Repository navigation
det/CICS: BLL cells and SERVICE RELOAD (OS/VS COBOL addressing, era 1) #5119
Description
What: support BLL cells (OS/VS COBOL base-locator-for-linkage: a LINKAGE SECTION 01 of pointers whose entries address the following 01s) and SERVICE RELOAD in the det translator, the oracle stub and the scanner's facts, per IBM's CICS/VS application programming documentation for OS/VS COBOL: EXEC CICS ADDRESS / GETMAIN / READ SET into a BLL cell, SERVICE RELOAD after changing a cell, the 4096-byte addressability per cell for long records. Today: 0 references in the translator, standards or register. Map it onto the pointer lane (#5041, #5042) rather than a parallel model.
Acceptance: BLL programs translate or are refused by name per member; register entries with IBM sources (what IBM leaves open is ASSUMED/REFUSED by name); hand-traced cics-crucible cases (no corpus program uses BLL yet); class sweep over the BLL idioms.
Context (owner-approved 2026-10-14): our CICS corpus is almost all eras 3-4 (CardDemo, CBSA, GenApp are modern demos) on top of well-modelled old COBOL semantics. Real estates are layered: 1980s programs LINK 1990s ones that call 2010s Db2 code in one region. Era model used here (approximate; the scanner should cite evidence, not dates):
| Era | ~When | CICS | COBOL | Tells |
|---|---|---|---|---|
| 0 | ~1969-80 | macro level (DFHPC/DFHFC, CSA/TCA) | ANS 68 | DFHxx macro calls, no EXEC CICS |
| 1 | ~1978-88 | CICS/VS 1.x command level | OS/VS COBOL (74) | BLL cells + SERVICE RELOAD, HANDLE CONDITION/AID, DL/I, FCT/PCT/PPT/DCT macro tables |
| 2 | ~1988-96 | CICS/MVS 2.x, CICS/ESA 3-4 | VS COBOL II / COBOL 85 | ADDRESS OF / SET ADDRESS, RESP/RESP2, EXEC SQL, CSD (RDO), DPL |
| 3 | ~1996-2005 | CICS TS 1-2 | COBOL for MVS/OS390, LE | MQ, EXCI, 3270 bridge, web |
| 4 | 2005+ | CICS TS 3+ | Enterprise COBOL | channels/containers, SOAP/JSON, async, named counters |
Order: after the current CI work and CICS blocker work (owner). Siblings: the five "era" issues filed together.
🤖 Generated with Claude Code
Activity
squid-protocol commentedon Oct 14, 2026
Siblings (era work, filed together 2026-10-14): #5117 era tags, #5118 estate-crucible era profiles, #5119 BLL cells / SERVICE RELOAD, #5120 resource macro tables, #5121 macro-level / Assembler refusals. Queued after the CI and CICS blocker work (owner).
🤖 Generated with Claude Code
squid-protocol commentedon Oct 14, 2026
Crucible-first backfill steps (cics-crucible, Sonnet ~20-30 min each)
Owner 2026-10-14: these cases are written as churn backfill while gitgalaxy CI is busy. cics-crucible has no gating CI, so a PR there costs almost nothing on the shared runner pool. Each step is one case with expected logs hand-derived from cited IBM docs (cics-crucible AGENTS.md), gated by
tools/validate.py, as one PR there. The tag and gitgalaxy pin bump are a separate step that needs owner OK.Owner decision first: cics-crucible's SPEC §2 fixes the reference region to Enterprise COBOL on CICS TS 6.x, and
syntax_check.py(cobc) can't speak to OS/VS COBOL. Decide whether era-1 BLL cells belong here (behind an additivelanguage: "osvs-cobol"key) or in estate-crucible's era strata (#5118). Nothing below is queued until then.language: "osvs-cobol", stating what BLL cells and SERVICE RELOAD mean, plus validator support. Cite: CICS/VS (or CICS/MVS 2.1) Application Programmer's Reference, OS/VS COBOL "BLL cells" and "SERVICE RELOAD" (find the publication number and section).Release (shared): batch the merged case PRs, run gitgalaxy's
cics_cruciblerunner against crucible main and review the baseline ledger. With owner OK, tag the next version after v0.5.0, bumpPINNED_REFintests/_cics_crucible_pin.py, and re-baseline: oneci-lane item when CI is calm.squid-protocol commentedon Oct 14, 2026
Owner decision (2026-10-14): era-1 BLL cells / SERVICE RELOAD crucible coverage goes in estate-crucible (#5118 S4 BLL cells, S5 SERVICE RELOAD, S6 the >4096-byte record), not cics-crucible. cics-crucible's reference region stays Enterprise COBOL on CICS TS 6.x. The cics-crucible C1-C3 steps above are withdrawn; this issue's translator, oracle-stub and scanner work is unchanged.
det/CICS BLL: read a program only on BLL evidence (SERVICE, SET of a …
det/CICS BLL: ruff / mypy clean (Part of #5119)
squid-protocol commentedon Oct 14, 2026
Review of PR #5159 (Part of this issue; BLL cells + SERVICE RELOAD lowered onto the pointer lane per IBM's conversion, register X56). Still open here after it lands:
🤖 Generated with Claude Code
det/CICS: OS/VS COBOL BLL cells and SERVICE RELOAD on the pointer lan…
BLL: register X56 (chained areas), RT-CICS-BLL-ADDRESS, e2e proofs, r…
bll: narrow the paragraph match for mypy (Part of #5119)
squid-protocol commentedon Oct 15, 2026
Siblings: #5197 (gaps from the PR #5196 review)
Merge origin/main (Part of #5119)
bll: initialise `through` before the guarded block (CodeQL py/uniniti…
Merge origin/main (Part of #5119)
det/CICS: BLL SET(cell) class and OS/VS COBOL chained storage areas (…
det/CICS: a BLL cell cleared by MOVE ZERO / LOW-VALUES is SET cell TO…