On Thu, Jan 11, 2024 at 04:10:14PM +0100, Ludovic Courtès wrote:
> Hello Guix!
> 
> Several of us have been fiddling with the ‘core-updates’ branch for a
> while.  I think there’s now consensus that the branch is really
> dedicated to core packages and (guix build …) modules, as embodied in
> the new ‘core-packages’ team¹.
> 
> We’ve updated GCC 11.x, glibc, binutils, and various packages from (gnu
> packages base).  Notable exceptions are Coreutils, Findutils, sed, and
> tar; I tried but that’s a bit more work, notably because their variants
> in commencement.scm would no longer build because their build scripts
> use sed patterns not supported by Gash-Utils.
> 
> Long story short: I’d like us to freeze and merge the branch ASAP,
> notably because the glibc graft on ‘master’ leads to a bad user
> experience.  I’m happy with the current state of the branch and wouldn’t
> mind postponing remaining upgrades for the next cycle.
> 
> Thoughts?
> 
> Remaining work includes: checking that cross-compilation targets still
> work after the recent Binutils updates, checking i586-gnu (GNU/Hurd) and
> other platforms, and possibly addressing the Gawk non-determinism
> issue².
> 
> Currently package subsets are built here:
> 
>   https://ci.guix.gnu.org/jobset/core-updates
>   https://guix.bordeaux.inria.fr/jobset/guix-core-updates
> 
> I don’t think I can commit to coordinating the stabilization effort
> though as I’m busy with other things this month.  Would anyone like to
> take the lead on this?
> 
> Happy updating!
> 
> Ludo’.
> 
> ¹ https://issues.guix.gnu.org/67880
> ² https://issues.guix.gnu.org/68378

There's a patch floating around somewhere to adjust the page size on
jemalloc on aarch64 to be at least 64k so that people running guix
software on apple silicon don't have issues.  I think we should add it
for core-updates so it doesn't get forgotten, I've seen it come up on
IRC at least once a week.

-- 
Efraim Flashner   <efraim@flashner.co.il>   רנשלפ םירפא
GPG key = A28B F40C 3E55 1372 662D  14F7 41AA E7DC CA3D 8351
Confidentiality cannot be guaranteed on emails sent or received unencrypted