Status for RT on `Rules-Requires-Root: no` transition

Niels Thykier <[email protected]> Tue, 14 Jan 2025 10:46:05 +0100
Newsgroups	gmane.linux.debian.devel.release,gmane.linux.debian.devel.dpkg.general
Message-ID	<[email protected]>
Hi Release team

Another update on the transition

# High level summary

  * 41 RC bugs still affect testing out of 229 bugs known.

  * Fixes for 18 RC bugs are expected later today, bringing us down
    to 23 RC bugs.

  * 13 RC bugs affect key packages in testing with fixes for 3 of them
    expected to migrate later today (bringing us down to 10 RC bugs for
    key packages tomorrow)
    - xorg, linux86, cross-toolchain-base-mipsen (waiting for linux,
      but linux seems like it will be ready to migrate tonight as well)

  * RC bugs for 7 key packages have been committed to git or had the
    maintainer confirm they would take care of it before the release.

  * Our safety net remains on 0 cases of misbuilds.

    (https://udd.debian.org/lintian-tag/wrong-file-owner-uid-or-gid)


# Remaining key package issues:


  * shim/shim-helpers-*/d-i: Waiting for the maintainer. All three
    expected to be resolved later and before the freeze.

  * gcc-X-cross-mipsen: The maintainer uploaded
    cross-toolchain-base-mipsen, so I have been trying to reach them
    to see if they are also planning a gcc-X-cross-mipsen as well
    (since that would in my view be better for all parties).

  * autopkgtests, sbuild, apt, krb5: Has patch committed to git

  * lintian: No signs of a patch yet nor a maintainer reaction.


# Remaining non-key package issues:

In total, 28 RC bugs are left for non-key packages. I am splitting them 
into two groups.

## Fixed in sid (19)

  * Fixes for 15 of those are expected to migrate later today.

  * 2 more are fixed in sid on day 1 (still pending build + autopkgtests
    results)

  * 1 more is fixed and waiting for python3-defaults to migrate.
    (I assume the transition will go through, so I am not worried for
     that one)

  * Finally, one more was fixed in sid but the maintainer upload had an
    unrelated regression. I filed #1092611 for the maintainer to resolve
    this.

Related UDD URL: 
https://udd.debian.org/bugs/?release=trixie&merged=ign&rttrixie-will-remove=ign&rttrixie-can-defer=ign&keypackages=only&fnewerval=30&flastmodval=7&fusertag=only&fusertagtag=rrr-no-as-default-issue&fusertaguser=niels%40thykier.net&allbugs=1&cpopcon=1&chints=1&ctags=1&cclaimed=1&cdeferred=1&caffected=1&crttags=1&cbritney=1&sortby=last_modified&sorto=asc&format=html#results

## Not fixed in sid (9)

  * One was fixed in sid and testing, but the bug was never closed.
    (#1089344). I decided not to retroactively fix the remaining numbers
    as I was afraid I would overlook some making the numbers in
    consistent (but it is fixed in sid + testing once the BTS updates).

  * Three packages had their NMU cancelled at the maintainers request and
    they are now waiting for the maintainer to resolve them or they will
    be auto-removed on the 21st of January. Whichever happens first.

  * Two packages had interaction between maintainer and me, where the
    maintainer was already planning an upload but got hit by key-expiry
    or wanted to postpone the upload (as they wanted to focus on
    something else first). I have offered to do a NMU for both in case
    it was helpful, but I am not

  * One had a pending auto-removal prior to the transition and another
    has a pending "Should this be removed from sid" suggestion. I have
    left those to their own devices.

  * One left with neither patch or maintainer reaction (avr-libc). I
    originally wrote it off as having a pending auto-removal from
    testing, but that seems to have been fixed without an upload. I might
    have a look at this one later.

Related UDD URL: 
https://udd.debian.org/bugs/?release=trixie&merged=ign&rttrixie-will-remove=ign&rttrixie-can-defer=ign&keypackages=ign&fnewerval=30&flastmodval=7&fusertag=only&fusertagtag=rrr-no-as-default-issue&fusertaguser=niels%40thykier.net&allbugs=1&cpopcon=1&chints=1&ctags=1&cclaimed=1&cdeferred=1&caffected=1&crttags=1&cautormtime=1&cbritney=1&sortby=last_modified&sorto=asc&format=html#results

# Final remarks

I believe only lintian and the gcc-X-cross-mipsen packages could be 
reason for a mild concern for closing the transition. Though, for 
gcc-X-cross-mipsen, the patches applied to gcc-X-cross should be doable 
to cherry-pick. For lintian, it can likely be resolved by disabling a 
test or two in its autopkgtests if push becomes shove.

Once those three plus apt, krb5, sbuild, and autopkgtests have been 
uploaded and migrated, I would consider this transition done. The shim + 
d-i cases are not actionable for me ("here by dragons" for NMU'ers) and 
the maintainers have confirmed they will handle them.

This statement is on the assumption that no more notable issues turn up 
(misbuilds, keypackages, or considerable numbers of non-key packages). 
Though, I am not worried on this front. The latest cases I am aware of 
were about 5-6 autopkgtests cases on the 5th of January (finite list), 
and then a test-rebuild identified a single case (krb5) on the 7th of 
January.

Best regards,
Niels


PS: The krb5 FTBFS was overlooked in my test-rebuilds because it (at the 
time) was affected by a tex-related build failure. Whatever that failure 
was, it was fixed outside krb5 between then and now. Unfortunate timing 
of my test-rebuilds. However, with unstable being unstable and at this 
scale, there would probably always have been some temporary regression 
in the way for some rebuilds.
OpenPGP_signature.asc (application/pgp-signature, 488 B)