-
Notifications
You must be signed in to change notification settings - Fork 4.2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fix release performance tests #50699
Conversation
Yes, this changes makes a lot of sense to me (at least until the runner becomes just a runner and loses any knowledge of the metrics) |
@youknowriad the fix worked, but the tests still failed. This time there seems to be an actual test failure: From https://github.com/WordPress/gutenberg/actions/runs/5074051713/jobs/9113812136:
The above is failing only for the Let me know if it rings a bell. I'm going to investigate this in a follow-up PR. /cc @kevin940726 |
What seems weird is that the "list" block seems to be selected which is probably why the paragraph is not shown in the inserter. So the question that needs solving is why that list block is selected? Also, it seems that the list block is in the "center" of the canvas. So potentially we might have have a step that says "click canvas" or "click editor" or something like that, that is selecting the list block inadvertently. I think we do have a step to "click the canvas" to enter "edit mode" in the site editor (in branches where the browse mode is available) but that click shouldn't select any block, it should just switch to "edit mode". |
Seems oddly similar to #48208 (comment). |
Huh, forgot about this one - it looks exactly the same! 😬 I'll try debugging the current issue, but I was also about to start migrating perf specs to Playwright - might be worth checking first if Playwright does a better job OOTB in terms of stability here. |
Alright, it passed after a rerun. 🎉 I'm going to focus on Playwright migration, then! |
What?
The disparity between the
performance.js
runner source and the tests' source is causing the release perf tests to fail. The error is that the tests are saving result artifacts under a different name than the perf script is expecting them to be, so it cannot find them. For example, we can look at the most recent release perf job. The command is:...which values for that run are:
performance.js
script is the branch that the job is dispatched from, which is actually a tag reftags/v15.8.0-rc.1
.wp/6.2
, hence the disparity.How?
The fix is to make the tests source the same as the runner source (
GITHUB_SHA
).Testing Instructions
The next release perf tests should pass.