QA » openQA Project

The dates point to https://github.com/os-autoinst/scripts/pull/335 as the culprit, but hard to tell what the problem is, as line 68 is really just a function header

It's reproducible with

./openqa-label-known-issues https://openqa.opensuse.org/tests/4425222

tinita wrote in #note-7:

It's reproducible with

./openqa-label-known-issues https://openqa.opensuse.org/tests/4425222

Calling the script with the old code still fails but silently

❯ dry_run=1 ./openqa-label-known-issues https://openqa.opensuse.org/tests/4425222
Requesting jobs/4425222 via openqa-cli
    ~/Documents/Work/qatools/repos/scripts on    add_requirements_to_run_script *1 ?2                                                                                                                                                                      
❯ echo $?
127

What is "special" with https://openqa.opensuse.org/tests/4425222 is that the CASEDIR uses a absolute tree branch which I think the openQA doesnt support, as it uses the #mybranch shortcut

In the existing (previous) code we have elif label_on_issues_from_issue_tracker "$id"; then
And the function label_on_issues_from_issue_tracker is expected to fail currently, because it just calls label-on-issue which passes if it wrote a comment, but fails if it didn't, just that the failure doesn't indicate something fatal. In which case it would try the next elif branch.
And because of the elif ... then the error is catched.

But the newly added line above calls label_on_issues_from_issue_tracker without an if/elif or || something , so that's why it's aborting the script.

I guesss if label_on_issues_from_issue_tracker in the new code fails, we still want to go through the rest of the code, e.g. call label_on_issues_without_tickets or handle_unreviewed.
So the code needs to be rearranged a bit.

As tina had mentioned in my last related PR and I mentioned https://github.com/os-autoinst/scripts/pull/342 the function needs to exit.

Btw. as this ticket is not estimated and has no acceptance criteria; can I misuse that for demanding more test coverage? :)

ybonatakis wrote in #note-10:

As tina had mentioned in my last related PR and I mentioned https://github.com/os-autoinst/scripts/pull/342 the function needs to exit.

Alternative approach https://github.com/os-autoinst/scripts/pull/343

jbaier_cz wrote in #note-14:

Btw. as this ticket is not estimated and has no acceptance criteria; can I misuse that for demanding more test coverage? :)

As discussed in the unblock we agree that it'd be best to have tests first and then see which approach works and Yannis is already looking into that.

Tests added and CI passes. Waiting for a new round of feedback

ybonatakis wrote in #note-17:

Tests added and CI passes. Waiting for a new round of feedback

https://github.com/os-autoinst/scripts/pull/342#issuecomment-2325027337

I'm trying to write down the logic I think we want:

investigate_issue¶

We write autoinst-log.txt and reason into the same file.
If we successfully got autoinst-log.txt (http 200 or 301), we continue with trying to label the test. DONE.

No autoinst-log.txt¶

• If we couldn't fetch autoinst-log.txt, check if there is a general issue. handle_unreachable performs various tests and should return non-zero to indicate that we shouldn't go on trying to label the test.
• If the http status was not 404, don't continue with labeling.
• If the job is too old, don't continue with labelling.
• If there is no reason as well, don't continue with labeling.

Only if the http status is 404, the job is not too old and the reason is set, continue with labelling.

As for now tests in CI fail but they work local. tests touch a part which it wasnt covered before. will try to fix

https://github.com/os-autoinst/scripts/actions/runs/10724123810/job/29739041706?pr=342 openqa-label-known-issues: line 63: hxselect: command not found is not from the test

https://github.com/os-autoinst/scripts/pull/342 under final review. I also went through the various unresolved comments to highlight open questions and resolve all addressed suggestions.

CI passes

https://progress.opensuse.org/issues/166772 for concern from the discussion https://github.com/os-autoinst/scripts/pull/342/files#r1745228802