On 21/09/2025 23:49, Bruno Haible wrote:
Collin Funk wrote:
I also pushed an update to the Gnulib submodule [1].
Note that there was a regression in Gnulib, that affects coreutils,
in the last two days [1]. I would recommend to wait until that is fixed,
then run a CI build on GitHub to verify the status.
Bruno
[1] https://lists.gnu.org/archive/html/bug-gnulib/2025-09/msg00282.html
I was about to say that the fchownat() changes looked too invasive at least.
That gnulib update pulled in:
* d947a2d927 pthread-once: Improve configuration.* e4c911bb44 tests: avoid grep
in init.sh
* 8ac8a9b4f2 tests: avoid test failure due to rmdir race
* 1f8f8f770c doc: Update for Android API level 37.
* b540161654 doc: Update for Android API level 36.
* eb2c639c9c chown, lchown: Fix regression on Cygwin 2.9.0 (regression today).
* 7b85c2a8da lchown: Fix compilation error on macOS (regression today).
* 12171837a9 chown: simplify CHOWN_CHANGE_TIME_BUG checking
* 162ce0b783 fchownat: fix security races and other bugs
* f7b84b35dd pthread-once: Fix link error on glibc < 2.34 systems (regr.
yesterday).
* cfef893be2 sig2str: Port to Android API level 36.
* c751b47168 qsort_r: Port to Android API level 36.
* 4bdd70e3cf lchmod: Port to Android API level 36.
* d941c98dbf pthread-once: Reduce link dependencies.
* 8ebd5a9aa7 nl_langinfo: Work around nl_langinfo multithread-safety bug on
macOS 26.
* ad4d97cbad fchownat: fix unlikely EOVERFLOW bug
* c48ee317b7 fchownat: fix missing HAVE_FCHOWNAT
* 92e7a7b130 fchownat: depend on fstatat
* dc1dc5630e stringeq: Fix use in C++ mode.
* ba34a934c4 map-c++-tests: Fix compilation error (regression 2025-09-17).
* 153cf1abe6 stack-trace: Add support for Oracle DeveloperStudio cc.
I won't cut a snapshot now, given Bruno's comment.
I think I'll revert the gnulib update for now,
and we can decide tomorrow what to release with.
cheers,
Padraig
