For the new Ubuntu Project Docs, I contributed a detailed write up on
what all the jargon means on the update-excuses page, and step by step
instructions on how to solve migration issues. I hope everyone working
on +1 migration reads it, finds it useful, and adds more to it over
time.
--> https://github.com/ubuntu/ubuntu-project-docs/pull/121
(Will follow up on review feedback when I return from PTO.)
### PS7 Proxy issue (LP: #2116119) ###
Andreas handed off to me awareness of this proxy bug that has been
affecting some packages. I read up on and summarized the discussion in
bug LP: #2116119, and identified a couple possibly affected packages.
Sounds like this has made some progress this week, but would be
worthwhile for future +1 maintainers to give themselves awareness of.
### Syncs ###
* node-nan (LP: #1967589)
- Upstream disabled the problematic test, so we can drop our
workaround
- sync'd 2.22.2-1
* unhide (LP: #2062269)
- Robie forwarded a workaround patch upstream last year.
- Upstream has implemented proper fixes, and put out a new release.
- Verified both fixes present in unhide version 20240510-2
- sync'd 20240510-2
### Merges ###
* sbd: Merge
- https://launchpad.net/~bryce/+archive/ubuntu/sbd-merge-v1.5.2-5/+packages
- https://code.launchpad.net/~bryce/ubuntu/+source/sbd/+git/sbd/+merge/489556
- Reviewed & approved by Lena
--> Uploaded to -proposed
### containerd-app ###
Filed LP: #2118738 with a possible fix for a containerd-app autopkgtest
issue. Essentially, the DEP8 test is firing up a server in the
background and then immediately starting the client, with the result
that sometimes the client is too quick and starts executing before the
server's port is ready for connections.
I've proposed a rough patch with some wait logic to prevent this
scenario, for Athos to look at when he returns from vacation.
As an aside, this client-server arrangement is not at all uncommon for
autopkgtests, and seems like it'd make sense to have some stock helper
code in autopkgtest to standardize the handling of such situations,
rather than having to reinvent the wheel each time?
### Retriggers ###
Reviewing noble's update-excuses, several packages had 'timeout failure'
and 'only unknown' issues, that typically just require retriggering. I
did so and they did indeed pass:
gimp-plugin-registry/s390x
fwts/amd64
glibc/amd64
lxc/ppc64el
I also did retriggers for all the following packages, and they're no
longer listed with any failures on update-excuses:
borgbackup
caddy
charliecloud
crowdsec
dbconfig-common
dotnet8
dpdk
fence-agents
gjs
gvfs
hamlib
imagemagick
lighttpd
llvm-toolchain-19
macaulay2
munin
openjdk-21-crac
pacemaker
sbuild
util-linux
vlc
yosys
### Existing update-excuse bugs ###
I worked my way up from the bottom of update-excuses looking at
update-excuse bugs, looking for obvious next actions. Many were out of
date, or still pending upstream attention. So, didn't find anything
easily actionable, but made notes where I could:
crun
- https://bugs.launchpad.net/ubuntu/+source/crun/+bug/2115813
- Bdrung already pulled the fix for this bug, but the migration
blockage is something different - a test case that fails on arm64.
- I found a possibly match in a closed upstream bug; they decided to
skip this specific test on this arch; perhaps we should too?
- I also added possible way to confirm it's that same test case.
ansible-core
- https://bugs.launchpad.net/ubuntu/+source/ansible-core/+bug/2116825
but this doesn't match current dep8 test log errors; I think this
might have been due to an apparmor constraint dir that affected
writing to /home/ubuntu by tests.
- The current dep8 test failure looks like possibly another instance
of the PS7 proxy issue (LP: #2116119), as ansible seems to want to
use pip to install things for its test.
- Upstream has an open PR for a workaround that might be worth
examining to carry as Ubuntu delta until it lands upstream?
- I've updated both bugs with these notes
brial
- https://bugs.launchpad.net/ubuntu/+source/sagemath/+bug/2063058
- Dependency on python-setuptools
- Newer version upstream but doesn't look like it corrects this...?
- Perhaps this version should be dropped from the archive until Debian
packages a newer release.
sks
- https://bugs.launchpad.net/ubuntu/+source/sks/+bug/1886992
- Build failure was fixed in debian
detox
- https://bugs.launchpad.net/ubuntu/+source/detox/+bug/2079767
- Fixed in debian
### Git-Ubuntu Snaps for non-amd64 architectures ###
Not really a +1 maintenance item, but this week Andreas and I also
sorted out an issue where git-ubuntu's snaps were not getting published
to latest/edge for architectures other than amd64. This turned out to
just be a credentials re-authorization that had expired. I
re-authorized it, and it's now producing snap builds for all
architectures to latest/edge:
https://snapcraft.io/git-ubuntu/releases
It feels like it's time to do a new formal release, so am thinking about
cutting a 1.2-rc1 release candidate next month.
### Tooling: Autopkgtest log parser ###
In between everything else, I started roughing together a simple parser
for DEP8 autopkgtest log files. I didn't too far (it's still mostly
TODOs) but I'm happy with the scaffolding. It's able to break the logs
up into their logical sections, and dump it as JSON.
https://git.launchpad.net/~ubuntu-server/+git/ubuntu-helpers/tree/bryce/parse-autopkgtest-log
Not sure if/when I'll get another chance to work on it, but making
mention of it in case anyone else thinks it of interest.
Bryce
--
ubuntu-devel mailing list
ubuntu-devel@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel
