintf, with broken return value,
even when compiled on mingw with a fixed snprintf, because the
only behavior which it wants to fix is %1$s handling. But this
means that the replacement libintl_snprintf has a broken return.
If one uses the 'snprintf-posix' module, then the gnulib
replacement kicks in, and does everything that libintl needs, so
on mingw, <libintl.h> specifically avoids overriding snprintf if
it detected that gnulib replaced snprintf. However, if one only
uses the 'snprintf' module and also uses libintl, this means
there are two problems:
1. The gnulib 'snprintf' module does not replace the mingw
snprintf function, because it has proper return values, while the
libintl.h header knows that %1$d is broken so snprintf must be
replaced, with the end result that the application gets the
libintl replacement snprintf with broken return values in spite
of the gnulib module.
2. Conversely, if the application did '#define snprintf snprintf',
that would be enough to make libintl avoid installing its own
replacement because libintl would see the define as a sign that
gnulib is happy with snprintf. However, if gnulib didn't enforce
%1$s, users can end up with translated strings that break when
passed to the native snprintf.
Happily, the gnulib snprintf replacement already guarantees %1$s
without needing any further preprocessor macros defined, and
without dragging in the LGPLv3+ bulk of snprintf-posix, so the
problem boils down to guaranteeing that gnulib will replace
snprintf if it lacks %1$s support. Basically, gnulib must
replace snprintf under all the same conditions as libintl, as
well as any other conditions of its own, if the libintl trick
of deferring to gnulib is to work correctly.