Repository navigation
🐛 gfi: promise an automatic merge only where the hive merges - #18832
hivecommons-hive[bot] merged 1 commit into
Conversation
The good-first-issue welcome comment and its agent prompt always said the hive merges green PRs from claimed contributors automatically. Where a repository's auto-merge is off (below L6, or repo_policies auto_merge: false) the contributor lane skips the repo and a maintainer merges, so the invite promised something the hive does not do there. gfi.Vars gains AutoMerge, filled per issue from SweepOptions.RepoAutoMerge, which gfiSweepOptions wires to Config.RepoAutoMergeEnabled. With it on the rendered text is byte-identical to before; with it off the heading, step 4, the footer link label and the agent prompt say a maintainer merges. Signed-off-by: Danathar <6772335+Danathar@users.noreply.github.com> Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BQYnPGFJx81c8JXR3o1sg1
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: The full list of commands accepted by this bot can be found here. DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
There was a problem hiding this comment.
Docs currency looks fine: the flow, off-wording, and .AutoMerge variable are documented and a changelog fragment is included. One out-of-scope note: the first-timer PR welcome (good-first-issue.md:259) still implies auto-merge unconditionally. Not blocking.
— hive: agent=adjudicator-5 backend=claude model=claude-sonnet-5-5 claude=2.1.326
|
Hive status — updated in place; edits do not notify. Full history is in the Activity Monitor. Review follow-ups: 1 open of 1 (below
|
Open PR overlap checkBase branch: Conflicts are checked with every PR merged onto the current tip of
|
Summary
The good-first-issue welcome comment always promised that the hive "merges green PRs from claimed contributors automatically", and its paste-to-agent prompt said the contributor lane "can auto-merge it once CI is green". Where a repository's auto-merge is off, the contributor lane skips that repository and a maintainer merges, so the invite promised something the hive does not do there.
The comment now promises an automatic merge only where the hive's contributor lane would merge. Elsewhere it says the hive reviews the PR and a maintainer merges it. Repositories the hive does merge get byte-identical text.
What happened
contributor-lane automerge sweep skipped reporepo=commonreason=repo-auto-merge-disabled, and approved contributor PRs there route to090-ready-for-human-merge.Cause
gfi.DefaultTemplatestep 4 andVars.AgentPromptare unconditional, andgfi.Varscarries nothing about merging.repoAutoMergeAllowed, wired fromConfig.RepoAutoMergeEnabled. That is false below L6, and at L6 whereproject.repo_policies[].auto_merge: false.The same wording gap was closed for kicks and the dashboard in #17543 and #17550.
What changed
gfi.Vars.AutoMerge(new): whether the hive merges this repository's contributor PRs itself. Filled per issue fromSweepOptions.RepoAutoMerge, whichgfiSweepOptionswires tocfg.RepoAutoMergeEnabled. When the func is unset the sweep keeps today's promise, so unwired callers render as before.gfistays stdlib-only.AutoMergefalse:AutoMergetrue nothing changes.good_first_issue.comment_templatecan use{{ if .AutoMerge }}.good-first-issue.md: the flow, the off wording, the new template variable) and a changelog fragment.Not in this PR
good_first_issue.reposalready restricts it.blob/v5.Tests
pkg/gfi/automerge_test.go:AutoMergetrue the default template renders byte-identical to the pre-change template, and the agent prompt is unchanged;.AutoMerge;cmd/hive/gfi_wire_test.go:gfiSweepOptionswires the func from config: anauto_merge: falserepo, a repo with no override at L6, and every repo below L6.Only
gofmt -lwas run locally (AGENTS.md bans localgo test/go build/go vet); this PR's CI is the test run.🤖 Generated with Claude Code
https://claude.ai/code/session_01BQYnPGFJx81c8JXR3o1sg1