On Sun, 14 Sep 2025, Kirill Makurin wrote:
When I was running `make check` with 32-bit msvcrt.dll I noticed one
more test failure, which was a bug in sem_timedwait32. The first patch
in this series adds a fix.
FWIW, I saw one spurious test failure with this patchset on github actions
as well:
FAIL: pthread_rwlock/rwlock6_t
==============================
Assertion failed: (rd2Result == 20), file
../../tests/pthread_rwlock/rwlock6_t.c, line 121
FAIL pthread_rwlock/rwlock6_t.exe (exit status: 1)
This happened on x86_64 once.
It's of course not a huge issue if it happens only once, but if we are
going to be running these tests on each push, it will start diminish the
value of all the CI if we have common spurious failures in there (even
though it probably goes away by restarting that job).
So ideally, we'd find and avoid any such spurious issues. (I haven't tried
to reproduce this one myself yet - I've been more than busy enough trying
to get the tests to run reliably in ARM environments. With all the latest
posted patches, they do run fine both in my local test environments and on
github actions finally.)
// Martin
_______________________________________________
Mingw-w64-public mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/mingw-w64-public
