* m4/pid_t.m4 (AC_TYPE_PID_T): Redefine this Autoconf macro
only for Autoconf versions 2.69 and earlier, since 2.70
will be fixed.
---
ChangeLog | 7 +++
m4/pid_t.m4 | 7 ++-
2 files changed, 13 insertions(+), 1 deletion(-)
diff --git a/ChangeLog b/ChangeLog
index 073f96870..c6b057e68 1
On 64-bit Windows, pid_t needs to be 64-bit large, i.e. intptr_t or 'long long'
or '__int64'.
This is because the return type of _spawnv* (when invoked with _P_NOWAIT) and
the argument of the _cwait function are 'intptr_t' (see [1][2]: "The return
value from an asynchronous _spawnvp or _wspawnvp (
> 2020-07-21 Bruno Haible
>
> inttypes: Fix PRI*PTR and SCN*PTR on 64-bit native Windows.
> * m4/inttypes.m4 (gl_INTTYPES_PRI_SCN): On 64-bit native Windows, make
> sure PRIPTR_PREFIX is defined to "ll", not "l".
Oops, there was a typo in this patch, leading to PRIdPTR = "lld
On 8/23/20 7:28 AM, Bruno Haible wrote:
Thanks for the report. I didn't know that __typeof__ does not work for all
kinds of expressions. (intprops uses __typeof__ as well.)
Almost all the places intprops.h uses __typeof__ are safe since they can't or
shouldn't be bitfields. (Pretty much anyth
Jörg Sonnenberger noticed that the new gnulib module 'supersede', when
writing to /dev/null in Solaris/Illumos zones, fails [1][2].
Per POSIX [3],
open ("/dev/null", O_TRUNC | O_WRONLY)
and
open ("/dev/null", O_CREAT | O_TRUNC | O_WRONLY)
should be equivalent. But in Solaris/Illumos zones, the
Hello Bruno,
>> In file included from regex.c:74:
>> ./regexec.c:1198:3: error: invalid application of 'typeof' to bit-field
>> DEBUG_ASSERT (state->halt);
>> ^
>
> Thanks for the report. I didn't know that __typeof__ does not work for all
> kinds of expressions. (intprops uses __typeof__ as
Hi Benno,
> Running a compile test of nano with the current state of gnulib
> (commit 452fee36, "verify: Do use __builtin_assume on clang")
> on FreeBSD (which happens to use clang), the compile failed:
>
> clang -DHAVE_CONFIG_H -I. -I.. -D_GNU_SOURCE -D_DEFAULT_SOURCE
> -I/usr/include/ncursesw
Paul Eggert wrote:
>+ libc-config: avoid Clang’s __diagnose_if__
>+ * lib/cdefs.h (__warndecl, __warnattr, __errordecl):
>+ For now, do not use __diagnose_if__ here, as this fails
>+ on Fedora 31 with Clang 9.0.1, with diagnostic
>+ "/usr/include/bits/stdio2.h:263:9: error:
Hi,
Running a compile test of nano with the current state of gnulib
(commit 452fee36, "verify: Do use __builtin_assume on clang")
on FreeBSD (which happens to use clang), the compile failed:
clang -DHAVE_CONFIG_H -I. -I.. -D_GNU_SOURCE -D_DEFAULT_SOURCE
-I/usr/include/ncursesw -g -O1 -march=
