time_rz.c and Android NDK r27

This new release of the Android NDK defines timezone_t in time.h whether or not it defines the related tzalloc and localtime_rz extensions, which being new APIs are controlled by __ANDROID_API__ as expected. Consequently, when building for earlier Android releases with this NDK, Emacs compilation r

Re: [gnu-prog-discuss] FSF office adddress update

Simon Josefsson via gnu-prog-discuss wrote: Jason Self writes: [...] 31 Milk Street # 960789 Note that according to USPS standardized abbreviations it should be Milk St. I can't find any occurance for this one -- presumably because few maintainers used time machines for preparing lic

Re: doc: Mention a gmtime_r problem

Eric Blake wrote: > But see https://austingroupbugs.net/view.php?id=1797#c6794; the Austin > Group is trying to work with the C committee to get this fixed in the > future (with Paul's help); among other things, it fixes gmtime() and > gmtime_r() to both have to set a value of either "GMT" or "UTC"

Re: doc: Mention a gmtime_r problem

On Tue, Sep 03, 2024 at 11:34:27AM GMT, Bruno Haible wrote: > POSIX specifies that gmtime_r sets the tm_zone field to "UTC". (Interestingly, > no such requirement exists for gmtime.) But see https://austingroupbugs.net/view.php?id=1797#c6794; the Austin Group is trying to work with the C committee

tests: Strenghten tests with invalid file descriptor

Let's see whether some kernel confuses the notion of "file descriptor" with the notion of "file descriptor or AT_FDCWD". A little bit of "let's try to provoke undefined behaviour internally" mindset... 😉 2024-09-03 Bruno Haible tests: Strenghten tests with invalid file descriptor.

Re: pthread_setname_np documentation

Hi Eric, > Hi, I'm writing about gnulib's documentation for pthread_setname_np: > https://www.gnu.org/software/gnulib/manual/html_node/pthread_005fsetname_005fnp.html > It mentions NetBSD 10.0 as a platform where the function has a > different signature, however, I recently discovered that there i

Re: FSF office adddress update

Jason Self writes: > On Tue, 03 Sep 2024 08:54:26 +0200 > Simon Josefsson via gnu-prog-discuss wrote: > >> I wonder if maybe some packages used the abbreviated version >> 'Franklin St' instead? > > Possibly; USPS standardized abbreviations would have Street shortened > to St. > >> Does anyone h

pthread_setname_np documentation

Hi, I'm writing about gnulib's documentation for pthread_setname_np: https://www.gnu.org/software/gnulib/manual/html_node/pthread_005fsetname_005fnp.html It mentions NetBSD 10.0 as a platform where the function has a different signature, however, I recently discovered that there is another platform

unlockpt tests: Avoid test failure on Haiku

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

doc: Mention a gmtime_r problem

POSIX specifies that gmtime_r sets the tm_zone field to "UTC". (Interestingly, no such requirement exists for gmtime.) The comments in tests/test-gmtime_r.c remind us that this is not always the case. 2024-09-03 Bruno Haible doc: Mention a gmtime_r problem. * doc/posix-functio