Django Fellow Report - Natalia

Week ending 2023-09-10 (Week 36)

Triaged

Reviewed

Authored

Other/Misc

  • Security reports review and analysis.
  • DjangoCon talk continued.

Week ending 2023-09-17 (Week 37)

Triaged

Reviewed

Authored

Other/Misc

  • Preparation for Django 5.0 feature freeze and alpha release.
  • Security reports review and analysis.
  • Removed spam comments from a few tickets.
2 Likes

Week ending 2023-09-24 (Week 38)

Triaged

Reviewed

Authored

Other/Misc

1 Like

Week ending 2023-10-01 (Week 39)

Triaged

Reviewed

Authored

Other/Misc

  • Finalized security patches for all relevant Django versions.
  • Sent prenotification and posted public announcement about the incoming security release planned for Oct 4th.
  • Recorded the first practice session of my DjangoCon US presentation. I need to make it shorter or talk faster!
3 Likes

Week ending 2023-10-08 (Week 40)

Triaged

Reviewed

Authored

Other/Misc

2 Likes

Week ending 2023-10-15 (Week 41)

Triaged

Reviewed

Authored

Other/Misc

2 Likes

Week ending 2023-10-22 (Week 42)

DjangoCon US conference week!

Conference report: I arrived at noon on Sunday, very tired after 30 hours of traveling. I had a few scares with flights and connections, but in the end, it all worked out. I met A LOT of wonderful people while in Durham, both existing “digital” acquaintances and people I had never met before. Everyone was so, so friendly and welcoming, and also very supportive.

During the week, I barely opened my laptop, so the bulk of my work involved meeting the community, engaging in very interesting conversations about Django—its strengths, weaknesses, and ways to help and make it better. We also discussed inclusion, and it was quite amazing that I had to wait in line for the ladies’ bathroom (may sound trivial, but I consider this an impactful conference achievement! I’ve never had to wait in the bathroom at a tech conference before).

I was astonished by the diversity of both the speakers and attendees, not just in terms of gender but also with the presence of many other minority groups. I enjoyed and learned from all the talks, but the Djangonauts, in particular, made me feel fortunate to be a part of this community. I also engaged in other interesting but less relevant conversations, such as discussing how to prepare mate and my life before becoming a fellow.

During the sprint, I split my time between onboarding and assisting new contributors and discussing Django documentation with Daniele Procida. We reviewed several contributing-related doc pages that I had personally found challenging while preparing for my role. Daniele provided a dozen excellent suggestions that we’ll work to incorporate soon. I’d like to express my gratitude to Lance (@lancegoyke) and Velda (@VeldaKiara) for stepping up and offering to assist in this effort.

In summary, a heartfelt thank you to each and every one of you! I feel privileged to have had the opportunity to attend and share my experiences with all of you. Thanks to the organizers, volunteers, those who engaged in discussions, those who provided support before, during, and after my presentation, and all the attendees who were friendly and respectful of our code of conduct.

6 Likes

@nessita I’ve heard numerous positive reports on your talk. It really hit home! :clap:

For Triaged tickets, could I request an outcome in brackets at the end of the line? e.g. ticket-34896 … (wontfix)

That would be helpful :pray:

Yes, absolutely! In fact I usually have those, I’ll debug what happened with the last batch. Thanks for pointing this out!

It was great to get to meet you at DjangoCon!

I also engaged in other interesting but less relevant conversations, such as discussing how to prepare mate and my life before becoming a fellow.

I wish I’d been around for those conversations, they seem pretty essential to me :smiling_face:

3 Likes

Week ending 2023-10-29 (Week 43)

I started the week feeling quite exhausted after DjangoCon US and the trip back home. I thought I could ease into the week, when Mariusz luckily reminded me (thank you!!!) that “today” (that was last Monday) was Beta Release Day. After a brief moment of panic and some mate, I reviewed and merged the pending release blocker, allowing for the subsequent 5.0 Beta Release, so it all worked out in the end.

To avoid reliving such a shocking episode, I’ve created a dedicated Google Calendar for Django Releases. The main purpose of this calendar is to track both past and upcoming Django releases, as well as send (me) timely reminders to allow sufficient preparation time for release tasks. If you believe this calendar could be valuable for you personally or for the broader community, let me know!

Triaged

Reviewed

Other/Misc

2 Likes

Week ending 2023-11-05 (Week 44)

Triaged

Reviewed

Authored

Other/Misc

4 Likes

Week ending 2023-11-12 (Week 45)

Triaged

Reviewed

Authored

  • https://github.com/django/django/pull/17462 - Fixed #34958 – Fixed isolation of messages_tests.tests.TestLevelTagstest_lazy.
    • This one was “fun” to do in that I found the experience rewarding once I solved it, but until then I spent a lot of time debugging what the heck was going on, with little success. While I did have the “shuffle seed” to reproduce the error when running the whole suite (YEY), it took me a lot of printing and runs to reduce the scope of the test run to a subset. Is there a procedure to methodically find this reduced set of troubling tests?
      (I tried --pair (probably misused) but I kept having No problem pair found :persevere:)

Other/Misc

  • Follow up on ticket-34830 made me go deeper in the report investigation leading to many TILs around our default (error) views and the template handling. Thanks Tim!

detect-test-pollution uses a bisect approach to find a pair of polluting tests. See its video tutorial for an explanation. It works with pytest but could be adapted to work with unittest/django’s test runner.

I’ve not used --pair, that’s interesting! O(n) though… :fearful:

1 Like

Ah no, I misunderstood --pair from its docstring. Just found the docs. It works like detect-test-pollution.

Maybe it’s not properly compatible with --shuffle? It seems like a good candidate for promotion to an actual Django feature rather than living only in runtests.py

1 Like

@adamchainz Thanks for the pointers!

What do you mean with promoting pair to an actual Django feature?

I suspect Adam means supporting it for manage.py test too.

Yes, and the bisect option!

Adding these options to manage.py test makes sense! I agree they are super helpful. Shall we create a separated topic to propose the (extension?) feature and gain consensus?