posix_spawn-internal: improvements on native Windows

2022-09-09 Thread Bruno Haible
Here are some improvements for the posix_spawn* implementation on native Windows. The first patch is a small fix (only visible if, among the file action, there is an _addopen of a file in O_APPEND mode, followed by a _adddup2 of that file descriptor). The next 4 patches optimize away most

spawn-pipe: Fix pipe-filter-* test hangs (regression 2020-12-24)

2022-09-09 Thread Bruno Haible
Since 2020-12-25, the 4 pipe-filter-* tests were hanging on native Windows. This patch, together with the previous one, fixes the issue. The change that triggered the malfunction was switching the spawn-pipe implementation from a code specific to native Windows to a generic code that uses

spawn-pipe: Fix possible hangs in programs that spawn several children

2022-09-09 Thread Bruno Haible
This patch fixes a possible hang in programs that use the 'spawn-pipe' module and also spawn other child processes. It is also needed for fixing the hanging pipe-filter-* tests on native Windows. 2022-09-09 Bruno Haible spawn-pipe: Fix possible hangs in programs that spawn several

pipe-filter-gi tests: Fix long-standing failure on native Windows

2022-09-09 Thread Bruno Haible
In a testdir with SPAWN_PIPE_IMPL_AVOID_POSIX_SPAWN set to 1 instead of 0, there is a test failure: FAIL: test-pipe-filter-gi2.sh This patch fixes it. Similar to the corresponding patch "pipe-filter-ii tests: Fix long-standing failure on native Windows" from 2021-06-15. 2022-09-09 Bruno

Re: Fwd: gnulib posix_spawn_file_actions_addclose

2022-09-09 Thread Bruno Haible
I committed: > 2022-09-07 Bruno Haible > > posix_spawn_file_actions_addclose tests: Avoid test failure on musl. > Reported by Valery Ushakov in > . > * modules/posix_spawn_file_actions_addclose-tests

Re: [PATCH] Never call gl_FUNC_GETCWD_PATH_MAX on Darwin

2022-09-09 Thread Kirill A. Korinsky
Thanks for the feedback. I guess that the right approach to fix it is described here: https://lists.gnu.org/archive/html/bug-gnulib/2022-01/msg00085.html -- wbr, Kirill signature.asc Description: Message signed with OpenPGP

Re: bool and C23

2022-09-09 Thread Bruno Haible
Paul Eggert wrote: > To help move ahead on this, I propose the attached patches to Gnulib. > They use a simpler version of the Autoconf macro, named gl_C_BOOL to > avoid collision with any future improvements to Autoconf. My hope is > that gl_C_BOOL can be renamed to AC_C_BOOL, or something