squid-protocol · Aug 30, 2026 · Aug 30, 2026
@@ -89,6 +89,38 @@ directly by `tri_comparison_chart.py` to render the `**` marker and award GitGal
 `docs/self_scan/how_to_investigate_a_discrepancy.md` for the methodology and the
 `tri-comparison-ledger-sweep` skill's manual-verification fallback for the procedure.
 
+### 2026-08-30: re-verified against the v1.2.0 corpus expansion
+
+`language-crucible` v1.2.0 (pin bumped in [#2478](https://github.com/squid-protocol/gitgalaxy/pull/2478))
+added one `.jcl` file — `cics-java-jcics-samples/etc_VSAM_DEFVSAM.jcl`, an IDCAMS VSAM-cluster
+define job — bringing JCL's corpus to **186 `.jcl` files across 6 repos** (185 pipeline-scanned).
+#2478 bumped the pin without refreshing `manual_verification.json`, so the 2026-08-29 record
+(`375 / 117`) went stale against the pinned corpus and JCL's precision panels dropped back to a
+bare `0/376` / `0/118` with no badge. This pass re-established the record.
+
+**Method.** The same three independent readings as the 2026-08-29 pass, over the whole v1.2.0
+corpus: GitGalaxy's own `func_start` / `class_start` regex; an independent line scanner sharing no
+implementation; and a `galaxyscope --db-only` pipeline cross-check (`struct_*` vs. `*_count` vs.
+`function_data` / `class_data` row counts), per file.
+
+| Signature | Corpus truth | Independent scanner | Pipeline | False positives | Misses |
+| :--- | :---: | :---: | :---: | :---: | :---: |
+| `func_start` (EXEC steps) | **376 / 376** | 376 | 376 | 0 | 0 |
+| `class_start` (JOB cards) | **118 / 118** | 118 | 118 | 0 | 0 |
+
+- The single new file contributes exactly **one EXEC step** (`//DEFINE EXEC PGM=IDCAMS`) and
+  **one JOB card** (`//DEFVSAM JOB`). Its instream IDCAMS control statements (`DELETE`,
+  `DEFINE CLUSTER`) between `//SYSIN DD *` and `/*` are correctly **not** matched — the same
+  SYSIN-payload rejection the 2026-08-29 pass verified, holding on new content.
+- Net effect of the v1.2.0 expansion is `+1 / +1`. No engine change, no new defect, no
+  regression: every other JCL count is identical to the 2026-08-29 reading, and per-file
+  `struct_*` still equals the raw regex count on every scanned file.
+- `DFH$SIP1.jcl` remains the one correctly-unscanned `.jcl` (SIP parameter deck, `*`-prefixed
+  lines, zero real signatures).
+
+**Chart records** (`manual_verification.json`, `2026-08-30`): Functions `376 / 376`, Classes
+`118 / 118`. Args unchanged — still `none`-granularity, no entry.
+
 ### 2026-08-29: re-verified against the v1.1.0 corpus expansion
 
 `language-crucible` v1.1.0 (pin bumped in [#2398](https://github.com/squid-protocol/gitgalaxy/pull/2398))

@@ -50,17 +50,17 @@
     },
     "jcl": {
       "class": {
-        "note": "class_start (JCL JOB cards) verified against the full language-crucible v1.1.0 corpus (185 .jcl files, 5 repos). Two independent checks over all 184 pipeline-scanned files: (1) GitGalaxy's class_start regex vs. an independent JOB-card grep (^//[A-Za-z0-9@#$][A-Za-z0-9@#$]{0,7}[ \\t]+JOB([ \\t]|$), no shared implementation, independently rejects //*-commented job cards) -- exact agreement, 117 == 117, zero per-file mismatches. The 12 cics-banking-sample-application-cbsa/CREL0*.jcl + CREDB2L.jcl files whose entire JOB card is commented out (//*NAME JOB ...) and the 2 cics-genapp sampcma/sampwui.jcl files whose jobname is an angle-bracket placeholder (//<CMASAPPL> JOB) are correctly NOT counted -- confirmed by direct source read. (2) real pipeline DB output: struct_class_start == class_count == 117 on every file (JCL is in _CLASS_START_NAMED_EXTRACTION_LANGS since #2203, so no brace-slice fallback). Every corpus file carries at most one JOB card. Zero false positives. Count moved 116 -> 117 when #2415 (aperture.py infra_path_pattern left-boundary) was fixed in the same PR, letting DBRMLIB.jcl through the perimeter -- raw-corpus and pipeline now agree at 117/117. See docs/language_status/jcl.md #9.",
-        "total": 117,
-        "verified": 117,
-        "verified_at": "2026-08-29",
+        "note": "class_start (JCL JOB cards) verified against the language-crucible v1.2.0 corpus (186 .jcl files, 6 repos; 185 pipeline-scanned, same DFH$SIP1.jcl zero-density exclusion as the function entry). Three checks over the whole corpus: (1) GitGalaxy's class_start regex vs. an independent JOB-card grep (^//[A-Za-z0-9@#$][A-Za-z0-9@#$]{0,7}[ \\t]+JOB([ \\t]|$), no shared implementation, independently rejects //*-commented job cards) -- exact agreement, 118 == 118, zero per-file mismatches. The 12 cics-banking-sample-application-cbsa/CREL0*.jcl + CREDB2L.jcl files whose entire JOB card is commented out (//*NAME JOB ...) and the 2 cics-genapp sampcma/sampwui.jcl angle-bracket placeholders (//<CMASAPPL> JOB) remain correctly NOT counted. (2) real pipeline DB output: struct_class_start == class_count == 118 on every file, class_data carries 118 rows (JCL is in _CLASS_START_NAMED_EXTRACTION_LANGS since #2203, no brace-slice fallback). (3) direct source read of the v1.2.0 delta: cics-java-jcics-samples/etc_VSAM_DEFVSAM.jcl contributes exactly one JOB card (//DEFVSAM JOB). Every corpus file carries at most one JOB card. Count moved 117 -> 118 on 2026-08-30 solely from that one new file (crucible v1.1.0 -> v1.2.0, gitgalaxy pin bump #2478). Zero false positives. See docs/language_status/jcl.md #9.",
+        "total": 118,
+        "verified": 118,
+        "verified_at": "2026-08-30",
         "verified_by": "Claude Sonnet 5"
       },
       "function": {
-        "note": "func_start (JCL EXEC steps) verified against the full language-crucible v1.1.0 corpus (185 .jcl files, 5 repos). Two independent checks over all 184 pipeline-scanned files: (1) GitGalaxy's func_start regex applied directly to raw source text, cross-checked line-for-line against an independent grep (^//[A-Za-z0-9@#$]{0,8}[ \\t]+EXEC([ \\t]|$), which shares no implementation with the primary regex and independently excludes //* comment lines) -- exact agreement, 375 == 375, zero per-file mismatches. (2) real pipeline DB output (galaxyscope --db-only): file_data.struct_func_start == function_count == 375 on every file, so neither prism.py's comment splitter nor detector.py's Mode A (_slice_by_labels) body slicing drops or corrupts any match before extraction. Spot-read against direct source for EXEC steps across itpll.jcl, CBL0001J.jcl, RUNASAM1.jcl including unnamed '// EXEC CICS,MEMBER=' statements (correctly matched with an empty step name). Zero false positives: every match is a real EXEC statement. Count moved 370 -> 375 when #2415 (aperture.py infra_path_pattern matched 'gen'/'lib' as filename substrings, dropping MAPGEN.jcl and DBRMLIB.jcl) was fixed in the same PR -- raw-corpus and pipeline now agree at 375/375. See docs/language_status/jcl.md #9.",
-        "total": 375,
-        "verified": 375,
-        "verified_at": "2026-08-29",
+        "note": "func_start (JCL EXEC steps) verified against the language-crucible v1.2.0 corpus (186 .jcl files, 6 repos; 185 pipeline-scanned -- cics-banking-sample-application-cbsa/DFH$SIP1.jcl is a bare SIP parameter deck with no // statements and is correctly aperture-excluded at zero-density). Three checks over the whole corpus: (1) GitGalaxy's func_start regex applied directly to raw source text, cross-checked line-for-line against an independent grep (^//[A-Za-z0-9@#$]{0,8}[ \\t]+EXEC([ \\t]|$), which shares no implementation with the primary regex and independently excludes //* comment lines) -- exact agreement, 376 == 376, zero per-file mismatches. (2) real pipeline DB output (galaxyscope --db-only): file_data.struct_func_start == function_count == 376 on every file, and function_data carries 376 rows, so neither prism.py's comment splitter nor detector.py's Mode A (_slice_by_labels) body slicing drops or corrupts any match. (3) direct source read of the v1.2.0 delta: cics-java-jcics-samples/etc_VSAM_DEFVSAM.jcl contributes exactly one EXEC step (//DEFINE EXEC PGM=IDCAMS); its instream IDCAMS SYSIN control statements (DELETE / DEFINE CLUSTER) are correctly NOT matched. Count moved 375 -> 376 on 2026-08-30 solely from that one new file (crucible v1.1.0 -> v1.2.0, gitgalaxy pin bump #2478). Zero false positives. See docs/language_status/jcl.md #9.",
+        "total": 376,
+        "verified": 376,
+        "verified_at": "2026-08-30",
         "verified_by": "Claude Sonnet 5"
       }
     },