community-hosted-pipeline tests should not ignore missing bundle

[clobrano]

If the submitted operator bundle has the wrong structure (e.g. the bundle folder inside operator-name/x.y.z/, like we did by mistake in #4199) the tests ignore the lack of bundle or catalog and terminates with success

[set-github-pr-title : set-github-status] + TITLE=
[set-github-pr-title : set-github-status] + echo 'Setting the PR title'
[set-github-pr-title : set-github-status] Setting the PR title
[set-github-pr-title : set-github-status] No bundles nor catalogs have been added/updated
[set-github-pr-title : set-github-status] + [[ '' != '' ]]
[set-github-pr-title : set-github-status] + [[ '' != '' ]]
[set-github-pr-title : set-github-status] + [[ '' != '' ]]
[set-github-pr-title : set-github-status] + echo 'No bundles nor catalogs have been added/updated'
[set-github-pr-title : set-github-status] + exit 0

full log: https://gist.github.com/rh-operator-bundle-bot/6e5f4598ee34371a84a11709b7aa0b58

I think not finding the bundle should result in an error instead.

[mporrato]

The behaviour of skipping the rest of the ci pipeline if there is no new bundle in the PR is by design. However the PR should not be automatically merged in this situation.

Thanks for reporting the issue: we will address discuss it.

Edit: Upon further thinking, automerge is still the right thing to do here: there are legitimate changes a maintainer can do outside of a bundle (for example updating the ci.yaml file or a readme file for the operator) that should be allowed to leverage the auto-merge feature.
The automatic merge for a situation like the PR you mentioned, even if it's not what the contributor intended, still does not prevent an additional fixup PR from being submitted and successfully processed.

[github-actions]

This issue is stale because it has been open for 30 days with no activity.