Skip to content

equivalence: cases for the 16 CICS programs that have none (6 estates) #5076

Closed

Description

@squid-protocol

16 CICS programs in 6 estates have no equivalence case at all (13 of them have no other gap), so they can never be proven. 8 are in burned estates.

Do: add an equivalence case per program (scenarios over its whole input class, per the pr-worker rules), starting with the 13 whose only gap is the missing case; list the 16 with estate and status in the PR. Burned estates only; never loosen a proof.

Acceptance: proof_blockers shows no "no equivalence case" entry for a burned-estate CICS program; each new case is proven or names its holes.

Measured on origin/main 36671d5 (2026-10-13) over the 8 local corpora + 10 census clones (12 with CICS): cics_census.py survey --baseline --census-corpora <census> then proof_blockers.py <survey dir> --census-corpora <census> --json. 103 of 125 CICS programs translate whole; only 14 are proven (all CardDemo BMS).

Owner-approved re-ranking of the CICS queue by estate/program reach (2026-10-13).

🤖 Generated with Claude Code

Activity

  1. added
    enhancementNew feature, sensor, or structural signature
    testingUnit, integration, and E2E pipeline verification
    legacy-modernizationCOBOL refractor, dead-code extraction, and JCL forging
    priority: highCore feature broken, but workarounds exist
    on Oct 14, 2026
  2. squid-protocol commented on Oct 14, 2026

    @squid-protocol
    OwnerAuthor

    Siblings: #5089 (the members PR #5087 left open: ZECS000 install symbols, nested RUN TRANSID, EIBFN literal fact, facade RETURN-length refusals).

  3. added
    tier:scoreChurn tier 3: translation scores / proofs of named programs
    on Oct 15, 2026
  4. added
    scope:coreBlocks a burned-estate program's next evidence rung, or the 4th-estate one-shot trial
    era:2Era CICS/MVS-ESA, COBOL II/85: RESP, SQL, DPL (~1988-96) (era_tags.py)
    on Oct 16, 2026
  5. squid-protocol commented on Oct 16, 2026

    @squid-protocol
    OwnerAuthor

    Acceptance met by PR #5279: re-measured on HEAD (proof_blockers, burned estates), no burned-estate CICS program carries a "no equivalence case" gap any more. The new cases are BNK1CCA (proven) and DELCUS (not proven, holes named: file change after SYNCPOINT ROLLBACK, now tracked on #5089). The other programs in the list were cased earlier and keep their named blockers (#5089, #4952/#5064, #4771). The issue's 16-program list was measured on an older commit and covered non-burned estates, which are out of scope by policy.

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

    enhancementNew feature, sensor, or structural signatureera:2Era CICS/MVS-ESA, COBOL II/85: RESP, SQL, DPL (~1988-96) (era_tags.py)legacy-modernizationCOBOL refractor, dead-code extraction, and JCL forgingpriority: highCore feature broken, but workarounds existscope:coreBlocks a burned-estate program's next evidence rung, or the 4th-estate one-shot trialtestingUnit, integration, and E2E pipeline verificationtier:scoreChurn tier 3: translation scores / proofs of named programs

    Projects

    No projects

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions