Open eppisapiafsl opened 4 months ago
First change I’d make is I would add the --failOnErrors
parameter to your danger ci
— from what I can see of your workflow, you expect GitHub Actions to report if Danger dies.
Your danger step might never have started to execute. An earlier step in the same job is still running, or some other (GitHub Actions) weirdness is happening.
@eppisapiafsl if you want more help, please share more of your GitHub workflow file. If that’s too complicated, I would dive in to GitHub Actions to watch the execution logs. You (or a repo-admin) can watch the full logs as each step in the workflow gets executed.
Watching on the PR-page can have some browser-weirdness / doesn’t show every event sequentially, it only shows the final result for the steps that are tagged as “required”. You want to prove that danger is even executing at all, and that may show log messages that help you debug the issue.
Thanks @fbartho , I didn't share the whole file as the previous steps are just "yarn install" setups 😅
I can confirm, that the Danger JS
GHA finished as expected but the webhook didn't post the message or get updated
I believe the merge_group
trigger doesn't populate the PR information, so Danger gets stuck trying to find the PR to post 😅
I believe the merge_group trigger doesn't populate the PR information
Ah! That’s probably it. In retrospect Merge Groups sort of include every PR that is in the group together, so maybe we’d want special handling so you could access the PR info for every PR involved?
If you read the GitHub Actions doc, there’s probably a schema/payload for the Merge Group event. That may be something that DangerJS could load in. — We have a danger.pr
payload so maybe there should be an array of danger.mergeGroup
?
I’m definitely motivated to support making merge groups more convenient, but the feature wasn’t generally available the last time I set a team up with a CI workflow.
Yeah, I am still diving into the migration I also found other issues as the merge group trigger doesn't support all the options that we have in the pull_request
We have a danger.pr payload so maybe there should be an array of danger.mergeGroup This would be amazing, so we provide feedback when the CI fails in the merge group
In the short term, what about a parameter to tell the Danger server to "do nothing"
danger --ci --merge-group
I could try with an empty danger file, but can be avoided with a parameter of if Danger recognizes the trigger
Describe the bug
We have configured "Danger" as a required step in the CI. It works as expected in Pull Request and After the merge process, but we face a problem configuring it for the new GitHub Merge Queue action (Configure to group 3 PRs at time)
To Reproduce Steps to reproduce the behavior:
jobs: danger_merge_group: runs-on: ubuntu-latest name: Danger JS steps: ...
The danger file to run in Merge group is quite simple 😅
Expected behavior
Danger populates the result of the GHA in the Merge Group action.
Screenshots If applicable, add screenshots to help explain your problem.
Your Environment
Additional context