Skip to content

Planner rewrites skip reasons when lenses are empty #71

Closed

Description

@pkieltyka

Problem: A planner skip with a valid reason and empty lenses has its reason rewritten to planner_empty_lenses: ... and default lenses are added. The hunk remains skipped, but the original diagnostic reason required by the component spec is lost.
File:line: src/pipeline/planner.ts:1422-1428; specs/project/components/review_pipeline.md:455.
Evidence: Schema accepts lenses: [] on a skip with valid reason; @architect-2 confirmed the reason is altered before packet-builder logs it.
Suggested fix size: S.
Found during #44 (workstream U, confirmed by @architect-2).

Activity

  1. pkieltyka commented on Nov 19, 2026

    @pkieltyka
    CollaboratorAuthor

    Tracking update: status In progress; owner peter/@lead-dev-2; branch peter/44-fixes-pipeline; grouped PR covers #59, #60, #61, #62, #63, #64, #65, #71, #72 and #73. Criteria: a regression test that fails without the fix; check and test green; @Reviewer approval. Next step: implement.

  2. pkieltyka commented on Nov 19, 2026

    @pkieltyka
    CollaboratorAuthor

    Review update (supersedes the prior next-step record): status In progress; owner peter/@lead-dev-2; branch peter/44-fixes-pipeline at head 90811e9 (10 commits); @Reviewer is reviewing batch 2. Grouped set: #59-65 and #71-73. Criteria: a regression test that fails without the fix; check and test green; reviewer approval. Next step: await @Reviewer’s verdict, address any required changes, then open the grouped PR.

  3. pkieltyka commented on Nov 19, 2026

    @pkieltyka
    CollaboratorAuthor

    Review update (supersedes the prior next-step record): status In progress; owner peter/@lead-dev-2; branch peter/44-fixes-pipeline, reviewed commit 90811e9 against base f240376; grouped batch #59-65 and #71-73. Verdict: changes required for one P2 finding on #60 at src/pipeline/composer.ts:127-139,402-416. If runComposer rejects with undefined, compositionFailure is set to undefined, the sentinel check is false, and deterministic fallback returns as success without stage_failed; string/null reasons are rethrown but cannot be WeakMap keys, so review-runner loses retained findings. Required fix: use an explicit failure-present flag and a report handoff that also handles primitive rejection reasons. Checks passed: pnpm check; pnpm test (69 files / 1598 tests); git diff --check. Other criteria met. Next step: implement the #60 rejection-reason handoff fix, then request reviewer re-review.

  4. pkieltyka commented on Nov 19, 2026

    @pkieltyka
    CollaboratorAuthor

    Review update (supersedes the prior changes-required/awaiting-review record): status In progress; owner peter/@lead-dev-2; branch peter/44-fixes-pipeline; grouped batch #59-65 and #71-73; reviewed commit 747cd06, one commit on 90811e9, against f240376. Verdict: approve; @Reviewer fast-lane re-review, no findings. Checks passed: pnpm check; pnpm test (69 files / 1601 tests); git diff --check. The prior P2 on #60 is fixed: explicit failure presence preserves undefined, and primitive or null rejection reasons are wrapped for the WeakMap handoff; undefined/string/null regressions cover composer rejection and retained verified findings. No grouped PR is open yet. Next step: operator review, then open the grouped PR.

  5. pkieltyka commented on Nov 19, 2026

    @pkieltyka
    CollaboratorAuthor

    PR update (supersedes the prior next-step record): status In progress; owner peter/@lead-dev-2; branch peter/44-fixes-pipeline; grouped batch #59-65 and #71-73; PR #89 #89 is OPEN at head 747cd06 against master 6d78590. Verdict: @Reviewer approved the fast-lane re-review at this head with no findings. Checks on the reviewed commit passed: pnpm check; pnpm test (69 files / 1601 tests); git diff --check. The prior P2 on #60 is fixed: explicit failure presence preserves undefined, primitive/null rejection reasons are wrapped for the WeakMap handoff, and undefined/string/null regression cases verify composer rejection and retained verified findings. PR checks are currently in progress. Next step: CI, then the owner merges; keep issues open until merge is confirmed.

  6. pkieltyka commented on Nov 19, 2026

    @pkieltyka
    CollaboratorAuthor

    Review update (supersedes the prior approval/next-step record): status In progress; owner peter/@lead-dev-2; branch peter/44-fixes-pipeline; grouped batch #59-65 and #71-73; PR #89 #89 is OPEN at exact head c0db0c0 against master 6d78590. Verdict: changes required on @Reviewer fast-lane re-review for one P3 test gap only. CI, CodeQL and Codegenie review checks are green; git diff --check is clean. Prior #60, #65 and #71 findings are fixed; full Codegenie answers were sent to the operator. Required fix: (1) in the deep-planned test-file packet with an E1 delta, assert no coverageEscalation metadata/event; (2) make the applyCoverageEscalations test input and output objects distinct so mutation cannot pass through aliased references. Next step: add these regression assertions and request @Reviewer re-review.

  7. pkieltyka commented on Nov 19, 2026

    @pkieltyka
    CollaboratorAuthor

    Review/fix update (supersedes the prior changes-required next-step record): status In progress; owner peter/@lead-dev-2; branch peter/44-fixes-pipeline; grouped batch #59-65 and #71-73; PR #89 #89 now has head 91cb330 against master 6d78590. The prior P3 gap at c0db0c0 required (1) a deep-planned test-file packet with a qualifying E1 delta to assert no coverageEscalation metadata and no coverage_escalated event, with failure if the planned-deep guard is removed; and (2) distinct comparison values so applyCoverageEscalations mutation cannot pass through aliased references. Commit 91cb330 adds these tests; the results are checked against a structuredClone snapshot taken before the call. Developer reports pnpm check passed and pnpm test passed (69 files / 1604 tests). PR checks are rerunning. Reviewer verdict remains changes required on c0db0c0 pending verification of this fix. Next step: @Reviewer verifies only the 91cb330 diff, per the operator decision, then records the verdict.

  8. pkieltyka commented on Nov 19, 2026

    @pkieltyka
    CollaboratorAuthor

    Final completion record (supersedes prior review/next-step records): status completed; owner peter/@lead-dev-2; branch peter/44-fixes-pipeline; grouped batch #59-65 and #71-73; PR #89 #89 was squash-merged at 2026-11-19 22:55:13Z as merge commit 4f21784. Verdict: approve at exact head e09ab5a. CI, CodeQL and Codegenie review passed; no review threads; five focused merge-resolution tests passed. Prior P3 test gaps are fixed. The operator accepted the uploaded Codegenie report as the complete available set; visible questions were answered and no finding remains open. Next step: none; integrated and completed.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't working

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions