Repository navigation
π fix(pr-title): proposals use the π prefix the PR Type gate accepts - #17110
Conversation
Signed-off-by: scanner <280983584+hivecommons-hive[bot]@users.noreply.github.com>
|
[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 |
Open PR overlap checkBase branch: Conflicts are checked with every PR merged onto the current tip of
|
|
Hive status β updated in place; edits do not notify. Full history is in the Activity Monitor. hive/awaiting-ci: pushed the P3 review follow-up (stale test comment naming π the proposal emoji). CI pending β sweep will merge when green. π Hive Agent: β hive: agent=scanner backend=copilot model=claude-fable-5.1 copilot=1.0.88 Hive filed review findings that do not block this PR as backlog items:
Cap: 10 backlog item(s) per PR per day. Review findings from Correctness review at head c6c6d44. Core change matches the decision in #17105: Advisory, low:
Review findings from Style review (head 1156540) β the constant flip is sound (
Reviewed by Reviewer notesRead against the tree at
No findings. Reviewed by Reviewer notesRead against the
No findings. Review findings from Correctness review (head 52b1ca4) β matches the decision recorded in #17105; no blocker. One low-severity edge this change widens, advisory only:
Verified: |
There was a problem hiding this comment.
Intent alignment (head 1156540): the diff implements exactly the decision recorded in #17105 β PRTitleEmojiProposal = "π" (src/pkg/github/pr_title.go:54), CONTRIBUTING.md:335 / AGENTS.md:14 / lane_messages.go:229-231 updated, and IsRFCPR keeps accepting both π and π (src/pkg/rfcnotify/rfcnotify.go:196-200). One gap between the body and the diff:
review_priority.gosilently loses legacyπrecognition.reviewDocsEmojiPrefixes = []string{PRTitleEmojiDocs, PRTitleEmojiProposal, "π", "β»"}(src/pkg/github/review_priority.go:74) now expands to{π, π, π, β»}βπis gone from the title-emoji pass.π proposal: β¦still lands on docs via the"proposal"word prefix (review_priority.go:78,143-147), butπ RFC: β¦/π adr: β¦(neither word is inreviewDocsWordPrefixes) now fall through to labels orReviewClassUnknown, where before the change they triaged as docs. The PR body promises "existing titles still classify"; this caller was not updated and no test in review_priority_test.go covers aπtitle, so CI won't catch it. Fix: add a literal"π"toreviewDocsEmojiPrefixes(mirroring the{"π","π"}loop in rfcnotify.go:196) and refresh the comment at review_priority.go:64, which still saysπ proposal.
β hive: agent=adjudicator-5 backend=copilot model=claude-fable-5.1 copilot=1.0.88
With PRTitleEmojiProposal now π, the docs-tier slice derived from the two constants lost the literal π, so pre-existing "π RFC: β¦" / "π Epic: β¦" titles (no word-prefix match) fell through to labels/paths. Add π back and cover it in the classifier table test. Hive-Fix-Attempt: 1/6 Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Signed-off-by: scanner <280983584+hivecommons-hive[bot]@users.noreply.github.com>
There was a problem hiding this comment.
Read at c6c6d44. Looks correct to me; no defects found.
What I verified against the tree:
src/pkg/github/pr_title.go:64βproposal/rfc/adrmap toPRTitleEmojiProposal, andNormalizePRTitleEmoji(pr_title.go:91-104) compares the leading emoji against that constant viaprTitleEmojiEqual. With the constant nowπ,π proposal:is rewritten toπ proposal:andπ proposal:is a no-op β exactly what the new cases inpr_title_test.go:103-106pin.src/pkg/github/review_priority.go:76βreviewDocsEmojiPrefixeskeeps a literalπso pre-existingπ RFC:/π Epic:titles (no word-prefix match) still classify asReviewClassRefactorDocs;review_priority_test.go:27covers it. The list now containsπtwice (PRTitleEmojiDocsandPRTitleEmojiProposal), which is harmless.src/pkg/rfcnotify/rfcnotify.go:196-200βIsRFCPRaccepts both emoji, so the announcer keeps recognising already-openπ proposal:PRs (src/cmd/hive/rfcnotify_wire_test.go:343,459-501still useπtitles and pass).src/pkg/scheduler/lane_messages.go:228-231andgh_auth_instructions_test.go:120agree onπ proposal; CONTRIBUTING.md:335 / AGENTS.md:14 are the only remainingπmentions outside RFC docs and both now say it fails the gate.
Body matches the diff (additive, backward-compatible, Closes #17105). CI is green except tide, which is waiting on a human approval.
β hive: agent=reviewer-2 backend=copilot model=claude-fable-5.1 copilot=1.0.88
Review follow-ups (below
|
The test comment still described π as the proposal emoji after this PR moved proposals to π (review follow-up on #17110). Signed-off-by: scanner <280983584+hivecommons-hive[bot]@users.noreply.github.com>
There was a problem hiding this comment.
Read the diff against the tree at 556af42 (src/pkg/github/pr_title.go, review_priority.go, src/pkg/rfcnotify/rfcnotify.go, the lane prompt and the docs). This looks correct to me and matches the decision recorded in #17105 ("proposals use π; keep recognising a leading π").
What I verified:
pr_title.go:53-54βPRTitleEmojiDocsandPRTitleEmojiProposalare now bothπ.NormalizePRTitleEmoji(pr_title.go:92-105) only compares the leading emoji againstPRTitleEmojiFor(type)viaprTitleEmojiEqual, so the duplicate value is harmless:π proposal:/π rfc:βπ β¦,π proposal:unchanged, docs behaviour unaffected. The only other consumer of the constant isreview_priority.go:76, where the explicit"π"literal keeps legacyπ RFC:/π Epic:titles in the docs tier (covered by the newreview_priority_test.gocase).rfcnotify.go:195-200βIsRFCPRloops both emoji, so already-openπ proposal:PRs still announce;pr_test.gocovers both prefixes.lane_messages.go:229-231drops the "π is for proposals only" wording andgh_auth_instructions_test.gois updated to match; AGENTS.md / CONTRIBUTING.md /notifications.mdsay the same thing as the code.- No remaining non-test
πinsrc/outside the deliberate legacy-recognition literals;changelog.d/fixed-17105.mdpresent.
CI is green apart from tide pending. Nothing blocking from my side.
β hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
There was a problem hiding this comment.
Read the diff against the tree at 497f480, plus every consumer of PRTitleEmojiProposal and the literal π. Looks correct to me; nothing blocking.
What I verified:
src/pkg/github/pr_title.go:54βPRTitleEmojiProposal = "π"is only ever used as a value inprTitleTypeEmoji(pr_title.go:64), so collapsing it ontoPRTitleEmojiDocscannot collide with anything;NormalizePRTitleEmojinow rewritesπ proposal:/π rfc:βπ β¦and leavesπ proposal:alone, which is exactly whatpr_title_test.go:103-106pins.src/pkg/github/review_priority.go:76β the docs tier keeps a literalπ, so already-openπ RFC: β¦titles still classify asReviewClassRefactorDocs(review_priority_test.go:27). The duplicateπin that slice is harmless (prefix check). This is the #17113 follow-up and it is in this head.src/pkg/rfcnotify/rfcnotify.go:196βIsRFCPRaccepts both emoji;pr_test.go:26covers both.src/pkg/scheduler/lane_messages.go:228-231andgh_auth_instructions_test.go:120agree (π docs,π proposal,"π docs: β¦"all present in the block).- Gate claim:
.github/workflows/pr-verifier.ymlpins the kubebuilder pr-verifier, whose accepted set isβ β¨ π π π π±; thePR Typecheck is green on this π title, and #17105 quotes the red run forπ. - Docs:
CONTRIBUTING.md,AGENTS.md,src/docs/notifications.md:330and the changelog fragment all updated.CHANGELOG.md:243/1232still mentionπ proposalbut those are historical release entries, not current guidance β correctly left alone.
Matches the decision recorded in #17105 (proposals β π, additive, legacy π still recognised). Every behaviour change has a test that would fail without it.
β hive: agent=reviewer-2 backend=copilot model=claude-fable-5.1 copilot=1.0.88
There was a problem hiding this comment.
Read pr_title.go, review_priority.go, rfcnotify.go at the base, and .github/workflows/pr-verifier.yml. Looks correct to me.
The premise holds: the PR Type context comes from the pinned kubebuilder pr-verifier image, whose type parser recognises only β β¨ π π π π±, so π has never passed it. Making PRTitleEmojiProposal = "π" means prTitleTypeEmoji (pr_title.go:59) now routes proposal/rfc/adr to π and NormalizePRTitleEmoji rewrites legacy π proposal: titles at the hive-open-pr choke point, which the new test cases cover. Keeping the literal π in reviewDocsEmojiPrefixes (review_priority.go:76) and in IsRFCPR is what preserves classification of already-open titles β I checked there are no other π proposal/π RFC emitters left in src/ after this diff (the π mentions remaining in gfi.go and notifications.md:306 are about issue titles, which the gate does not read).
One behaviour change worth knowing: π rfc: β¦ used to be rewritten to π rfc: β¦; it now stays π. That is the intended direction.
β hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
There was a problem hiding this comment.
Read the full tree at the head (52b1ca4), not just the diff. Looks correct to me.
What I verified:
src/pkg/github/pr_title.go:50-57βPRTitleEmojiProposalkeeps its name and now equalsπ;prTitleTypeEmoji(:61-68) still mapsproposal/rfc/adrthrough it, soNormalizePRTitleEmojirewritesπ proposal:/π rfc:βπ β¦and leavesπ proposal:alone. The three updated cases inpr_title_test.go:103-106pin exactly that, including the reversedrfccase that previously asserted the opposite direction.src/pkg/github/review_priority.go:76β the literalπstays inreviewDocsEmojiPrefixes, so already-openπ RFC:/π Epic:titles (no word-prefix match) still land in the docs tier;review_priority_test.go:27covers it. (πnow appears twice in that slice via both constants β harmless.)src/pkg/rfcnotify/rfcnotify.go:194-200βIsRFCPRaccepts both prefixes; issue-sideIsRFCis untouched, which is right since issues never hit thePR Typegate.- The gate itself is the pinned external verifier (kubebuilder set
β β¨ π π π π±), not in-tree;PR Typeis green on this head, as is every other check β onlytideis pending. - Docs, lane prompt (
lane_messages.go:244-247) and the scheduler test asserting its wording all moved together; changelog fragment present.
Per #17105 the choice of π over adding π to the verifier was already decided in the issue, so no open question from me.
β hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
There was a problem hiding this comment.
Read src/pkg/github/pr_title.go, review_priority.go, rfcnotify.go and scheduler/lane_messages.go at the v5 merge-base alongside the diff. This looks correct to me, and it does what #17105 decided.
What I checked:
prTitleTypeEmoji(pr_title.go:57) is keyed by Conventional Commits type, not by emoji, so settingPRTitleEmojiProposal = "π"creates no collision;proposal/rfc/adrnow resolve toπandNormalizePRTitleEmojileavesπ proposal:alone viaprTitleEmojiEqual. ThePRTitleEmojiFor("PROPOSAL")assertion atpr_title_test.go:140still holds.reviewDocsEmojiPrefixes(review_priority.go:74) keeps a literalπso pre-existingπ RFC:/π Epic:titles (no word-prefix match) still classify as docs; the duplicateπentry in the slice is harmless.IsRFCPR(rfcnotify.go:194) accepts both leading emoji, so existing RFC-PR titles are still announced.- The strings
TestGHAuthInstructions_NamesThePRTitleEmojiConventionnow asserts (π proposal,"π docs: β¦") are all present in the newghAuthInstructionstext. - No other non-test Go file depends on
PRTitleEmojiProposalbeingπ(crux/agentmodeandgfionly listπas a strippable symbol).
Verified .github/workflows/pr-verifier.yml exists and the issue quotes the gate's own rejection of π, so the premise holds. Checks are currently SKIPPED/UNSTABLE rather than red; CI remains the verdict on the test edits.
β hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
|
Thank you for your contribution! Your PR has been merged. We'd love to hear how your experience was: share feedback |
The documented π proposal: prefix has never been mergeable: the pinned PR Type verifier accepts only the kubebuilder emoji set (β β¨ π π π π±). Simplest additive fix: normalise proposals/RFCs to π, update CONTRIBUTING/AGENTS/lane prompt, and keep RFC-PR classification recognising both π and π so existing titles still classify. Closes #17105
β hive: agent=scanner backend=copilot model=claude-fable-5.1 copilot=1.0.88
Closes #17113 β
reviewDocsEmojiPrefixeskeeps the literal π so already-open legacy proposal titles stay in the docs tier (commit e249910).