+1 maintenance

Steve Langasek steve.langasek at ubuntu.com
Mon Jan 31 21:29:46 UTC 2022

There were some gaps in +1 maintenance coverage for Foundations the week of
Jan 24, so I jumped in to do a bit more.

I spent some time going deeper on fixing NBS blockers from

Removal blocked by glewlwyd, which is fixed in -proposed and only blocked by
an update of gnutls28 which was blocked on an amd64 autopkgtest of systemd. 
There was a systemd autopkgtest running, but it had been going for 6h+ when
the median time for a successful autopkgtest run on amd64 is around 1h15m,
so I re-triggered.  My re-trigger failed but an auto-rerun succeeded, so
this unblocked - although per Ɓukasz's report there may have been another
rebuild required.

No-change rebuilds had already been triggered for reverse-dependencies, but
two of these reverse-dependencies had preexisting build failures (knxd,
purify).  The purify build failure was already filed as Debian bug #1001528
and the package is removed from Debian testing, so I've removed that package
from jammy.  The knxd build failure is not reproducible in Debian because
Debian is behind on glibc versions compared to Ubuntu, so I've patched this
package, uploaded and forwarded the fix to Debian as bug #1004310.

The gnuradio transition has thankfully moved along quite a bit since the
last time I looked at it, given that when the new gnuradio was uploaded to
Debian unstable probably half the reverse-dependencies failed to build. 
Remaining blockers for removal are gr-limesdr and sdrangelove.  gr-limesdr
is blocked on new python3-defaults making python3.10 the default; I looked
at the blocking autopkgtest failures for this, but also I removed gr-limesdr
from the release pocket in the meantime since it has no reverse-dependencies
and the fixed version is certain to get back in for release.  And
sdrangelove FTBFS, as reported in Debian bug #1002166, so also removed.

Took a peek at the first two blocking autopkgtests, both of which required
sourceful changes.  It doesn't appear there has been any mass bugfiling yet
in Debian for this transition, and Graham has been on top of retries for any
failures that are shallow, so I moved on to lower-hanging fruit.

One revdep, biboumi, which has been blocked in -proposed on a ppc64el build
failure of botan for close to 2 months.  Since no one else has made any
headway on it, I dove in and after struggling to get a working ppc64el VM on
hardware that supports jammy (POWER9), put some work into isolating the
particular tests that were failing.  This is all captured in
https://bugs.launchpad.net/ubuntu/+source/botan/+bug/1959021; quite still
some ways away from actually resolving the problem, as the botan testsuite
is fairly opaque.  Since the tests started failing after a no-change rebuild
of botan against openssl 3, it's possible this is a regression in openssl
itself rather than in botan, and I've also confirmed the failures are still
there when running against a build of openssl from trunk.  This will take
some more time and effort to get to the bottom of.

Steve Langasek                   Give me a lever long enough and a Free OS
Debian Developer                   to set it on, and I can move the world.
Ubuntu Developer                                   https://www.debian.org/
slangasek at ubuntu.com                                     vorlon at debian.org
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <https://lists.ubuntu.com/archives/ubuntu-devel/attachments/20220131/f657580a/attachment.sig>

More information about the ubuntu-devel mailing list