+1 Maintenance report

[Christian Ehrhardt]

Hello,
last week I was on +1 Duty, well I was supposed to :-/
Sadly many things tried to keep me away from it. After checking with
Mclemenceau for that to be "ok", I still tried to get as much done as I could
fit in between. This might only be worth maybe ~1.25day overall, but at least
I was able to tackle a few things.

## Segyio [1]

Vorlon left this bug from last weeks +1, but while starting to look at it
I found Ginggs already had tackled that. I helped to ensure upstream is aware
(for bonus confusion it is an embedded source in sigyio from catch2) and the bug
status [1] is up to date to reflect all that.

## parlatype [2]

FTBFS due to whitespace damage, reported to Debian [3] and filed as LP
update-excuse bug [2]. Uploaded a fix to get it resolved for now and expecting
it will become a sync later on.

This fix migrated in Ubuntu with the fix and got merged into Debian.
On the weekend the fix was uploaded to Debian including that and I synced
the package again.

## nbd [4]

I realized too late how young the age of the last upload was and had created
and submitted two fixes already. This is pretty much WIP in Debian and
uploaded to unstable every time. We can wait until it resolves there, and LP
but tracking and explaining that was filed.

As expected over the next few days it was resolved.
Never look at too young uploads I guess.

## ldc [5]

This was an FTBFS for almost two weeks and the reason is that the LDC project
did not yet convert from the ORCv1 to the ORCv2 API for the jit compiler.
Thereby it can't build with llvm >=12 which is our default. Build this against
llvm-11 until upstream has resolved that (known issue for them).

For Ubuntu I uploaded a fix to build with llvm-11 for not, but to really
resolve this mid term I filed a bug with Debian outlining the backgrounds
(considered experimental and to be dropped by upstream) and proposed a PR there
to do so.

Builds look good now, but it still hangs in the new queue of jammy.

## emscripten

I looked into the TFBFS of this but it seemed flaky and failing for different
reasons every time. Sometimes taking 10h to do so made it hard to debug and
then I found that Debian still uploads more versions to fix similar (but not
the same) test issues on their side.
I think we need to let Debian settle this and once concluded there but still
failing for us have another look.

## glew [7]

I've found that glew is listed as blocking dependency 45 times.
As usual the chain is long, there are:
- vtk7 also wait on gdal, but that is only listed as it was recnetly build
  and waits to be published
- vtk7 which has a test issue on armhf
- rss-glx which FTBFS

I've found rss-glx to be fixed in Debian [6] filed a LP tracker about it [7]
rebased and uploaded a merge of that new version to resolve this issue.
This worked fine and is at least one less to block glew.

## pg/post* [8]

I have found that some universe packages around the wider postgres context
have issues and thought I might be able to help those to get out of the way
of other transitions which implicitly depend on them e.g. proj, pdal,
pgrouting and indirectly even openssl.

I did some debugging on a ppc64 host, but no matter which combination
the issue was not reproducible there. So I eventually filed [8] to document
the summary so far and brought it to the attention of the postgresql packagers
if they recognize anything.


----

[1]: https://bugs.launchpad.net/ubuntu/+source/segyio/+bug/1951658
[2]: https://bugs.launchpad.net/debian/+source/parlatype/+bug/1951833
[3]: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1000388
[4]: https://bugs.launchpad.net/ubuntu/+source/nbd/+bug/1951839
[5]: https://bugs.launchpad.net/ubuntu/+source/ldc/+bug/1951845
[6]: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=984322
[7]: https://bugs.launchpad.net/debian/+source/rss-glx/+bug/1951968
[8]: https://bugs.launchpad.net/ubuntu/+source/postgis/+bug/1952604

P.S. as usual I'm not listing the many small cases of retriggering with or
without changed package sets.


-- 
Christian Ehrhardt
Staff Engineer, Ubuntu Server
Canonical Ltd