You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Investigate workflow job failing on main: test / test (job 3) #71350
🐛 Error Message:
failure: Process completed with exit code 1.
warning: Cache not found for keys: Linux-node-modules-d404f80f8707543f66c27d64638a51d61ac2d639c73bef19e481a3aa2b3de6ce
⚠️Action Required⚠️:
🛠️ A recent merge appears to have caused a failure in the job named test / test (job 3).
This issue has been automatically created and labeled with Workflow Failure for investigation.
This doesn't really look like a coincidence to me, and I wouldn't be too too surprised to learn that #71047 caused test failures. I'm going to open a revert and see if tests pass
@roryabraham Okay, I think I’m starting to understand the core of the issue. First of all, in the old PR, only the first commit triggered CI, which is why I thought everything was fine. I created a new PR, and now I can reproduce the issue, but only on CI.
The root cause seems to be the bump in the Babel target of the compiler from 18 to 19, which resulted in longer tests. In most tests, the increase in time was small, but for this particular test, it was significantly larger for some reason. I will continue investigating why changing the compiler target caused the tests to take longer.
However, I still think this change should remain, as it appears to have been an oversight that it wasn’t updated in the first place.
Here is the new draft just to confirm my theory: #71364
Activity
roryabraham commentedon Sep 26, 2025
This doesn't really look like a coincidence to me, and I wouldn't be too too surprised to learn that #71047 caused test failures. I'm going to open a revert and see if tests pass
kstroz commentedon Sep 26, 2025
@roryabraham Okay, I think I’m starting to understand the core of the issue. First of all, in the old PR, only the first commit triggered CI, which is why I thought everything was fine. I created a new PR, and now I can reproduce the issue, but only on CI.
The root cause seems to be the bump in the Babel target of the compiler from 18 to 19, which resulted in longer tests. In most tests, the increase in time was small, but for this particular test, it was significantly larger for some reason. I will continue investigating why changing the compiler target caused the tests to take longer.
However, I still think this change should remain, as it appears to have been an oversight that it wasn’t updated in the first place.
Here is the new draft just to confirm my theory:
#71364
GH Action with version 18: https://github.com/Expensify/App/actions/runs/18031830367/job/51309790846
GH Action with version 19: https://github.com/Expensify/App/actions/runs/18032094310/job/51310632930
Small update:
roryabraham commentedon Sep 26, 2025
We can continue with follow-up in #71215