Repository navigation
Revert A1 (#5615): restore green main — EqualsClaim discovery-corpus closure-poison - #5627
Merged
Merged
Conversation
briansrls
added a commit
that referenced
this pull request
Jun 24, 2026
… keystone; operator+parent SIGNED). CONTEXT: A1 (#5615, the Infra/Structural CI-failure classifier + floor_exit_blocks_merge verdict surface) was REVERTED by #5627 to clear an outage — its closure expansion poisoned EqualsClaim res (#5651) * Revert "Revert "CI A1: infra-vs-structural classifier + green-on-main merge-blocking verdict (single authority, fail-closed) (#5615)" (#5627)" This reverts commit c7962ab. * WIP: A1 RE-LAND on #1 with merge-freshness verdict LIVE (CI-lockdown ARM-A ke * WIP: A1 RE-LAND on #1 with merge-freshness verdict LIVE (CI-lockdown ARM-A ke * WIP: A1 RE-LAND on #1 with merge-freshness verdict LIVE (CI-lockdown ARM-A ke * WIP: A1 RE-LAND on #1 with merge-freshness verdict LIVE (CI-lockdown ARM-A ke * Merge fix: §6 is single-authority in the .dag plan (doc is generated) main's batch-B fanned docs/plans/ci-merge-freshness.md INTO a structured .dag plan (gunbc.plans.ci_merge_freshness) enrolled as a generated artifact (generated_artifact.dag: committed doc must == expected_plan_md(plan), checked by the floor's generated_artifact_drift_gate). My §6 honest-marking landed only in the doc during conflict resolution -> would drift-red post-merge (semantic merge-skew). Fix: add §6 to ci_merge_freshness_body() (the authority) and regenerate the doc from expected_plan_md so committed == rendered. Verified by execution: generated_artifact_drift_witnesses + plan_lens witnesses both true; doc now byte-equals the rendered plan. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * WIP: A1 RE-LAND on #1 with merge-freshness verdict LIVE (CI-lockdown ARM-A ke * Restore dsl-v2-defork-audit.md §2A (do not ship the regen-deletion) main_wet regen deleted §2A (the .dag authority lacks it); that content is #5652's Q2 resolver-wall analysis and must not be lost. Restore the doc to origin/main so my branch is faithful. The §2A-in-.md-not-in-.dag drift is a pre-existing fleet main-red (escalated to lockdown mgr), NOT mine to fix by deletion. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Brian Searls <briansrls@gunb.ai> Co-authored-by: Claude Opus 4.8 (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.
REVERT-TO-GREEN for an ACTIVE main-red outage (design-authority ruling, bright-stag-194; incident lead crisp-carp-603).
Incident
main has been RED since A1's merge (4fceebf, #5615) → cascading B3 (e235d14, #5617) → #5625 (886d9ec HEAD). Last green = 2e32aed (#5615's base). Real failed step
gunbc ci (.dag witnesses + gates)(structural, NOT infra).Root cause
discovery-corpus (batch 2, dsl-first) fails resolving
src/v2/workflow/affected_set_floor_runner_test.dag:69/77/85—type 'EqualsClaim' not found in scope— EVEN THOUGH EqualsClaim is defined+exported insrc/v2/std/verification.dag:173, imported correctly, with NO module collision. = a closure-resolution interference (merge-skew / #5295 frontier-index cache-poison class): A1 green on its base + sibling green on its base, jointly-red on main. Slipped through because A1's green-on-main verdict shipped DORMANT (the exact class it targets).This PR
Reverts A1 (#5615) ALONE — applied clean, no overlap with B3/#5625. CI here is the single-root test: green = A1's closure-poison was the sole root (B3+#5625 go green once A1 is removed). A1 re-lands after the closure-skew is diagnosed off the critical path AND its own verdict is enforcing enough to have caught this.
Do NOT touch the load-bearing resolver to fix-forward under the red (gated). Diagnosis proceeds on a separate branch.