Dedup ecmascript/python/typescript catalog fn families (per-file parameterized cores)

[briansrls]

Summary

Mechanical per-file dedup of stamped catalog/primitive-fn families in the three cited upstream language models. Each file gains parameterized cores; every importer-visible name stays as a thin wrapper. No cross-language shared module (per operator lane brief).

Size accounting (wc -l, main → branch)

File	Before	After	Δ
ecmascript.dag	1489	1337	−152
python.dag	1669	1557	−112
typescript.dag	3684	3658	−26
total	6842	6552	−290

Raw diff: 467 insertions / 757 deletions across the three files.

Family → per-file core

ecmascript.dag

• ecmascript_primitive_facts_node ← 5× ecmascript_*_facts_node
• ecmascript_primitive_spec_facts ← 5× ecmascript_*_spec_facts
• ecmascript_primitive_algebra_witness_node ← 3× ecmascript_*_algebra_witness_node
• ecmascript_catalog_surface_symbol_in_rows ← 6× per-catalog *_catalog_surface_symbol (+ dissolved *_catalog_surface_step, *_catalog_surface_spelling_eq, *_catalog_surface_fold_false)
• ecmascript_primitive_fact_bundles_from_catalog_rows ← 6× ecmascript_primitive_bundles_from_*_catalog

python.dag

• python_primitive_facts_node ← 5× python_*_facts_node
• python_primitive_algebra_witness_node ← 3× python_*_algebra_witness_node
• python_g1_2_pair_axis_fact_bundle_entries ← 5× python_g1_2_*_fact_bundle_entries
• python_catalog_surface_symbol_in_rows ← 5× per-catalog *_catalog_surface_symbol (+ dissolved step/spelling_eq/fold_false helpers)
• python_per_language_fact_bundle_entries_from_catalog_rows ← 5× python_g1_2_*_catalog_entries

typescript.dag

• ts_primitive_algebra_witness_node ← 4× ts_*_algebra_witness_node (number/bigint/bool/string)
• ts_catalog_surface_symbol ← inlined fold (dissolved ts_catalog_surface_symbol_step)
• ts_numeric_primitive_fact_bundles ← inlined fold (dissolved ts_primitive_bundles_from_catalog_step)

Dropped names are internal step/spelling helpers only; all importer-facing fn names grep-clean.

Test plan

Executed locally (CTRL_BUILD_BYPASS_SHIMS=1, release build):

claim_batch --source-root src/v2 --entry src/v2/test/claim/manual/python_wave1_grammar_claim.dag --function python_wave1_grammar_validates_fixture
→ PASS (resolve 2638–3147ms; witness 44–48ms)

claim_batch --source-root src/v2 --entry src/v2/test/claim/manual/mvp_int_cross_target_coercion_test.dag --function witness_cross_target_rust_i32_to_python_int_accepts
→ PASS (resolve 5861ms; witness 60ms)

claim_batch --source-root src/v2 --entry src/v2/test/claim/manual/typescript_language_model_anchor.dag --function ts_anchor_number_algebra_inhabitance
→ resolve PASS (2680ms; 33 modules / 1655 closure items) — entry is data, not Bool witness

All three language modules resolve clean in witness closures touching ecmascript.dag, python.dag, and typescript.dag. Zero .rs changes.

[gunbai-bot]

CI investigation (6a0d392) — no PR fix warranted

Checked run 28775068966:

Job	Result	Notes
rust_tests	pass (42m31s)	v1 rust gate green
ci	fail	gunbc ci (.dag witnesses + gates) timed out after 135 minutes

Failure line: The action 'gunbc ci (.dag witnesses + gates)' has timed out after 135 minutes.

Last gantt marks before timeout show full-corpus compile still in progress (compile.reconcile.done at ~52 min wall; job killed at 135m cap). No witness/gate failure attributable to this PR's three-file dedup diff — the floor never finished compiling.

Matches systemic compile-wall called out by manager (same timeout on main/#6311/#6313). No push planned — another push would reset reviews and start another doomed 135m run.

PR substance unchanged: green base (#6311 merge), −290 lines, claude+cursor APPROVE on HEAD.

— sent from valiant-ant-848