-
-
Notifications
You must be signed in to change notification settings - Fork 717
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
Problems with CI #132
Comments
There's probably some flakiness in the tests. I've seen it more often in the browser tests, but could be in the unit tests as well. We might want to change the github action config so that the tests retry once in case of a failure? |
I have mixed feelings about rerunning the test automatically. The problem is that the action cannot distinguish between a flaky test and one which fails because it has to fail. I suggest we trigger tests manually if we want to rerun them. Like this we can spot problems with the tests faster. Maybe we can keep this issue to keep track flaky test runs. |
Here is another false negative:
This is in CI -> macos-latest |
And in https://github.com/maplibre/maplibre-gl-js/runs/2339962584?check_suite_focus=true macos-latest:
and test-browser:
|
And in CI -> test-browser -> test-chrome
https://github.com/maplibre/maplibre-gl-js/pull/137/checks?check_run_id=2364438111 |
https://github.com/maplibre/maplibre-gl-js/pull/140/checks?check_run_id=2380509547
ci -> macos-latest -> unit-test |
https://github.com/maplibre/maplibre-gl-js/pull/140/checks?check_run_id=2380738229
ci -> windows-latest -> yarn install |
When I run the tests, this is the only one error message I get, while running the tests.
This are my versions:
|
I've seen this test fail too, seems like this specific zoom test is flaky...? |
Sorry, my bad, these are different tests, but related to zoom... I don't remember exactly which one failed in my case, but it was related to zoom... |
@HarelM No, it has nothing to do with your code. The error occurs before. |
After a merge to
|
|
I think we should flag the tests that we see that sometimes fail and and some retry mechanism or something. Timing can changed based on machine load, so these tests should be more resilient... |
Here's another failure, no code changes were actually made (only changes in typing). the following is the test that failed:
Also this test:
|
Problem with the release workflow: https://github.com/maplibre/maplibre-gl-js/runs/3514866759?check_suite_focus=true
|
The problem is with the ' around the file name in lint-css when running on Windows. |
Looks like that was added by @petr-pokorny-1 in #95 "First jobs perform tests. It runs on Windows runner since Linux runners don't have GPU, which is required by many tests. Second job runs on Linux and creates builds minified library, creates github release, uploads release assets, updates version in npm package and submits package in the npm registry." |
Thanks for the explanation @msbarry. I think it is a real problem that we have different CI strategies for |
Definitely agree @wipfli - I set up the ci and analyze workflows not the release one. I seem to remember it was tricky to factor out common reused steps. Another option might be to chain them, so you can trigger a release workflow off of a successful build? |
I opened another issue on it - to split the steps to different yml files, after that you can define the triggers for each yml file I guess, maybe... |
This issue is stale because it has been open 60 days with no activity. Remove stale label or comment or this will be closed in 7 days. |
This issue was closed because it has been stalled for 30 days with no activity. |
Poor GitHub seems to have a hard time will all our CI scripts. |
I'm no longer experiencing this issues. |
I submitted pull request #131 and in the first test run
unit-test
of the actionCI
failed withSo this was https://github.com/maplibre/maplibre-gl-js/actions/runs/738748443.
@nyurik then asked me to add a line to the changelog, which I did and which then triggered another run of the tests which went through without any error. This second run which worked was https://github.com/maplibre/maplibre-gl-js/actions/runs/738794426.
Any idea why the test might be unreliable?
The text was updated successfully, but these errors were encountered: