Skip to content

Update pytest and selenium deps#17033

Open
stevejalim wants to merge 6 commits intomainfrom
2025-02-12--dep-updates--pytest
Open

Update pytest and selenium deps#17033
stevejalim wants to merge 6 commits intomainfrom
2025-02-12--dep-updates--pytest

Conversation

@stevejalim
Copy link
Copy Markdown
Contributor

@stevejalim stevejalim commented Feb 12, 2026

This catches up with straggling version bumps that historically were painful around pytest-selenium, but I figured it was time to bite the bullet

And.... integration tests are... seemingly happy, but have some error annotations, oddly: https://github.com/mozilla/bedrock/actions/runs/21957292026 and https://github.com/mozilla/bedrock/actions/runs/21959245804 - are these false positives or tolerable negatives @maureenlholland @wen-2018 ?

@stevejalim stevejalim requested a review from a team as a code owner February 12, 2026 16:31
@stevejalim stevejalim changed the base branch from main to 2025-02-12--dep-updates February 12, 2026 16:32
@stevejalim stevejalim marked this pull request as draft February 12, 2026 16:45
Marks applied to fixtures have no effect and are now an error in
pytest 9 (PytestRemovedIn9Warning). All test functions using this
fixture already have their own @pytest.mark.cdn decorator.
Newer Selenium enforces W3C WebDriver spec compliance, rejecting
vendor-specific capabilities (avoidProxy, tags) at the top level.
These must be nested under sauce:options.
Base automatically changed from 2025-02-12--dep-updates to main February 12, 2026 18:11
@stevejalim stevejalim marked this pull request as ready for review February 13, 2026 10:37
@stevejalim stevejalim requested a review from a team as a code owner February 13, 2026 10:37
@wen-2018
Copy link
Copy Markdown
Collaborator

And.... integration tests are... seemingly happy, but have some error annotations, oddly:

For https://github.com/mozilla/bedrock/actions/runs/21957292026/job/63424850468, it looks like the submit button was clickable but timed out. It might be caused by https://github.com/mozmeao/springfield/blob/main/media/js/newsletter/form-utils.es6.js#L90. If we are using failure@example.com and the errorCallback() and the submit click can happen at the same time, causing the click hang. We can probably improve the test to be async

@stevejalim
Copy link
Copy Markdown
Contributor Author

And.... integration tests are... seemingly happy, but have some error annotations, oddly:

For https://github.com/mozilla/bedrock/actions/runs/21957292026/job/63424850468, it looks like the submit button was clickable but timed out. It might be caused by https://github.com/mozmeao/springfield/blob/main/media/js/newsletter/form-utils.es6.js#L90. If we are using failure@example.com and the errorCallback() and the submit click can happen at the same time, causing the click hang. We can probably improve the test to be async

Did you mean to link to the springfield repo there?

@wen-2018
Copy link
Copy Markdown
Collaborator

Did you mean to link to the springfield repo there?

Oops, I meant to link the same file in bedrock: https://github.com/mozilla/bedrock/blob/main/media/js/newsletter/form-utils.es6.js#L185

@stevejalim
Copy link
Copy Markdown
Contributor Author

stevejalim commented Feb 18, 2026 via email

@wen-2018
Copy link
Copy Markdown
Collaborator

I can look into this the second half of my day 👍

@stevejalim stevejalim requested a review from janbrasna March 16, 2026 11:52
@janbrasna
Copy link
Copy Markdown
Collaborator

I can't speak to the issues with the WNP/Flare newsletter validation — but understand the flakiness is annotated on the Playwright parts (appears to be loading issue, some pages were not returned, some buttons couldn't be clicked; in only one of four browser runs, I'll try to inspect the traces to see what the runner saw, If there's any clue in there), whereas this only touches the Selenium side of things.

I previously did successfully upgrade the deps springfield#361 without any impact (the reasons for downpinning seem to no longer exhibit), the only wrinkle was then when e.g. urllib/requests security bump landed, it clashed with the subdeps and that pytest+selenium stack needed to be reverted back to be able to update the vulnerable prod deps.

So I'm all for this, let's see if the newer major versions have better dep constraints or release at a faster cadence to cover the vulns as fast as the prod deps to match.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants