Details
-
Improvement
-
Status: Resolved (View Workflow)
-
Major
-
Resolution: Fixed
Description
There should be an option by which the normal suppression of Multibranch projects for origin PRs would be disabled, so you would get two projects:
- one building the head of the branch
- one building the speculative merge commit.
There could also be an option to suppress the origin branch project when a PR is filed from the branch. This would cause the origin branch project to be deleted once the PR is filed. (It could be resurrected if the PR is closed but the branch not deleted.) This option would necessitate a new API in github-branch-source to filter branches.
Attachments
Issue Links
- blocks
-
JENKINS-34728 Pull Request pipeline fails to checkout successive commits
-
- Resolved
-
-
JENKINS-35843 Pull requests from branches (not forks) should show up as pull requests
-
- Resolved
-
- is blocked by
-
JENKINS-36574 Add option to choose static notification context
-
- Resolved
-
- is duplicated by
-
JENKINS-33530 Pull Requests "submitted from origin repository, skipping" not clear why this is skipping
-
- Resolved
-
- is related to
-
JENKINS-34120 Do not suppress unmergeable PRs
-
- Resolved
-
-
JENKINS-33237 Failure to check out commit from a closed & reopened PR
-
- Resolved
-
-
JENKINS-34931 Build merge commit instead of PR head
-
- Resolved
-
-
JENKINS-36283 PR builds customization
-
- Resolved
-
-
JENKINS-35991 Control of BB PR behavior on multi-branch pipelines
-
- Resolved
-
- links to
Well, they are related in that all reasonable options about what to build and based on what reference ought to be offered.
Tricky because the GitHub API documentation is frustratingly vague and incomplete. When getting a pull request, there is an undocumented .base.sha which seems to be the common ancestor between the PR branch and the base branch—very different from pull/*/merge and merge_commit_sha, which both seem to represent a virtual merge of the PR head onto the current base branch head (or, rather, a time-delayed snapshot thereof, and as noted in
JENKINS-33237not a very reliable one).