Monday 17 June 2024

Re: +1 Maintenance Report

(Sorry, I sent incomplete)

Hello everyone,

Sorry for the delay...  I was waiting for the resolution (reviews/acceptance) of some of the tasks I did
(learning also .
As Athos commented, I shadowed him on the last days of his shift rotation (thanks, Athos! It was a pleasure).

These are the fixes I made:

* python-chemspipy: A FTBFS: when tests were running, six module wasn't found
 (provided by python3-six package, not listed as (b)dependency in d/control). But the
 package was a sync from Debian and there it was built OK [1]. Going deeper, python3-six
 was a implicit dependency of python3-request via its direct dependency python3-urllib3. 
 We have a different version that the one used in sid/trixie:
     Ubuntu: 2.0.7-1 [2], synced from experimental
     Debian (sid/trixie): 1.26.18-2 [3]
  six module was embedded into python3-urllib3 until this last version (see changelog entry).

* geophar: FTBFS. simpy is embedded in upstream source code, but that is not the case in Debian
  (so the same is true for Ubuntu). Since the last version we updated, there have been new tests covering
  the existence of these embedded simpy files: that test doesn't make sense in Debian, nor in Ubuntu. 
  I adapted the check to avoid this test.  Also, assertEquals is deprecated [4], so I changed it by refreshing
  the existing correspondent patch. I created an MR in salsa for this [5]. I didn't forward it to upstream, because
  in the Debian tracker, the patches are marked as "need to be submitted to upstream" and I left it to the maintainer
  as the original author of the patch.

* python-phpserialize: FTBFS: since python3.12, the assert_ method no longer exists [6]. We could see a deprecation
  warning for this in previous builds, like this one for Focal [7]: 
       test_object_hook (tests.PhpSerializeTestCase) ... /<<PKGBUILDDIR>>/tests.py:92: DeprecationWarning: Please use assertTrue instead.
         self.assert_(b'WP_User' in x)
  In Debian, they still use Python3.11 (so they have the warning, not the error like us) [8]. I forwarded the patch I did to Debian [9], 
  but in the meantime, I was changing assertTrue by assertIn as pointed out by Benjamin Drung, and Sergio Durigan was faster :). So,
  it's synced to Oracular already.

python-awkward: I spent sometime with this, but finally Athos made it.

I want to thank Athos and Paride for the sponsorship and Benjamin Drung and Daniel Drapper for their reviews too.

Now, yes, the report is complete (I blame the keyboard shortcut that triggered the send). Thanks everyone for reading!



On Mon, Jun 17, 2024 at 10:04 AM Miriam Espana Acebal <miriam.espana@canonical.com> wrote:
Hello everyone,

Sorry for the delay...  I was waiting for the resolution (reviews/acceptance) of some of the tasks I did.
As Athos commented, I shadowed him on the last days of his shift rotation (thanks, Athos!).

These are the fixes I made:

* python-chemspipy: A FTBFS: when tests were running, six module wasn't found
 (provided by python3-six package, not listed as (b)dependency in d/control). But the
 package was a sync from Debian and there it was built OK [1]. Going deeper, python3-six
 was a implicit dependency of python3-request via its direct dependency python3-urllib3. 
 We have a different version that the one used in sid/trixie:
     Ubuntu: 2.0.7-1 [2], synced from experimental
     Debian (sid/trixie): 1.26.18-2 [3]
  six module was embedded into python3-urllib3 until this last version 




On Tue, Jun 11, 2024 at 12:32 AM Brian Murray <brian@ubuntu.com> wrote:
On Fri, Jun 07, 2024 at 06:03:35PM -0300, Athos Ribeiro wrote:
> I did +1 maintenance from 2024-06-03 to 2024-06-07.
>
> I start the week by running the ubuntu-archive-tools find-proposed-cluster
> script. There was nothing relevant there at this point of the cycle.
>
> Then I started looking at individual packages, no hard rules, but I was
> trying to focus on the bottom half of the list. Coincidently, the first
> 2 packages I looked at had infrastructure related failures. I then
> proceeded to run the archive tools retry script:
>
> ./retry-autopkgtest-regressions --log-regex='unexpected eof from the testbed'
>
> I sticked to this regex for the 5 days I was working on +1 maintainance and
> found dozens of (non-duplicated) occurrences each day.

I expect that the majority of these were on arm64 and a result of
virtual machines taking an extraordinarily long period of time to be
provisioned in PS6. This has been reported in RT 164415 and is
documented at our autopkgtest service page[1].

[1] https://discourse.ubuntu.com/t/autopkgtest-service/34490

--
Brian Murray

--
ubuntu-devel mailing list
ubuntu-devel@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel


--
Canonical-20th-anniversary

Miriam España Acebal

Software Engineer II - Ubuntu Public Cloud/Server

Email:

miriam.espana@canonical.com

Location:

Spain  (GMT+2)


canonical.com

ubuntu.com




--
Canonical-20th-anniversary

Miriam España Acebal

Software Engineer II - Ubuntu Public Cloud/Server

Email:

miriam.espana@canonical.com

Location:

Spain  (GMT+2)


canonical.com

ubuntu.com