QA » openQA Project

Observation¶

In the investigation tab of failed jobs there is an entry "diff_to_last_good", e.g. in https://openqa.opensuse.org/tests/1930034#investigation which is empty but should show differences of vars.json of corresponding jobs.

Expected result¶

Commonly jobs should have some difference, at least the difference in build number

Problem¶

This seems like a regression, maybe due to https://github.com/os-autoinst/openQA/pull/3993 or https://github.com/os-autoinst/openQA/pull/4361 or https://github.com/os-autoinst/openQA/pull/4366 which changed related code areas.

Suggestions¶

As we have a test in https://github.com/os-autoinst/openQA/blob/afbb8ee357472fcc37c36c9d80159d67eca19115/t/10-jobs.t#L469 that ensures that the difference is still correctly calculated maybe this is a regression on UI side only. I suggest to try out locally with a command like NOT_HEADLESS=1 perl -d t/ui/01-list.t, step through the test until chromium starts and see if the investigation info for failed tests reproduces the problem there. Then, go back to an older git state and check there, then bisect and fix.