On Cygwin 2.9.0 I'm seeing a test failure:
FAIL: test-xstrtol.sh
This patch avoids the failure, acknowledging that the results here
are different on different Cygwin versions.
2025-01-05 Bruno Haible
xstrtol, xstrtoll tests: Avoid test failure after 2024-07-25 c
ch adds a workaround.
2024-12-24 Bruno Haible
access, euidaccess tests: Avoid test failure in Cygwin 3.5.5.
* tests/test-access.h [__CYGWIN__]: Include , ,
.
(is_administrator): New function.
(is_root): New macro.
(test_access): Reenable F_OK te
On mingw-w64 with UCRT option, I see this test failure:
FAIL: test-c32isprint.sh
../../gltests/test-c32isprint.c:178: assertion 'is != 0' failed
FAIL test-c32isprint.sh (exit status: 1)
This patch fixes it.
2024-12-22 Bruno Haible
c32isprint te
On MSYS2, there is also this test failure:
FAIL: test-posix_spawn-chdir
read output is not the expected output
FAIL test-posix_spawn-chdir.exe (exit status: 1)
This patch fixes it.
2024-12-22 Bruno Haible
posix_spawn_file_actions_addchdir tests: Avoid
On AIX 7.1 and 7.3.1, I see this test failure:
FAIL: test-realloc-posix
../../gltests/test-realloc-posix.c:34: assertion 'p != NULL' failed
FAIL test-realloc-posix (exit status: 134)
The cause is my change from 2024-10-27:
<https://lists.gnu.org/arc
buffer.exe (exit status: 3)
Fixed as follows. Thanks for the report.
2024-09-26 Bruno Haible
string-buffer tests: Avoid test failure on native Windows.
Reported by Michele Locati in
<https://lists.gnu.org/archive/html/bug-gnulib/2024-09/msg00135.html>.
* t
On Haiku, I see this test failure:
FAIL: test-unlockpt
===
../../gltests/test-unlockpt.c:36: assertion 'unlockpt (-1) == -1' failed
Abort
FAIL test-unlockpt (exit status: 149)
Since POSIX does not require failure here (see
<https://pubs.opengroup.org/onlinepu
st_ctime.
2024-09-02 Bruno Haible
utime, futimens tests: Avoid test failure on Haiku.
* tests/test-utimens-common.h (checkable_atime): On Haiku, return false.
diff --git a/tests/test-utimens-common.h b/tests/test-utimens-common.h
index f9fdb9a867..a59e65df8c 100644
--- a/tests/test-utime
0x00C9) = 0x00E9
towlower(0x00E9) = 0x00E9
towupper(0x00C9) = 0x00C9
towupper(0x00E9) = 0x00C9
This is the cause for a test failure that I see with MSVC:
FAIL: test-mbscasestr2.sh
=
C:\cygwin64\home\bruno\testdir-all-for-mingw\gltests\test-mbscasestr2.c:56:
asse
On Haiku, I'm seeing this test failure (in a testdir of
posix_openpt grantpt pt_chown
):
FAIL: test-posix_openpt
===
../../gltests/test-posix_openpt.c:58: assertion '0 <= slave' failed
Abort
FAIL test-posix_openpt (exit status: 149)
This patch fixes it:
ts.
2024-08-12 Bruno Haible
pthread-rwlock: Attempt to avoid test failure on some more machines.
* tests/test-pthread-rwlock-waitqueue.c (STEP_INTERVAL): Bump the
minimum interval to 20 ms on macOS.
>From ef4c430e2d012bf2edaa13f28fd0042169845222 Mon Sep 17 00:00:00 2001
_GETOPT_PREFIX gets defined in config.h.
So, this patch fixes it. I don't remember why I had added this special casing
of mingw in 2009.
2024-07-23 Bruno Haible
getopt-posix, getopt-gnu tests: Avoid test failure on mingw.
* tests/test-getopt.h (test_getopt): Don't
The test-time test fails on FreeBSD 14.0/x86_64, in a VirtualBox VM, but
only if 2 or more CPUs are assigned to the VM.
This patch fixes it.
2024-06-28 Bruno Haible
time: Fix test failure on FreeBSD.
* m4/time.m4 (gl_FUNC_TIME): Guess no for FreeBSD in general.
* doc
[Ccing bug-gnulib]
Collin Funk wrote:
> > CentOS 7 will be supported until 2028.
> > I've updated Gnulib's end-of-life.txt:
>
> I remember we based our Python 3.7 dependency on CentOS almost being
> almost end of life [1]. Does this change anything there? Or will
> installing some python-3.7 packa
Bruno Haible writes:
>> Would it make sense to use __illumos__ here?
>
> No, it is too new. [1]
Makes sense. Feel free to change the ones I added previously to __sun if
you would like.
Collin
On Solaris 11 OpenIndiana and Solaris 11 OmniOS, I now (after the c32rtomb
sanity check addition) see a test failure:
FAIL: test-c32width
===
../../tests/test-c32width.c:98: assertion 'c32width (0x20369) == 2' failed
../../build-aux/test-driver: line 114: 15481: Abor
Hi Collin,
> > -#if !(defined __FreeBSD__ || defined __DragonFly__)
> > +#if !(defined __FreeBSD__ || defined __DragonFly__ || defined
> > __sun)
>
> Would it make sense to use __illumos__ here?
No, it is too new. [1]
Bruno
[1] https://lists.gnu.org/archive/html/bug-gnulib/202
Hi Bruno,
Bruno Haible writes:
> -#if !(defined __FreeBSD__ || defined __DragonFly__)
> +#if !(defined __FreeBSD__ || defined __DragonFly__ || defined __sun)
Would it make sense to use __illumos__ here?
Not sure if you tested this on regular Oracle Solaris.
Collin
On a Solaris 11 OmniOS machine with a zh_CN.GB18030 locale, libunistring
(or a testdir without module 'uchar-c23') also shows this test failure:
FAIL: test-c32ispunct.sh
../../gltests/test-c32ispunct.c:283: assertion 'is != 0' failed
../../gltests/t
FAIL test-c32isblank.sh (exit status: 1)
The cause is that here, iswblank (L'\f') returns true, not false (as expected).
This patch avoids the failure.
2024-06-08 Bruno Haible
c32isblank tests: Avoid test failure on NetBSD 10.0.
* tests/test-c32isblank.c (main): Disable
On mingw, I see this test failure:
FAIL: test-mbfile.sh
../../gltests/test-mbfile.c:49: assertion 'mb_len (next) == 2' failed
Stack trace:
0xbd502c13 main
../../gltests/test-mbfile.c:49
The cause is that setlocale has not installed an UTF-8 locale, bu
ss.c:37
FAIL test-euidaccess.exe (exit status: 1)
Likewise on MSVC.
This patch fixes it.
2024-05-25 Bruno Haible
euidaccess: Fix test failure on native Windows.
* lib/euidaccess.c: Don't include .
(euidaccess): On native Windows, invoke access, not _access.
diff
On the GitHub Cygwin machines (Cygwin 3.3.6 and 3.5.3), I see this test
failure:
FAIL: test-sethostname2
===
setting a too long hostname succeeded.
FAIL test-sethostname2.exe (exit status: 1)
The cause: Cygwin apparently does not reject a too long hostname argument.
While
On Cygwin 3.4.6 and older, but not on Cygwin 3.5.3, I see this test failure:
FAIL: test-poll.exe
This patch works around it, like we already did for AIX and native Windows.
2024-05-23 Bruno Haible
poll tests: Avoid test failure on Cygwin.
* tests/test-poll.c (test_pipe
Collin Funk wrote:
> I didn't know they were run as root (CI is new to me).
That's why I start all CI runs with these commands:
uname -a
id
env | LC_ALL=C sort
pwd
It helps understanding the environment :) Sometimes I need
to add
echo $PA
On 5/21/24 6:45 AM, Bruno Haible wrote:
> The CI runs the Solaris 11 and Solaris 11 OmniOS tests as root, and shows
> these two test failures:
>
> FAIL: test-access
> FAIL: test-euidaccess
>
> This occurs because although the file does not have the 'x' bit set,
> for root faccessat() returns 0.
>
5-21 Bruno Haible
access, euidaccess tests: Avoid test failure for root user on Solaris.
* tests/test-access.h (test_access): On Solaris, for the root user,
don't expect X_OK permissions to be absent.
* doc/posix-functions/access.texi: Mention the Solar
Simon Josefsson wrote:
> > + int alarm_value = 50;
> >signal (SIGALRM, SIG_DFL);
> > - alarm (10);
> > + alarm (alarm_value);
>
> Nice trick, but doesn't the compiler optimize away this?
When I'm compiling for debugging, I use CFLAGS="-ggdb", that is,
implicitly -O0. (Debugging at -O2 is s
Bruno Haible writes:
> + int alarm_value = 50;
>signal (SIGALRM, SIG_DFL);
> - alarm (10);
> + alarm (alarm_value);
Nice trick, but doesn't the compiler optimize away this? Maybe a
'volatile' is needed.
/Simon
signature.asc
Description: PGP signature
for QEMU-emulated
environments, and using the gdb-overridable idiom from
https://git.savannah.gnu.org/gitweb/?p=gnulib.git;a=commitdiff;h=a8d9f4635238f845bb76269a54fbb2c55255d365
.
2024-05-09 Bruno Haible
unistr/u8-strstr tests: Avoid test failure with ASAN.
* tests/unistr/test
723==ABORTING
FAIL test-free (exit status: 134)
The cause: this test assumes glibc malloc(), not some ASAN wrappers.
2024-05-09 Bruno Haible
free tests: Avoid test failure with ASAN.
* tests/test-free.c (main): Skip mmap/munmap based test if ASAN is
enabled.
diff --git
On NetBSD 10.0/i386, I see this test failure:
FAIL: test-expl
===
../../gltests/test-exp.h:81: assertion 'err > - err_bound / TWO_MANT_DIG && err
< err_bound / TWO_MANT_DIG' failed
FAIL test-expl (exit status: 134)
The expl() function is just more impreci
With cc from Oracle Developer Studio 12.6 on Linux, I see a test failure:
FAIL: test-mcel
===
FAIL test-mcel (exit status: 1)
This compiler does not implement the rule (from ISO C 99 § 5.1.2.2.3) that
the implicit return value from main() is 0. How to reproduce
Bruno Haible
strtod, strtold tests: Avoid a test failure on native Windows.
* modules/strtod-tests (Depends-on): Add setlocale.
* modules/strtold-tests (Depends-on): Likewise.
diff --git a/modules/strtod-tests b/modules/strtod-tests
index cbc8030ca4..b6fc4585bf 100644
--- a
On OpenBSD 7.4/arm64, I'm seeing this test failure:
$ ./test-fenv-round
../../gltests/test-fenv-round.c:53: assertion 'p < a' failed
Abort trap (core dumped)
As usual, the problem is that the soft-float functions
(__divtf3, __eqtf2
On NetBSD 9.3 (x86_64, arm64, powerpc) I see a test failure:
../../gltests/test-get-rusage-data.c:66: assertion 'value3 > value1' failed
FAIL test-get-rusage-data (exit status: 134)
This patch acknowledges the NetBSD behaviour, like the one on FreeBSD and
OpenBSD.
2024-01-18
On FreeBSD 12.2/sparc64 I see this test failure:
$ ./test-time
../../gltests/test-time.c:46: assertion 'tt3 >= tv2.tv_sec' failed
Abort trap (core dumped)
This patch fixes it.
2024-01-11 Bruno Haible
time: Fix test failure on FreeBSD 12.2/sparc64.
* m4/time.m4
Yesterday I did:
> (structptr_to_funcptr, funcptr_to_structptr): New functions/macros.
> (xcopy_structptr): Renamed from xcopy_funcptr. Mark as inline.
> (COPY_FUNCPTR, CODE): Use structptr_to_funcptr, funcptr_to_structptr.
This produces gcc -Wstrict-aliasing warnings on Linux/po
That patch caused some compiler warnings. Fixed like this:
2024-01-10 Bruno Haible
jit/cache tests: Avoid compiler warnings on OpenBSD.
* tests/jit/test-cache.c (main): New variables size_of_return1,
size_of_return2.
diff --git a/tests/jit/test-cache.c b/tests/jit/tes
ant), the best workaround is to
include the precompiled code of 'return1' (and 'return2') in the test program.
Done as follows:
2024-01-09 Bruno Haible
jit/cache tests: Avoid test failure on OpenBSD.
* tests/jit/test-cache.c (return1_code, return2_code): N
This patch fixes a test failure on Linux/hppa and NetBSD/hppa.
The cause of the failure was that on 32-bit hppa platforms, a function pointer
is not merely a 'struct func *', but a 'struct func *' + 2.
2024-01-09 Bruno Haible
jit/cache tests: Avoid test
for this
specific compilation unit.
[1] https://lwn.net/Articles/732201/
[2] https://lwn.net/Articles/732202/
[3] https://news.ycombinator.com/item?id=17246537
2024-01-08 Bruno Haible
jit/cache tests: Avoid test failure on OpenBSD.
* modules/jit/cache-tests (Files): Add m4/
memory.
Also, some of the #if conditions were incorrect:
- '#if defined __hppa__' is needed instead of '#if __hppa__'.
- Allow for __hppa__ to be set in __hppa64__ mode, and allow for __ia64__
to be set in 32-bit ia64 mode.
2024-01-08 Bruno Haible
jit/cac
On NetBSD 8.0/sparc I see a test failure of test-fenv-except-trapping-2.sh.
This patch avoids the failure.
2023-11-05 Bruno Haible
fenv-exceptions-trapping: Avoid test failure on NetBSD/sparc.
* tests/test-fenv-except-trapping-2.c (main): Skip the 'long double'
On some older Linux distros, I see the fenv-exceptions-trapping-2 test fail.
The reason are older versions of gcc, which generate different code for
floating-point comparisons.
2023-11-05 Bruno Haible
fenv-exceptions-trapping: Avoid test failure on older systems.
* tests/test
In a testdir of all of gnulib, on Mac OS X 10.5.8, I see this test failure:
FAIL: test-update-copyright.sh
This is the log file:
$ cat test-update-copyright.sh.log
--- - 2023-10-03 10:28:56.0 +0200
+++ update-copyright.test-ex-stderr 2023-10-03 10:28:56.0 +0200
@@ -1,2
On AIX 7.3.1, with ibm-clang as compiler, I see a unit test failure:
FAIL: test-aligned_alloc
../../gltests/test-aligned_alloc.c:52: assertion 'aligned2_blocks[i] != NULL'
failed
FAIL test-aligned_alloc (exit status: 134)
The problem here is that ibm-clang, w
This patch avoids two long-standing test failures on MSVC.
2023-07-28 Bruno Haible
mbmemcasecmp, mbmemcasecoll: Avoid test failure on MSVC.
* tests/test-mbmemcasecmp.h (test_utf_8): Disable two tests on platforms
that don't have the upper/lower mappings for &
Andrew Schulman wrote:
> So hooray for automated testing! It found a problem.
And that wasn't the only bug that this unit test found. A bigger one
was reported in [1] and fixed in [2].
Bruno
[1] https://lists.gnu.org/archive/html/bug-gnulib/2020-03/msg00085.html
[2] https://lists.gnu.org/archive
> Hi,
>
> Andrew Schulman wrote:
> > I'm building argp for Cygwin. The build succeeds, and all of the tests
> > pass except for one, test-argp-2.sh.
>
> For me, on Cygwin, this test passes.
>
> > The contents of test-suite.log are below.
> > ...
> > The error is pretty trivial, and I think it's
Hi,
Andrew Schulman wrote:
> I'm building argp for Cygwin. The build succeeds, and all of the tests
> pass except for one, test-argp-2.sh.
For me, on Cygwin, this test passes.
> The contents of test-suite.log are below.
> ...
> The error is pretty trivial, and I think it's incorrect. If I
> unde
I'm building argp for Cygwin. The build succeeds, and all of the tests
pass except for one, test-argp-2.sh. The contents of test-suite.log
are below.
The error is pretty trivial, and I think it's incorrect. If I
understand the source in lib/argp-help.c correctly
(https://git.savannah.gnu.org/gitwe
On cfarm105.cfarm.net, I'm seeing this test failure:
../../build-aux/test-driver: line 112: 3135129 Aborted "$@" >>
"$log_file" 2>&1
FAIL: test-sigdescr_np
This patch fixes it.
2023-06-25 Bruno Haible
sigdescr_np te
Hi/2.
Bruno Haible wrote:
> KO Myung-Hun wrote:
>>> dirfd is specified at
>>> https://pubs.opengroup.org/onlinepubs/9699919799/functions/dirfd.html
>>> Which says that if "The dirp argument does not refer to a valid directory
>>> stream"
>>> the function may fail with error EINVAL. It may alterna
KO Myung-Hun wrote:
> > dirfd is specified at
> > https://pubs.opengroup.org/onlinepubs/9699919799/functions/dirfd.html
> > Which says that if "The dirp argument does not refer to a valid directory
> > stream"
> > the function may fail with error EINVAL. It may alternatively simply crash
> > ("the
Bruno Haible wrote:
> KO Myung-Hun wrote:
>>> @@ -78,11 +82,17 @@ _gl_unregister_dirp_fd (int fd)
>>> int
>>> dirfd (DIR *dir_p)
>>> {
>>> +#if GNULIB_defined_DIR
>>> + int fd = dir_p->fd_to_close;
>>
>> dirfd() allows to take an invalid `dir_p' ?
>
> No; this must be a misunderstanding.
>
KO Myung-Hun wrote:
> > @@ -78,11 +82,17 @@ _gl_unregister_dirp_fd (int fd)
> > int
> > dirfd (DIR *dir_p)
> > {
> > +#if GNULIB_defined_DIR
> > + int fd = dir_p->fd_to_close;
>
> dirfd() allows to take an invalid `dir_p' ?
No; this must be a misunderstanding.
dirfd is specified at
https://p
Hi/2.
Bruno Haible wrote:
> On native Windows, I see a test failure:
>
> FAIL: test-fdopendir
>
>
> ../../gltests/test-fdopendir.c:76: assertion 'dup2 (fd, fd) == -1' failed
> FAIL test-fdopendir.exe (exit status: 3)
>
> The cause
Running a testdir on Solaris 10/SPARC with Sun C, I see a test failure:
$ ./test-hamt
$ echo
On native Windows, I see a test failure:
FAIL: test-fdopendir
../../gltests/test-fdopendir.c:76: assertion 'dup2 (fd, fd) == -1' failed
FAIL test-fdopendir.exe (exit status: 3)
The cause of the problem is that POSIX [1] says:
"Upon calling closedir() the
On native Windows, in a testdir of module 'fchdir' (but not in a testdir of
all of Gnulib), I'm seeing a test failure:
test-fchdir.c:74: assertion 'fchdir (fd) == 0' failed
What happens? fchdir.c contains logic for keeping track which file
descriptor corresponds to which
On 32-bit mingw I see this test failure:
FAIL: test-expm1
../../gltests/test-expm1.h:49: assertion 'y >= - x' failed
FAIL test-expm1.exe (exit status: 3)
The failure disappears when compiling without -O2. This is a hint that,
again, it's an issue w.r.t. the 6
1)
Since our documentation says that POLLHUP is not fixed by this module
anyway (it does not work on AIX either), let me disable this part of the
test.
2023-04-25 Bruno Haible
poll tests: Avoid test failure on native Windows.
* tests/test-poll.c (test_pipe): Disable the POL
I did:
> + set_binary_mode (STDOUT_FILENO, O_BINARY);
Oops, that produces a compilation error (reported by the 'grep' CI and
the 'sed' CI):
../../gltests/test-dfa-match-aux.c: In function ‘main’:
../../gltests/test-dfa-match-aux.c:61:20: error: ‘STDOUT_FILENO’ undeclared
(first use in this func
test-select.exe (exit status: 1)
Note that only the first test ("Invalid nfd test") fails.
For both functions, the implementation is Gnulib's override.
So, the fix is easy. It makes these two tests succeed.
2023-04-24 Bruno Haible
select, pselect: Fix test failure on
On mingw 10, I see this test failure:
FAIL: test-string-desc.sh
=
../../gltests/test-string-desc.c:113: assertion 'string_desc_write (3, s1) ==
0' failed
FAIL test-string-desc.sh (exit status: 1)
Apparently, in the Cygwin environment, it is not possible to pa
one:
2023-04-22 Bruno Haible
dfa tests: Fix test failure on mingw differently.
Suggested by Paul Eggert.
* tests/test-dfa-invalid-char-class.sh: Revert last change.
* tests/test-dfa-match-aux.c: Include binary-io.h.
(main): Put stdout into binary mode.
On 4/20/23 15:20, Bruno Haible wrote:
How to get further? Should I add this specific mktime call as a test
case to the 'mktime' module's tests?
If you're interested in pursuing that I suppose it wouldn't hurt.
The Microsoft documentation
https://learn.microsoft.com/en-us/cpp/c-runtime-library
Hi Paul,
As the nstrftime test is failing on several platforms [1], I'm trying to
debug it.
Now I'm on mingw 10. The output of test-nstrftime is sometimes
PST8PDT,M3.2.0,M11.1.0: expected "1969-12-31 16:00:00 -0800 (PST)", got
"1969-12-31 16:00:00 (PST)"
MST7: expected "1969-12-31 17:00:00 -07
Paul Eggert wrote:
> > +LC_ALL=C ${CHECKER} test-dfa-match-aux '[[:foo:]]' a > tmp 2>&1
> > +LC_ALL=C tr -d '\r' < tmp > out
>
> This sort of change means we won't catch stray CRs in the output.
>
> Is there some way we can instead compile or run MinGW apps so that they
> instead follow the POSI
On 4/20/23 09:39, Bruno Haible wrote:
+LC_ALL=C ${CHECKER} test-dfa-match-aux '[[:foo:]]' a > tmp 2>&1
+LC_ALL=C tr -d '\r' < tmp > out
This sort of change means we won't catch stray CRs in the output.
Is there some way we can instead compile or run MinGW apps so that they
instead follow the
On mingw 10, I see this test failure:
FAIL: test-readtokens.sh
--- exp 2023-04-20 06:20:32.200323300 +0200
+++ out 2023-04-20 06:20:32.232259800 +0200
@@ -1 +1 @@
-a:b:c:
+a:b:c:
--- exp 2023-04-20 06:20:32.278874900 +0200
+++ out 2023-04-20 06:20:32.310364100 +0200
On mingw 10, I see this test failure:
FAIL: test-parse-duration.sh
../../gltests/test-parse-duration.sh: line 64: test: 38898367
: integer expression expected
test-parse-duration fatal error: 38898367
is not 38898367
FAIL test-parse-duration.sh (exit status: 1
On mingw 10, I see this test failure:
FAIL: test-dfa-invalid-char-class.sh
--- exp 2023-04-20 06:15:56.778614400 +0200
+++ out 2023-04-20 06:15:56.857719400 +0200
@@ -1 +1 @@
-dfaerror: invalid character class
+dfaerror: invalid character class
FAIL test-dfa
On Cygwin 3.4.6, I see this test failure:
FAIL: test-c32isgraph.sh
../../gltests/test-c32isgraph.c:125: assertion 'is != 0' failed
../../gltests/test-c32isgraph.sh: line 13: 57088 Aborted (core
dumped) LC_ALL=$LOCALE_FR ${CHECKER} ./test-
ilures is a problem in Gnulib's test. Fixed like this:
2023-04-16 Bruno Haible
posix_spawn_file_actions_addclose tests: Avoid test failure on Cygwin.
* tests/test-posix_spawn_file_actions_addclose.c (main): Skip an
unportable test also on Cygwin.
On Slackware 14, I see this test failure:
FAIL: test-striconveh
=
../../gltests/test-striconveh.c:510: assertion 'length == strlen (expected)'
failed
FAIL test-striconveh (exit status: 134)
The conversion produces a question mark instead of the expected U+FFFD
On mingw, in a testdir of vasnwprintf-posix, I see a test failure:
../../gltests/test-vasnwprintf-posix.c:4209: assertion 'wcscmp (result,
L"000303c 33") == 0' failed
FAIL test-vasnwprintf-posix.exe (exit status: 3)
This patch fixes it.
2023-04-07 Bruno Haible
On Haiku, I see a test failure of test-string-desc-quotearg.
The reason is that on this platform, the C locale has UTF-8 encoding,
and therefore the quotearg.c logic uses U+2018 and U+2019 as quotation
marks, instead of the ASCII double-quote.
2023-04-05 Bruno Haible
string-desc
On OpenBSD 7.2, I see a test failure for vasnwprintf. The cause is that
we use %n in the swprintf format string, and OpenBSD disabled that in 2021.
We have already stopped using %n on OpenBSD in the snprintf format string
in 2020, see
https://lists.gnu.org/archive/html/bug-gnulib/2020-10/msg00013
Also on AIX 7.2, I see the 'time' test fail. Here too, for ca. 5 milliseconds
after gettimeofday() reports a new second, time(NULL) reports the previous one.
This patch fixes it.
2023-03-17 Bruno Haible
time: Fix test failure on AIX 7.2.
* m4/time.m4 (gl_FUNC_TI
rts a new second, it takes about 5 msec
until time(NULL) reports that new second as well.
This patch fixes it, by implementing time(NULL) based on gettimeofday().
Note that the mingw gettimeofday() and the gnulib replacement are equivalent.
2023-03-17 Bruno Haible
time: Fix test failu
On Solaris 11, I see a test failure:
../../gnulib-tests/test-nl_langinfo2.c:100: assertion 'strlen (fr_T_FMT_AMPM)
== 0 || strcmp (fr_T_FMT_AMPM, "%I:%M:%S") == 0' failed
../../gnulib-tests/test-nl_langinfo2.sh: line 12: 17260: Abort(coredump)
The problem is that
d in through
ssh: getgid() and getegid() both return (gid_t)(-1). But the st_gid of
freshly created files is reasonable.
This patch avoids the test failures. I could also make the entire test SKIPped,
but that would be exaggerated.
2023-02-10 Bruno Haible
chown, lchown, fchownat tests:
On 1/17/23 06:43, Bruno Haible wrote:
On Android 11, within Termux, I see this test failure:
FAIL: test-nstrftime
PST8PDT,M3.2.0,M11.1.0: expected "1969-12-31 16:00:00 -0800 (PST)", got "1969-12-31
17:00:00 -0700 (PDT)"
CET-1CEST,M3.5.0,M10.5.0/3: exp
Eric Blake wrote:
> > Apparently the SIGPIPE handler gets set to SIG_IGN, either by 'make' or
> > by the test-driver.
>
> Known regression in make 4.4:
> https://savannah.gnu.org/bugs/index.php?63307
Thanks. Indeed, I can reproduce the test failure also on a glibc
On Tue, Jan 17, 2023 at 03:21:52PM +0100, Bruno Haible wrote:
> On Android 11, I'm seeing a test failure:
>
>
> Apparently the SIGPIPE handler gets set to SIG_IGN, either by 'make' or
> by the test-driver.
Known regression in make 4.4:
https://savannah.gnu.org/b
Hi Paul,
On Android 11, within Termux, I see this test failure:
FAIL: test-nstrftime
PST8PDT,M3.2.0,M11.1.0: expected "1969-12-31 16:00:00 -0800 (PST)", got
"1969-12-31 17:00:00 -0700 (PDT)"
CET-1CEST,M3.5.0,M10.5.0/3: expected "1970-01-01 01:00:00
On Android 11, I'm seeing a test failure:
FAIL: test-execute.sh
=
test-execute-main: test-execute-child subprocess failed
../../gltests/test-execute-main.c:138: assertion 'ret == 127' failed
Aborted
test-execute.sh: test case 3 failed
../../gltests/test-exe
On Android 11, I'm seeing this test failure:
FAIL: test-spawn-pipe.sh
../../gltests/test-spawn-pipe-child.c:134: assertion '! is_open
(STDERR_FILENO)' failed
test-spawn-pipe.sh: iteration 4 failed
../../gltests/test-spawn-pipe-child.c:134: asser
On Android 11 with Termux, I see this test failure:
FAIL: test-posix_spawn-fchdir
=
cannot open directory: Permission denied
FAIL test-posix_spawn-fchdir (exit status: 1)
The reason is that the "/" directory cannot be opened. But other directories,
like
in the libtool wrapper's find_executable function.)
2022-09-14 Bruno Haible
posix_spawn-* tests: Fix test failure on mingw when libtool is in use.
* tests/test-posix_spawn-inherit0.c (CHILD_PROGRAM_FILENAME): Add the
EXEEXT suffix.
* tes
In a testdir of nearly all of gnulib, I see a test failure:
FAIL: test-pipe-filter-gi2.sh
The logs reveal that the test-pipe-filter-gi2 test 3 fails, with status 141.
141 being 128 + SIGPIPE, the suspicion falls on the 'sigpipe' module. And
indeed, the failure occurs only if t
fails.
This patch fixes both issues.
2022-09-04 Bruno Haible
count-leading-zeros: Fix a link error on 32-bit MSVC and a test failure.
* lib/count-leading-zeros.h: Correct syntax for #pragma intrinsic.
(COUNT_LEADING_ZEROS): Fix the return value
27; gets linked against
an old libunistring under /usr, and the symbol uc_general_category_byname
is taken from there, whereas the UC_CATEGORY_* variables are linked statically
from the local libgnu.a.
This patch fixes it.
2022-02-09 Bruno Haible
unictype/category-byname: Fix test failu
With nvc 22.1, I observe a test failure:
FAIL: test-log
==
../../gltests/test-log.h:68: assertion 'err > - err_bound / TWO_MANT_DIG && err
< err_bound / TWO_MANT_DIG' failed
This patch works around it, by allowing a slightly larger error bound.
Paul Eggert wrote:
> > I can't determine whether it's a GCC or a QEMU bug.
>
> Surely it is a QEMU bug, as QEMU isn't properly implementing the
> z/Architecture Principles of Operation (POP).
Thanks; I didn't know about this document.
> Page 7-25 of the current
> edition (SA22-7832-12) says th
On 9/12/21 8:28 AM, Bruno Haible wrote:
I can't determine whether it's a GCC or a QEMU bug.
Surely it is a QEMU bug, as QEMU isn't properly implementing the
z/Architecture Principles of Operation (POP). Page 7-25 of the current
edition (SA22-7832-12) says that the AGR instruction treats its o
Paul Eggert wrote:
> That value of 'sum' is wrong; it should be most-negative value {tv_sec =
> -9223372036854775808, tv_nsec = 0} because A and B are both that value,
> and 'sum = timespec_add (a, b)' is supposed to be saturated addition.
>
> My guess is that timespec-add.c's line 49 'INT_ADD_W
Paul Eggert wrote:
> Which version of GCC are you using?
s390x-linux-gnu-gcc-10 (Ubuntu 10.3.0-1ubuntu1~20.04) 10.3.0
> I suspect the bug is in the earlier call to timespec_add, as mentioned
> above.
I'll debug that call...
Bruno
1 - 100 of 396 matches
Mail list logo
