basis-build: repair interrupted releases, group master runs by commit - #5
Merged
Merged
Conversation
The guard added in #1 could wedge the publish job. `gh release create` uploads assets to a draft and publishes it last, so an interrupted run leaves a draft; `gh release view` finds drafts, so every later run exited 0 and the commit never got a release. A draft was never public, so it is now deleted and rebuilt. A published release missing an asset fails loudly instead of being kept, and a complete one is still kept. Master runs also group by commit SHA: a group holds only one pending run, so a per-ref group could still drop a run queued behind a long build and leave that commit without a release. Ports the fixes review found on BasisResearch/cvc5#16 and BasisResearch/z3#2. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Follow-up to #1, which shipped a guard that can wedge the publish job.
gh release createuploads assets to a draft and publishes it only once every upload finishes, so a run interrupted mid-upload leaves a draft behind.gh release viewfinds drafts, so #1's guard then exited 0 on every later run and that commit never got a release. A draft was never public and nobody can have pinned it, so it is now deleted and rebuilt. A published release missing any of its assets fails with::error::(delete it by hand to rebuild) rather than being silently kept, and a complete published release is still kept untouched.Master runs now also get a concurrency group per commit SHA. A group keeps only one pending run and cancels the rest, so the per-ref group from #1 could still drop a run queued behind a long build and leave that commit without a release. PR runs still group by ref and cancel superseded builds.
Review found these on the cvc5 and z3 ports of the same guard (BasisResearch/cvc5#16, BasisResearch/z3#2); this brings tlaplus in line. The jq asset check was verified against the real
basis-9f3709e669release, which reports all three assetsuploaded. The publish job does not run on pull requests, so as with #1 the publish path only exercises on merge.🤖 Generated with Claude Code