Triage/Migrations report (2020-01-16)

Bryce Harrington bryce.harrington at
Thu Jan 16 22:54:35 UTC 2020

On Thu, Jan 16, 2020 at 02:49:54PM +0100, Paride Legovini wrote:
> # Triage
> 13 bugs to triage, worth mentioning:
> ##
> Package: apache2
> Title: Apache DBD Auth not working with mysql Focal 20.04
> Looks like a valid issue, affecting Eoan too. I added steps to
> reproduce. Marked Triaged, importance: High, and subscribed the team.
> ##
> Package: nut
> This one was Incomplete and got expired by the Launchpad Janitor,
> however the bug is *almost* Fix Released as a version fixing the bug is
> currently in focal-proposed (in proposed for 2 days). I set it to
> Triaged and subscribed the team; to be set to Fix Released once nut
> migrates.
> # Migrations
> python-cffi:
>  - missing dep: libffi
> python-tempita: candidate

This depends on python3-nose, which still uses py2 for generating its
code docs.  Nose is blocking other python3 packages from migrating as
well.  I've been working on a fix for nose today, that I think should
unclog things, and hopefully let tempita, blinker, and maybe other
packages complete transition.

> openssh:
>  - known issue (cert date in the future not in the future, LP #1859013).
> python-boto:
>  - regression on patroni
>  - I think Bryce triggered a rebuild already
>  - Test failure unlikely to be really related to python-boto
>  - probable badtest?

Right, boto is blocked on patroni, which appears to have a python2-ism
in one of its test cases, as shown in the build log:

dcs = 'localhost:8500'
        parsed = urlparse(dcs)
        scheme = parsed.scheme
        elif scheme not in DCS_DEFAULTS:
            raise PatroniCtlException('Unknown dcs scheme: {}'.format(scheme))

The problem is that the urlparse module is renamed to urllib.parse in Python 3.

I think that if the test_parse_dcs() routine is updated to use
urllib.parse instead, this test case may pass.

It looks like that's the only test case that failed, so hopefully that'd
be the only change.  (Running 2to3 over the codebase can be helpful in
spotting other issues with incomplete py2 conversions; sounds like it
would have caught this case.)

Btw, one other thing worth checking into... it looks like python-boto is
getting sync'd from debian whereas previously we'd been keeping some
delta for the autopkgtest.  I didn't find a sync request bug report for
python-boto, so it might be worth doublechecking that the delta is
indeed droppable.


> re2c:
>  - regression on i386
>  - Bryce triggered a rebuild twice
>  - filed update-excuse bug LP: #1859980
> pacemaker: candidate
> net-snmp: candidate
> haproxy:
>  - missing builds
>  - LP: #1858485
> ocfs2-tools:
>  - regression on s390x
>  - related: LP: #1745155
> Paride
> -- 
> ubuntu-server mailing list
> ubuntu-server at
> More info:

More information about the ubuntu-server mailing list