Pretty please?

@mvadkert has recently created a pull request to use the latest message format:

• https://github.com/CentOS-PaaS-SIG/upstream-fedora-pipeline/pull/123

This should include running and queued messages as well. So when this is merged we should be able to add links to pending jobs.

Not the message format, just the topics name really. The content is still incorrect. I will propose a patch this week to get the content right, after that we need patches to pagure to fix that.

Trying with a simple approach how to fix this:

https://pagure.io/Fedora-Infra/loopabull-tasks/pull-request/1

@mvadkert, if I undestand it correctly, your proposed patch above should probably resolve the remaining part of #1 as well, right? There will be just a single flag displayed for the Fedora CI results so including the commit hash will not be necessary anymore.

@psss it is true, but maybe I would opt in to also display the ref that verifies it is true that the result is for current head. Without the ref, we just hope it is true :)

OK, makes sense to be a bit more specific.

still waiting for review from maintainers :/

@churchyard seems pending now works :) can you try pls?

I am closing this :) Reopen if needed

FTR https://pagure.io/Fedora-Infra/loopabull-tasks/pull-request/1

was merged

I see pending links! Thanks.

They are not updated once finished:

https://src.fedoraproject.org/rpms/pipenv/pull-request/13

@churchyard I know :( I politely asked @pingou to take a look what could be wrong :( I lack access to debug :(

Thanks for reopening!

Also, no pending links any more: https://src.fedoraproject.org/rpms/python-setuptools/pull-request/11

Pending link appeared after all.

found a typo in the code, retesting on a PR

It works! Thank You.

The pending link shows no more. The failure link is shown after the job fail (and I assume a success one works the same).

Confirmed. Here's a recent example:

• https://src.fedoraproject.org/rpms/did/pull-request/5
• https://jenkins-continuous-infra.apps.ci.centos.org/blue/organizations/jenkins/fedora-f29-pr-pipeline/detail/fedora-f29-pr-pipeline/42/pipeline
• https://apps.fedoraproject.org/datagrepper/id?id=2019-4c4b5cae-47a3-426b-bdbb-d46d48d8fa3c&is_raw=true&size=extra-large

hmm @pingou please any idea what is going on here?

@mvadkert honestly, nope, I've tried looking at if the messages makes it to this and what's going on beside that, but I've not been able to dive deeply into this and thus I don't know what's going on. There are no errors in the logs so there is nothing obvious or I looked at the wrong place/logs.

Sorry I can't be of more help right now.

@pingou would it make sense to propose the code @bkabrda has made for flag updating? That should have more consistent logging directly on pagure. This is a crucial piece of the feedback loop. I would love to help somehow, but not sure how, as we do not have any access to debug this :/ There should have been nothing changed in ci-resultsdb-updater or messages that would break this.

Updating that the issue should be finally fixed with this fix from @pingou:

https://infrastructure.fedoraproject.org/cgit/ansible.git/commit/?id=8e7ceb59f0fac38975292dbb4dd09b33beb07649

Confirmed, pending is now shown as expected.