(shadowed) +1 maintenance report

Utkarsh Gupta utkarsh.gupta at canonical.com
Mon Jun 14 14:28:02 UTC 2021


This week I partly shadowed Christian on his +1 maintenance.

Whilst I don't have the core-dev powers yet, Christian & Gianfranco
sponsored uploads & did re-triggers on my request. So a huge thanks
to them!

As Christian mentioned, transitions were actively being driven and
things were already being looked at, so I decided to work on things
which didn't get enough attention, hoping that my Ruby experience
could help. :)

Here's what I did:

### schleuder ###

This has been lurking around since so long. The armhf failure
isn't reproducible and passes locally everytime, even with
different armhf setups. Passes in Debian, too. Since this has
become so hard to debug & blocks other packages, added a hint[1]
& has now migrated.

[1]: https://code.launchpad.net/~utkarsh/britney/+git/britney/+merge/404025

### ruby-httpclient ###

ruby-httpclient had some HTTP_PROXY issues (which were LP builders
specific) and some test_proxy_ssl which resulted in FTBFS randomly
(cf: debbug #861456). We hit some in the past and decided to disable
on the Debian side and not forward them because these work fine
locally & only hangs in specific environments, more so in
LP's builders.

Anyhow, uploaded the fixed package, ruby-httpclient/2.8.3-3,
to Debian which built fine on all architectures & the package

### ruby-gitlab-pg-query ###

This package didn't build on any architecture (ever) because of FTBFS
as it accesses the internet during the build (cf: debbug #981878).
This was initially packaged because src:gitlab needed it but
now gitlab has switched to using the original source, ruby-pg-query,
and stopped relying on its fork, therefore, this isn't needed by any
package anymore. Opened LP: #1931257, requesting its removal; thanks
to Iain Lane for processing that.

### python-django-debug-toolbar ###

Did a loop-rebuild on amd64 to ensure that it wasn't actually a
legit failure. Asked Christian to re-trigger
python-django-debug-toolbar/1:3.2.1-1 for amd64 and it passed &

### ruby-loofah ###

The package is blocked by the regression induced in
ruby-rails-html-santizier because of some API changes in ruby-loofah.
The tests weren't adjusted and were recently patched upstream.
Cherry-picking those patches makes it happy and so got it uploaded
to Debian to avoid "delta"; thanks Hideki for doing that.

Simultaneously, uploaded newer ruby-loofah, too, to Debian and both
of them are happy. ruby-rails-html-sanitizer already migrated and
ruby-loofah should soon do, too. Asked Gianfranco to re-trigger the
tests (w/ triggers) on all architectures and it passed, as expected.

### ruby-stackprof | ruby-ferret | ruby-hiredis ###

glibc had these three blockers^.
The first two needed a re-trigger on amd64 & the last one needed
a re-trigger on armhf. I did a loop-rebuild to ensure there are
no legit failures, turns out there weren't. So I requested Gianfranco
to do it for me & all of them passed & migrated.

### ruby-excon ###

Long-due failure on s390x. Not reproducible; passes locally & also in
a s390x canonistack VM. So opened an issue upstream, explaining the
same, and added a hint[2]. Hopefully it will migrate soon.

[2]: https://code.launchpad.net/~utkarsh/britney/+git/britney/+merge/404113

- u

More information about the ubuntu-devel mailing list