Release override results in jobs being created over and over again #9
Loading…
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
PR #7 seems to result in jobs being created over and over again.
Example Workflow:
Is there an easy way to clear the job queue until this is fixed?
I would recommend stopping the runner. And then either manually cancel each of them. Or do that with a SQL update. There is no API yet to do the same.
Okay, sounds good!
It is very difficult (if at all possible?) to prevent infinite loops. In this case you seem to have a workflow triggered by the creation of a release and that workflow creates a release, which leads to an infinite loop. Or am I reading this incorrectly?
That is correct. I create a release and the idea was that this workflow would overwrite the release with release artifacts. Which is why I opened issue #8 to make this experience a little better.
Is there another workflow that I can run that only uploads release artifacts to forgejo releases?
What about using the workflow you have with
instead?
forgejo-release
will create the release for this tag but will not recreate the tag. You don't need to override then.That should work for now. It just means that I have to create a tag and edit releases after for manual release notes.
Creating a tag is a good way to say: "I want a release on that commit". You could also have a step before
forgejo-release
that extracts the release notes from a file in the repo if that's easier than doing it manually afterwards.That’s actually a great compromise. A changelog.md will ensure I have a portable version of release notes!
I'm not sure why, but it turns out it's been creating a new release over and over again for a while now after I cleaned it up the first time.
Is
actions/forgejo-release@v2
bugged?Turns out it's from the older commits? I stopped the runner and migrated to a new repo, so not sure how they picked up these jobs again. Was it maybe from migrating releases? Ugh... lol.
Okay, so I guess after making a new repository, deleting the old, and then renaming the new one to the old one; the runner had a bunch of jobs queued up again after starting it back up.
I wrote a quick and dirty tamper monkey script to click cancel on the several hundred pending jobs in queue.
This seems to have cleared it all up!
However, there has got to be a way to prevent a loop like this- and if it's not possible a clear all jobs button would be incredibly useful.