Bug#1141940: transition: glibc

Hi Emilio,

On 2026-08-10 12:56, Emilio Pozuelo Monfort wrote:
> Control: tags -1 confirmed
> 
> On 12/07/2026 19:20, Aurelien Jarno wrote:
> > Package: release.debian.org
> > Severity: normal
> > X-Debbugs-Cc: [email protected], [email protected]
> > Control: affects -1 + src:glibc
> > User: [email protected]
> > Usertags: transition
> > 
> > Dear release team,
> > 
> > I would like to request a transition slot for glibc 2.43. It has been
> > available in experimental for a few months and appears to be in good
> > shape. It has built successfully on all release architectures and on
> > most ports architectures.
> > 
> > One important change that causes compatibility issues is that, when
> > using the C23 standard (default with GCC 15), the functions bsearch,
> > memchr, strchr, strpbrk, strrchr, strstr, wcschr, wcspbrk, wcsrchr,
> > wcsstr and wmemchr now return a pointer to a const-qualified type when
> > the input argument is a pointer to a const-qualified type. A possible
> > quick workaround if we need to fix a package for the transition is to
> > force the ISO C level with -std=c17 or -std=gnu17 or use GCC 14.
> > An archive rebuild has been done, and the corresponding bugs have been
> > reported [1], as well as the ones causing autopkgtest failures. We are
> > now down to less than 40 bugs.
> > 
> > On the autopkgtest side, the experimental pseudo-excuses look good
> > overall, and the remaining ones due to the C23 change, issues in
> > unstable or britney selecting the wrong set of packages.
> > 
> > As glibc is using symbol versioning, there is no soname change. That
> > said a few packages use internal libc symbols and have to be
> > rebuilt for this transition. Here is the corresponding ben file:
> >    title = "glibc";
> >    is_affected = .depends ~ /libc[0-9.]* \(<</;
> >    is_good = .depends ~ /libc[0-9.]* \(<< 2.44\)/;
> >    is_bad = .depends ~ /libc[0-9.]* \(<< 2.43\)/;

Looking back at my notes, I realized there is also the socket-wrapper to 
binNMU, as it encodes the expected size of struct tcp_info in the tests 
instead of dynamically determining it at runtime.

> > Speaking about migration, this glibc version adds a few new GLIBC_2.43
> > symbols, to support ISO C23 (free_sized, free_aligned_sized,
> > memset_explicit and memalignment), and to wrap to Linux kernel syscalls
> > (mseal and openat2). These symbols are unlikely to be used widely at
> > this stage, and thus this transition should not block many packages to
> > migrate to testing.
> > 
> > Thanks for considering.
> 
> Please go ahead, and bump the remaining bugs to serious.

Thanks, I have just uploaded it and bumped the remaining bugs to 
serious.

Matthias it also means that you can upload cross-toolchain-base and 
cross-toolchain-base-ports based on glibc 2.43. Thanks!

Regards
Aurelien