libc6-dev:amd64 overwrites files from libc6:i386 and aborts update when upgrading to mantic

The issue isn't because of libc6-dev:i386, which is a multiarch package and should cope just fine with sharing an identical file with its :amd64 counterpart. The issue is in the biarch package libc6-dev-i386:amd64.

What happened is that the fortran header was moved into an arch-specific directory in the -dev packages for the 2.38, and instead the biarch packages inserted a symlink in the generic path. That particular change was picked up when I merged the Debian changes, but I didn't notice the symlink part. What this means is that the 2.38 biarch packages need to declare a Breaks+Replace relationship on the -dev packages << 2.37.

This will be fixed in an upcoming SRU. Meanwhile, to be able to upgrade the easiest thing that comes to mind is to remove the biarch package before upgrade and reinstall afterwards. Sorry for the inconvenience!

Thanks, for some reason running `apt-get --fix-broken install` multiple times in a row fixed it. I have no explanation.

I got a similar conflict bug with libgutils but this other one is not fixed with multiple `apt-get --fix-broken install`, see bug 2039279 .

On Sat, 14 Oct 2023 at 02:40, Thomas Debesse <email address hidden>
wrote:

> Thanks, for some reason running `apt-get --fix-broken install` multiple
> times in a row fixed it. I have no explanation.
>

AIUI this isn't really surprising, there is no problem once all packages
are at their newest versions, we just need to add metadata so that dpkg
unpacks them in the right order.

Alright, after a chat with Steve, it makes more sense to handle this via an upgrader quirk than an SRU of glibc in Mantic, although it *is* handled in the package that's been uploaded to Noble.

I'm realizing that I didn't try to reproduce this. After doing so in a container, the issue didn't present itself, presumably because ubuntu-release-upgrader actually uses --force-overwrite by default precisely to avoid these issues.

May I ask how exactly you tried to upgrade?

This bug was fixed in the package glibc - 2.38-3ubuntu1

---------------
glibc (2.38-3ubuntu1) noble; urgency=medium

  * debian/patches/git-updates.diff: update from upstream stable branch
    Dropped changes, superseded by the upstream git updates:
    - debian/patches/CVE-2023-4911.patch: terminate immediately if end of
      input is reached in elf/dl-tunables.c.
    - d/p/u/0001-Fix-leak-in-getaddrinfo-introduced-by-the-fix-for-CV:
      Cherry-picked to fix a regression in one of the previous CVE fixes
  * Merge 2.38-3 from Debian experimental
    Dropped changes, included in Debian:
    - debian/patches/hurd-i386/git-powerpc-longjmp.diff: Fix build after chk
      hidden builtin fix.
  * Drop d/p/lp2032624.patch as advised by upstream.
    Downstream users will have to actually implement those types or stop
    pretending they're GCC. (LP: #2032624)
  * d/p/lp2031495.patch: fix test suite on armhf for -prof variant
    (LP: #2031495)
  * d/control.in/i386: fix math-vector-fortran.h file move (LP: #2039234)

 -- Simon Chopin <email address hidden> Mon, 23 Oct 2023 18:54:07 +0200