Cygwin 1.7.1 placed symlinkat and unlinkat in the wrong headers
(due to a bug in the Linux man pages).
* lib/unistd.in.h (includes): Pick up headers needed for cygwin.
* doc/posix-functions/unlinkat.texi (unlinkat): Document the bug.
* doc/posix-functions/symlinkat.texi (symlinkat): Likewise.
Signed-off-by: Eric Blake <ebb9@byu.net>
On cygwin, compiling with -DGNULIB_POSIXCHECK failed because
the getsubopt link warning definition interfered with the
inclusion of the system header. The fix, as always in these
types of problems, is to ensure that system headers are
completely included before doing any overrides.
* lib/stdlib.in.h (includes): Move unistd inclusion sooner, to
avoid conflicting with system getsubopt.
* doc/posix-functions/getsubopt.texi (getsubopt): Document the
bug.
Signed-off-by: Eric Blake <ebb9@byu.net>
Portions of this commit are commented out because they tickle
glibc bugs. For a real-life example of the bug:
$ env -ua -:
env: invalid option -- :
Try `env --help' for more information.
$ env -:
env: invalid option -- :
Try `env --help' for more information.
$ env -+
env: invalid option -- +
Try `env --help' for more information.
$ env -ua -+
Try `env --help' for more information.
Notice that when -+ is not given as the first argument, the
error message is mistakenly suppressed.
* modules/getopt-posix-tests (Depends-on): Add dup2.
* tests/test-getopt.c (ASSERT): Avoid stderr.
(main): Move stderr to a temporary file.
* tests/test-getopt.h (getopt_loop): No longer manipulate opterr.
Instead, add parameter to inform caller if output occurred.
(test_getopt): Adjust all tests to expect silence, and add new
tests of leading ":".
* doc/glibc-functions/getopt_long.texi (getopt_long): Document
glibc shortcomings with leading "-:" or "+:" in optstring.
* doc/glibc-functions/getopt_long_only.texi (getopt_long_only):
Likewise.
* doc/posix-functions/getopt.texi (getopt): Likewise.
Signed-off-by: Eric Blake <ebb9@byu.net>
futimens/utimensat on Linux fails to bump ctime if mtime is
UTIME_OMIT and atime is specified.
* tests/test-utimens-common.h (check_ctime): Define.
* tests/test-utimens.h (test_utimens): Expose the Linux bug.
* tests/test-futimens.h (test_futimens): Likewise.
* tests/test-lutimens.h (test_lutimens): Likewise.
* doc/posix-functions/futimens.texi (futimens): Document the bug.
* doc/posix-functions/utimensat.texi (utimensat): Likewise.
Signed-off-by: Eric Blake <ebb9@byu.net>
Borrow ideas from dup_cloexec and dup3 to implement F_DUPFD and
F_DUPFD_CLOEXEC. Support querying the inheritance status via
F_GETFD, but for now, no support for changing with F_SETFD.
The remaining portions of fcntl fail with EINVAL.
* m4/fcntl.m4 (gl_FUNC_FCNTL): Also build fcntl.c on mingw.
* lib/fcntl.c (fcntl) <F_DUPFD, F_DUPFD_CLOEXEC, F_GETFD>: Provide
replacement for mingw.
* modules/fcntl (Description): Update.
(Depends-on): Add dup2.
* m4/fcntl_h.m4 (gl_FCNTL_H_DEFAULTS): Add witness.
* modules/fcntl-h (Makefile.am): Substitute it.
* lib/fcntl.in.h (fcntl): Update declaration.
(F_DUPFD, F_GETFD): New macros, when needed.
* doc/posix-headers/fcntl.texi (fcntl.h): Update documentation.
* doc/posix-functions/fcntl.texi (fcntl): Likewise.
* tests/test-fcntl.c (check_flags, main): Enhance test for items
we now guarantee.
Signed-off-by: Eric Blake <ebb9@byu.net>
fcntl(0,F_DUPFD,10000000) mistakenly failed with EMFILE
instead of EINVAL, and fcntl(0,F_DUPFD,-1) mistakenly passed.
* m4/fcntl.m4 (gl_REPLACE_FCNTL): New macro.
(gl_FUNC_FCNTL): Use it. Test for F_DUPFD bug.
* lib/fcntl.c (rpl_fcntl) <F_DUPFD>: Work around it.
<F_DUPFD_CLOEXEC>: Reduce calls to _gl_register_dup.
* doc/posix-functions/fcntl.texi (fcntl): Document it.
Signed-off-by: Eric Blake <ebb9@byu.net>
Implement F_DUPFD_CLOEXEC. The unit test still fails on systems
with other fcntl bugs (such as cygwin 1.5 mishandling F_DUPFD,
or mingw lacking fcntl altogether). Passes on Linux, both with
and without kernel support, and on cygwin 1.7.
* modules/fcntl (Files): List new files.
(configure.ac): Run a test.
* m4/fcntl.m4 (gl_FUNC_FCNTL): New file.
* lib/fcntl.c (rpl_fcntl): Likewise.
* m4/fcntl_h.m4 (gl_FCNTL_H_DEFAULTS): Add witness defaults.
(gl_FCNTL_H): Always replace fcntl.h.
* modules/fcntl-h (Makefile.am): Substitute witnesses.
* lib/fcntl.in.h (fcntl): Declare replacement.
(F_DUPFD_CLOEXEC, GNULIB_defined_F_DUPFD_CLOEXEC): New macro when
needed, plus a witness.
* doc/posix-functions/fcntl.texi (fcntl): Document this.
* doc/posix-headers/fcntl.texi (fcntl.h): Likewise.
* tests/test-fcntl.c: New file.
* modules/fcntl-tests: Likewise.
Signed-off-by: Eric Blake <ebb9@byu.net>
Requiring that {un,}setenv gracefully reject NULL is just a waste
of processing power; POSIX agreed to this argument in
http://austingroupbugs.net/view.php?id=185
so we no longer worry whether a native implementation handles NULL.
* m4/setenv.m4 (gl_FUNC_SETENV_SEPARATE): Test handling of "" but
not NULL.
* tests/test-setenv.c (main): Relax test.
* tests/test-unsetenv.c (main): Likewise.
* doc/posix-functions/setenv.texi (setenv): Document this.
* doc/posix-functions/unsetenv.texi (unsetenv): Likewise.
Signed-off-by: Eric Blake <ebb9@byu.net>
POSIX requires 'echo foo > bar; m4 -Dfoo=1 bar -Dfoo=2 bar' to
output '1' then '2'. To achieve this, m4 relies on the GNU
getopt{,_long} extension of a leading '-'. However, BSD getopt
fails to honor this extension when POSIXLY_CORRECT.
Also, BSD getopt fails to reparse POSIXLY_CORRECT from the
environment even when a reset is requested (whether by
optreset=1 or by optind=0).
* m4/getopt.m4 (gl_GETOPT_CHECK_HEADERS): Test for the bug.
* tests/test-getopt.c (main): Check POSIXLY_CORRECT first, to
flush out BSD bug.
* tests/test-getopt.h (test_getopt): End lists with NULL.
* tests/test-getopt_long.h (test_getopt_long): Likewise.
(test_getopt_long_posix): Enhance test.
* modules/getopt-posix-tests (Depends-on): Add stdbool.
* doc/glibc-functions/getopt_long.texi (getopt_long): Mention
getopt-gnu.
* doc/glibc-functions/getopt_long_only.texi (getopt_long_only):
Likewise.
Signed-off-by: Eric Blake <ebb9@byu.net>
Cygwin 1.5.x mistakenly failed with EINVAL for a duration longer
than 49.7 days (2**32 milliseconds). Meanwhile, the existing
code for HAVE_BUG_BIG_NANOSLEEP would infloop, instead of return
failure, for invalid arguments.
* lib/nanosleep.c (rpl_nanosleep) [HAVE_BUG_BIG_NANOSLEEP]:
Fix logic bug when nanosleep fails. Work around cygwin 1.5.x
bug.
(getnow): Delete, not needed.
* m4/nanosleep.m4 (gl_FUNC_NANOSLEEP): No longer require
LIB_CLOCK_GETTIME.
* modules/nanosleep (Depends-on): Add intprops and verify. Drop
clock-time, gettime.
* doc/posix-functions/nanosleep.texi (nanosleep): Document the
bug.
* modules/nanosleep-tests: New test.
* tests/test-nanosleep.c: New file.
Signed-off-by: Eric Blake <ebb9@byu.net>
On cygwin 1.5.x, sleep amounts larger than 49.7 days (2**32
milliseconds) failed instantly, but with a garbage return
value from uninitialized memory.
* lib/sleep.c (rpl_sleep): Work around the bug.
* m4/sleep.m4 (gl_FUNC_SLEEP): Detect the bug.
(gl_PREREQ_SLEEP): Delete unused macro.
* modules/sleep (Depends-on): Add verify.
* m4/unistd_h.m4 (gl_UNISTD_H_DEFAULTS): Add default.
* modules/unistd (Makefile.am): Substitute witness.
* lib/unistd.in.h (sleep): Update prototype.
* doc/posix-functions/sleep.texi (sleep): Document the bug.
* tests/test-sleep.c (main) [HAVE_DECL_ALARM]: Test it.
* modules/sleep-tests (Depends-on): Check for alarm.
Signed-off-by: Eric Blake <ebb9@byu.net>
chown(name,geteuid(),-1) failed to update the change time if
name was already owned by the current effective user. Work
around it by using chmod, which does not have this bug.
Unfortunately, lchown has the same bug, but OpenBSD 4.0 lacks
lchmod and lutimes, so there is no way to affect ctime without
unlinking and recreating the symlink, which is too dangerous.
* lib/chown.c (rpl_chown): Work around the bug.
* lib/lchown.c (rpl_lchown): Attempt to do likewise.
* m4/chown.m4 (gl_FUNC_CHOWN): Test for ctime bug.
* m4/lchown.m4 (gl_FUNC_LCHOWN): Check for lchmod.
* modules/chown (Depends-on): Add stdbool.
* modules/lchown (Depends-on): Likewise.
* doc/posix-functions/chown.texi (chown): Document the bug.
* doc/posix-functions/lchown.texi (lchown): Likewise.
* tests/test-lchown.h (test_chown): Relax test.
Signed-off-by: Eric Blake <ebb9@byu.net>
POSIX requires setenv(NULL,"",0), setenv("a=b","",0),
unsetenv(NULL), and unsetenv("a=b") to fail with EINVAL, but
many BSD implementations lack validation. The gnulib
replacement for void unsetenv did not do validation, and the
replacement for setenv was out of sync with glibc. Also, some
BSD implementations of setenv("a","==",1) eat the leading '='.
See also some recent Austin Group interpretations on environ:
http://austingroupbugs.net/view.php?id=167http://austingroupbugs.net/view.php?id=185
* lib/setenv.c (setenv) [!HAVE_SETENV]: Resync from glibc.
(setenv) [HAVE_SETENV]: Work around bugs.
* lib/unsetenv.c (unsetenv) [HAVE_UNSETENV]: Work around bugs.
* m4/setenv.m4 (gl_FUNC_SETENV_SEPARATE, gl_FUNC_UNSETENV): Check
for bugs.
(gl_FUNC_SETENV): Write in terms of gl_FUNC_SETENV_SEPARATE.
* m4/environ.m4 (gl_ENVIRON): Avoid expand-before-require.
* m4/stdlib_h.m4 (gl_STDLIB_H_DEFAULTS): Update defaults.
* modules/stdlib (Makefile.am): Update substitutions.
* lib/stdlib.in.h (setenv, unsetenv): Update prototypes.
* doc/posix-functions/setenv.texi (setenv): Document the bugs.
* doc/posix-functions/unsetenv.texi (unsetenv): Likewise.
* modules/setenv-tests: New test.
* modules/unsetenv-tests: Likewise.
* tests/test-setenv.c: New file.
* tests/test-unsetenv.c: Likewise.
Signed-off-by: Eric Blake <ebb9@byu.net>
Solaris 9 fchownat(dir,"name/",uid,gid,flag) has same bugs as
chown and lchown.
* lib/fchownat.c (rpl_fchownat): Work around Solaris bug. Avoid
penalizing glibc chownat when only lchownat is broken.
* m4/openat.m4 (gl_FUNC_FCHOWNAT): Replace fchownat if there are
trailing slash bugs.
* doc/posix-functions/fchownat.texi (fchownat): Document the bug.
* modules/openat-tests (Files): Include more files.
(Depends-on): Add mgetgroups, sleep, stat-time.
(configure.ac): Add additional checks.
(Makefile.am): Build new test.
* tests/test-fchownat.c: New file.
Signed-off-by: Eric Blake <ebb9@byu.net>
FreeBSD 7.2 mistakenly succeeds on getgroups(-1,ptr) (POSIX
requires EINVAL failure since -1 is less than the proper result).
* lib/getgroups.c (rpl_getgroups): Work around the bug.
* m4/getgroups.m4 (gl_FUNC_GETGROUPS): Detect the bug.
* doc/posix-functions/getgroups.texi (getgroups): Document it.
* tests/test-getgroups.c (main): Fix buffer overrun.
Signed-off-by: Eric Blake <ebb9@byu.net>
These days, most systems already declare getgroups with gid_t*.
But in the rare case that GETGROUPS_T is still int but gid_t
is short, the user should not have to uglify their code; let
the replacement hide all the magic.
Tested by configuring with ac_cv_type_getgroups=uint64_t on a
platform with 32-bit gid_t, and ignoring compiler warnings.
However, since we don't replace setgroups, the GETGROUPS_T
workaround is still needed there for now.
* lib/getgroups.c (rpl_getgroups): Change signature. Copy array
an element at a time if GETGROUPS_T is wrong size.
* lib/getugroups.h (getugroups): Change signature.
* lib/unistd.in.h (getgroups): Likewise.
* m4/getgroups.m4 (gl_FUNC_GETGROUPS): Use replacement if
signature needs fixing.
* m4/getugroups.m4 (gl_GETUGROUPS): No longer need
AC_TYPE_GETGROUPS.
* modules/group-member (Depends-on): Add getgroups.
* lib/group-member.c (group_info, get_group_info): Use gid_t.
(group_member): Rely on getgroups replacement.
* lib/getugroups.c (getugroups): Use gid_t.
* tests/test-getgroups.c (main): Likewise.
* NEWS: Mention the signature change.
* doc/posix-functions/getgroups.texi (getgroups): Mention the
problem with signature.
* doc/glibc-functions/setgroups.texi (setgroups): Mention that
GETGROUPS_T is still useful for setgroups.
Signed-off-by: Eric Blake <ebb9@byu.net>
Solaris 9 mknod("name/",mode,dev) mistakenly creates "name" for
non-directory mode. FreeBSD 7.2 mknod("dangling/",mode,dev)
mistakenly creates the target of the symlink if run as root.
FreeBSD and OpenBSD mknod("fifo",S_IFIFO|mode,0) fails for non-root.
Use of mknod caused link failures on mingw.
* modules/mknod: New file.
* m4/mknod.m4 (gl_FUNC_MKNOD): Likewise.
* lib/mknod.c (mknod): Likewise.
* m4/sys_stat_h.m4 (gl_SYS_STAT_H_DEFAULTS): Set witness
defaults.
* modules/sys_stat (Makefile.am): Substitute them.
* lib/sys_stat.in.h (mknod): Declare replacement.
* MODULES.html.sh (Support for systems lacking POSIX:2008):
Document it.
* doc/posix-functions/mknod.texi (mknod): Likewise.
* modules/mknod-tests: New test.
* tests/test-mknod.c: Likewise.
Signed-off-by: Eric Blake <ebb9@byu.net>
Solaris 9 mkfifo("name/",mode) mistakenly creates "name".
FreeBSD 7.2 mkfifo("dangling/",mode) mistakenly creates a fifo
at the target of "dangling". Mingw lacks named pipes altogether,
but this at least avoids link failures.
* modules/mkfifo: New file.
* m4/mkfifo.m4 (gl_FUNC_MKFIFO): Likewise.
* lib/mkfifo.c (mkfifo): Likewise.
* m4/sys_stat_h.m4 (gl_SYS_STAT_H_DEFAULTS): Set witness
defaults.
* modules/sys_stat (Makefile.am): Substitute them.
* lib/sys_stat.in.h (mkfifo): Declare replacement.
* MODULES.html.sh (Support for systems lacking POSIX:2008):
Document it.
* doc/posix-functions/mkfifo.texi (mkfifo): Likewise.
* modules/mkfifo-tests: New test.
* tests/test-mkfifo.h (test_mkfifo): New file, borrowed in part
from test-mkfifoat.c.
* tests/test-mkfifo.c: New file.
Signed-off-by: Eric Blake <ebb9@byu.net>
readlink("link-to-symlink/",buf,len) mistakenly read the contents
of "symlink", rather than failing.
* m4/readlink.m4 (gl_FUNC_READLINK): Also detect FreeBSD bug with
slash on symlink.
* doc/posix-functions/readlink.texi (readlink): Document the bug.
* tests/test-readlink.h (test_readlink): Enhance test.
Signed-off-by: Eric Blake <ebb9@byu.net>
symlink(name,"dangling/") mistakenly created a symlink at the
target of "dangling".
* m4/symlink.m4 (gl_FUNC_SYMLINK): Also detect FreeBSD bug with
slash on symlink.
* doc/posix-functions/symlink.texi (symlink): Document the bug.
* tests/test-symlink.h (test_symlink): Enhance test.
Signed-off-by: Eric Blake <ebb9@byu.net>
link("link-to-file/","a") mistakenly created "a" as a link to "file".
* m4/link.m4 (gl_FUNC_LINK): Also detect FreeBSD bug with slash on
symlink.
* doc/posix-functions/link.texi (link): Document the bug.
* tests/test-link.h (test_link): Enhance test.
* tests/test-linkat.c (main): Update caller.
Signed-off-by: Eric Blake <ebb9@byu.net>
open("link-to-file/", O_RDONLY) mistakenly succeeds. The previous
patch was enough to fix utimens when no fd is involved, but this is
necessary for futimens to pass.
* m4/open.m4 (gl_FUNC_OPEN): Also detect FreeBSD bug with slash on
symlink.
* doc/posix-functions/open.texi (open): Document the bug.
* doc/posix-functions/utimes.texi (utimes): Likewise.
* tests/test-open.h (test_open): Add parameters, and test symlink
handling.
* tests/test-open.c (main): Adjust caller.
* tests/test-fcntl-safer.c (main): Likewise.
* modules/open-tests (Depends-on): Add stdbool, symlink.
* modules/fcntl-safer-tests (Depends-on): Likewise.
* tests/test-openat.c (main): Add test-open tests.
Signed-off-by: Eric Blake <ebb9@byu.net>
Like Solaris 9, FreeBSD 7.2 mistakenly allows stat("link-to-file/").
Unlike Solaris, it correctly forbids stat("file/"). A number of
interfaces are affected (such as utimes), but replacing stat is
enough to catch several by reusing the Solaris 9 fixes.
* m4/stat.m4 (gl_FUNC_STAT): Also detect FreeBSD bug with slash on
symlink.
* doc/posix-functions/stat.texi (stat): Document the bug.
* tests/test-stat.h (test_stat_func): Add argument.
* tests/test-stat.c (main): Adjust caller.
* tests/test-fstatat.c (main): Likewise.
* modules/stat-tests (Depends-on): Add stdbool, symlink.
Reported by Jim Meyering.
Signed-off-by: Eric Blake <ebb9@byu.net>
utimes("file/",times) mistakenly succeeds. This commit doesn't fix
utimes, but does make utimensat be careful before calling utimes.
The test is now enhanced to test trailing slashes and directories.
Meanwhile, cygwin 1.5 stat() on a directory changes atime (it does
a readdir under the hood to populate st_nlink), so only mtime of
a directory is reliable enough for testing. Cygwin 1.7 no longer
has this problem, because it no longer wastes time on st_nlink.
* lib/utimens.c (fdutimens, lutimens): Force a stat if platform
has trailing slash bugs.
* tests/test-lutimens.h (test_lutimens): Enhance test.
* tests/test-utimens.h (test_utimens): Likewise.
* doc/posix-functions/utime.texi (utime): Document the bug.
* doc/posix-functions/utimes.texi (utimes): Likewise.
* doc/posix-functions/utimensat.texi (utimensat): Likewise.
* doc/glibc-functions/futimesat.texi (futimesat): Likewise.
* doc/glibc-functions/lutimes.texi (lutimes): Mention utimens.
* doc/posix-functions/futimens.texi (futimens): Mention
limitation.
Signed-off-by: Eric Blake <ebb9@byu.net>
Provide utimensat where it is missing, and rpl_utimensat to work
around ENOSYS and EINVAL bugs in older Linux kernels.
* modules/utimensat: New file.
* lib/utimensat.c (utimensat): Likewise.
* m4/utimensat.m4 (gl_FUNC_UTIMENSAT): Likewise.
* lib/utimens.c (utimensat): Avoid recursion into rpl_utimensat,
so we can work around Linux bugs.
* m4/sys_stat_h.m4 (gl_SYS_STAT_H_DEFAULTS): Add witnesses.
* modules/sys_stat (Makefile.am): Substitute them.
* lib/sys_stat.in.h (utimensat): Declare it.
* MODULES.html.sh (systems lacking POSIX:2008): Mention module.
* doc/posix-functions/utimensat.texi (utimensat): Likewise.
* modules/utimensat-tests: New test.
* tests/test-utimensat.c: Likewise.
Signed-off-by: Eric Blake <ebb9@byu.net>
Provides futimens where it is missing, and rpl_futimens to work
around bugs in older Linux kernels.
* modules/futimens: New file.
* lib/futimens.c (futimens): Likewise.
* m4/futimens.m4 (gl_FUNC_FUTIMENS): Likewise.
* lib/utimens.c (futimens): Avoid recursion into rpl_futimens, so
we can work around Linux bugs.
* m4/sys_stat_h.m4 (gl_SYS_STAT_H_DEFAULTS): Add witnesses.
* modules/sys_stat (Makefile.am): Substitute them.
* lib/sys_stat.in.h (futimens): Declare it.
* MODULES.html.sh (systems lacking POSIX:2008): Mention module.
* doc/posix-functions/futimens.texi (futimens): Likewise.
* modules/futimens-tests: New test.
* tests/test-futimens.c: Likewise.
Signed-off-by: Eric Blake <ebb9@byu.net>
dup2(1,1000000) needs to fail with EBADF (per POSIX), not
EMFILE, based on our usage of it in other modules.
* m4/dup2.m4 (gl_FUNC_DUP2): Detect bug.
* doc/posix-functions/dup2.texi (dup2): Document it.
Reported by Nelson H. F. Beebe and Jim Meyering.
Signed-off-by: Eric Blake <ebb9@byu.net>
