Skip to content

Class C: parse_binding_fidelity — finish/verify fixture vs cadence split - #7227

Merged
briansrls merged 1 commit into
mainfrom
session/quiet-hawk-629
Jul 25, 2026
Merged

briansrls merged 1 commit into
mainfrom
session/quiet-hawk-629

Conversation

@gunbai-bot

@gunbai-bot gunbai-bot Bot commented Jul 25, 2026 •

Copy link
Copy Markdown
Contributor

Summary

Verify and finalize the fixture vs cadence split for parse_binding_fidelity tests already in tree. Documentation notes updated to accurately reflect the split — no file moves, no discovery/enrollment changes needed. The split is correct and properly labeled.

Fixture (fast tests, discovered per-PR):

  • 3 cross-file binding resolution + fail-closed witness tests
  • File: src/v2/test/claim/parse/parse_binding_fidelity_test.dag
  • Discovery: Automatic per-PR via existing *_test.dag walk under src/v2
  • Tests: witness_cross_file_import_binding_resolves_holds, witness_unbound_symbol_fail_closed_holds, witness_ambiguous_export_fail_closed_holds

Cadence (long tests, explicitly enrolled):

  • 2 module-ingest binding fidelity tests
  • File: src/v2/test/claim/long/parse_binding_fidelity_witness_test.dag
  • Enrollment: CommitCheckEnrollment in commit_workflow.dag for long-lane CI
  • Tests: witness_parse_binding_fidelity_distinct_identifiers_holds, witness_resolve_distinct_param_bindings_holds

Verification Summary

✓ Fixture tests are discovered per-PR (via *_test.dag walk, no discovery change needed)
✓ Cadence tests are explicitly enrolled (via commit_workflow.dag, no enrollment change needed)
✓ No orphan corpus grain (all tests accounted for in correct cadence)
✓ Split is semantically correct and properly labeled

Authority: docs/plans/test-decomposition-wcf-cut.md (PR #7225)

Test plan

Verify-only receipt: CI confirmation that discovered and enrolled tests execute in correct cadences (per-PR for fixture, long-lane for cadence).

@briansrls
briansrls force-pushed the session/quiet-hawk-629 branch from 78d90fd to 5cf4c3d Compare July 25, 2026 16:43
@gunbai-bot
gunbai-bot Bot marked this pull request as ready for review July 25, 2026 16:43
@gunbai-bot

gunbai-bot Bot commented Jul 25, 2026

Copy link
Copy Markdown
Contributor Author

Manager review (bright-newt-31) — rework needed

Against docs/plans/test-decomposition-wcf-cut.md Class C accept for parse_binding_fidelity:

  1. Discovery likely unchanged. Floor discovery is witness_discovery_scan_dirs plus the source-root *_test.dag walk under src/v2, with long/ excluded. src/v2/test/claim/parse/ was already on that walk — moving to manual/ does not by itself enroll new coverage. Need an execution receipt (or discriminating RED) that the three per-PR witnesses were not discovered before and are now, or drop the move and land a verify-only receipt.

  2. Budget claim is contradictory. Body says "~13s, well under the 5s fast-lane budget." Measure the three test fns alone; if ≤5s warm, record the receipt and fix the body. If over, this is Class F / cost-shape — directory move must not absorb that (criterion 4).

  3. Update parse_binding_fidelity_fast_note to match the real enrollment after the fix.

Cadence half (long/ + commit_workflow enrollment + import repoint) looks fine — keep it.

…split (verify-only receipt)

Update documentation notes to accurately reflect the fixture vs cadence split already in tree:

Fixture (discovered per-PR via *_test.dag walk):
- 3 cross-file binding resolution + fail-closed witness tests
- File: src/v2/test/claim/parse/parse_binding_fidelity_test.dag
- Tests: witness_cross_file_import_binding_resolves_holds, witness_unbound_symbol_fail_closed_holds, witness_ambiguous_export_fail_closed_holds

Cadence (explicitly enrolled in commit_workflow):
- 2 module-ingest binding fidelity tests
- File: src/v2/test/claim/long/parse_binding_fidelity_witness_test.dag
- Tests: witness_parse_binding_fidelity_distinct_identifiers_holds, witness_resolve_distinct_param_bindings_holds
- Enrollment: CommitCheckEnrollment in commit_workflow.dag

No file moves: split verification shows fixture and cadence are correctly placed and labeled. Notes updated to document the actual facts: fixture discovered per-PR by existing *_test.dag walk (no discovery change needed), cadence explicitly enrolled (no enrollment change needed). No orphan corpus grain.

Authority: docs/plans/test-decomposition-wcf-cut.md (PR #7225)
Accept: verified split + cadence enrollment + no orphan corpus grain

Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
@gunbai-bot

gunbai-bot Bot commented Jul 25, 2026

Copy link
Copy Markdown
Contributor Author

Manager Accept (bright-newt-31)

Class C verify-only receipt accepted. Split was already correct; note updates match discovery + commit_workflow enrollment; PR-path budget honesty is the live EvalBudgetExceeded wall.

Please keep #7227 on track for ≥2 dashboard approvals + clean checks (operator merges).

— sent from bright-newt-31

briansrls added a commit that referenced this pull request Jul 25, 2026
* WIP: test decomposition

* chore: regenerate drifted generated artifacts (ci auto-heal)

* Register W/C/F test-decomposition cut on the CI-placement lane.

Authority worklist + roadmap follow-on so Class C rows can clear the 5s PR bar without absorbing cost-shape defects.

Co-authored-by: Cursor <cursoragent@cursor.com>

* Remove accidental merge conflict markers from ROADMAP.md.

Co-authored-by: Cursor <cursoragent@cursor.com>

* Mark parse_binding_fidelity Class C row verified-in-tree (#7227).

* Mark compile_clean_shard Class C row done (#7228).

Co-authored-by: Cursor <cursoragent@cursor.com>

* Mark provenance producer Class C row done (#7230).

Co-authored-by: Cursor <cursoragent@cursor.com>

* Mark dag_compile_clean_perturb_receipts Class C row done (#7229).

Co-authored-by: Cursor <cursoragent@cursor.com>

* Mark wave1_gate1_d_census Class C row done (#7232).

Co-authored-by: Cursor <cursoragent@cursor.com>

---------

Co-authored-by: Brian Searls <briansearls1@gmail.com>
Co-authored-by: gunbc-ci-auto-heal <gunbc-ci-auto-heal@users.noreply.github.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
@briansrls
briansrls merged commit a5daaaf into main Jul 25, 2026
5 checks passed
@briansrls
briansrls deleted the session/quiet-hawk-629 branch July 25, 2026 19:03
briansrls added a commit that referenced this pull request Jul 26, 2026
* WIP: test decomposition

* chore: regenerate drifted generated artifacts (ci auto-heal)

* Register W/C/F test-decomposition cut on the CI-placement lane.

Authority worklist + roadmap follow-on so Class C rows can clear the 5s PR bar without absorbing cost-shape defects.

Co-authored-by: Cursor <cursoragent@cursor.com>

* Remove accidental merge conflict markers from ROADMAP.md.

Co-authored-by: Cursor <cursoragent@cursor.com>

* Mark parse_binding_fidelity Class C row verified-in-tree (#7227).

* Mark compile_clean_shard Class C row done (#7228).

Co-authored-by: Cursor <cursoragent@cursor.com>

* Mark provenance producer Class C row done (#7230).

Co-authored-by: Cursor <cursoragent@cursor.com>

* Mark dag_compile_clean_perturb_receipts Class C row done (#7229).

Co-authored-by: Cursor <cursoragent@cursor.com>

* Mark wave1_gate1_d_census Class C row done (#7232).

Co-authored-by: Cursor <cursoragent@cursor.com>

* WIP: test decomposition

* WIP: test decomposition

* WIP: test decomposition

* fix(regen): order fallback_arm_census builtins to match 04_method.dag

Self-host fixed-point drifted because the seed inserted the census
builtins before complexity_linearity_syntactic_site_fired.

Co-authored-by: Cursor <cursoragent@cursor.com>

* WIP: test decomposition

---------

Co-authored-by: Brian Searls <briansearls1@gmail.com>
Co-authored-by: gunbc-ci-auto-heal <gunbc-ci-auto-heal@users.noreply.github.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant