Bug#834855: transition: glibc 2.24
Control: forwarded -1 https://release.debian.org/transitions/html/glibc-2.24.html
Control: tags -1 confirmed
On 19/08/16 23:47, Aurelien Jarno wrote:
> Package: release.debian.org
> Severity: normal
> User: release.debian.org@packages.debian.org
> Usertags: transition
>
> Dear release team,
>
> We would like to get a transition slot for glibc 2.24. It is currently
> available in experimental and has been built successfully on all
> official architectures. For the debian-ports architectures the
> situation also looks good except for hppa with multiple regressions in
> the testsuite due to the use of gcc-6 instead of gcc-5. This is being
> worked on, and I don't think we should not block the transition on that.
>
> The glibc 2.24 is built using gcc-6 instead of gcc-5, the transition is
> therefore important for the gcc-5 removal. This will be the version
> shipped with Stretch, so this is the last major change before the
> release. Of course we plan to improve things by fixing additional bugs,
> mostly by tracking the upstream 2.24 stable branch as long as possible.
>
> As the glibc is using symbol versioning, there is no soname change. That
> said a few packages are using libc internal symbols and have to be
> rebuilt for this transition:
> - apitrace
> - bro
> - dante
> - libnih
> - libnss-db
> - unscd
>
> Here is the corresponding ben file:
> title = "glibc";
> is_affected = .depends ~ /libc[0-9.]* \(<</;
> is_good = .depends ~ /libc[0-9.]* \(<< 2.25\)/;
> is_bad = .depends ~ /libc[0-9.]* \(<< 2.24\)/;
>
> This version will also make a few packages to fail to build from source
> due to the removal of deprecated part of the API. The existing binaries
> still work (the ABI hasn't changed), so there this doesn't cause any
> issue for upgrades. Bugs have been reported against these packages:
> #833727 - emacs24: FTBFS with glibc 2.24
> #833965 - tcsh: uses deprecated BSD union wait type
> #833968 - csh: uses deprecated BSD union wait type
> #833970 - netkit-rsh: uses deprecated BSD union wait type
> #833971 - xautolock: uses deprecated BSD union wait type
> #833972 - socks4-server: uses deprecated BSD union wait type
> #833995 - gridengine: embedded tcsh copy uses deprecated BSD union wait type
> While it has been done relatively recently they all contain a patch, so
> I don't think we should consider them as blockers for the transition.
>
> In addition a few new symbols have been added that might prevent a few
> other packages to transition to testing if they pick up the new symbols.
> They are quick_exit for libc.so and nextdown*/nextup* for libm.so.
Sorry for the delay, I was away.
Go ahead.
Cheers,
Emilio
Reply to: