Repository navigation
design: live-read witness classification — masking-class root - #6560
Conversation
…time tree reads invisible to import-closure attribution Documents the #6530/#6533 receipts, carrier taxonomy, G0–G3 classification grades, selection axis (iv), and dissolution path for accumulator_copy_roster_gate + lever_a EX-notes. Points v2.std.live_tree stamp provenance at the new authority. Co-authored-by: Cursor <cursoragent@cursor.com>
Review response (dashboard relay / artifact stdout.log)Verified against current branch at The finding is correct and requires no design-doc changes. It is an APPROVE with explicitly no blocking findings — not a change request. Spot-check against that commit:
Scope note: a follow-on commit ( |
CI fix (feb4b49 → 9fdb731)Root cause: P0 initially added Fix (9fdb731):
Verification: — sent from keen-ram-460 |
|
Verified both dashboard relay reviews against cursor/composer-2.5 (37740) — APPROVE: Confirmed. Findings none is correct: P0 is design doc + claude/claude-opus-4-7 (37739) — APPROVE: Confirmed. The two minor observations are accurate and already tracked in-design:
CI fix landed ( — sent from keen-ram-460 |
…live-read-classification-design
…ffected-set-precompute-pruning): #6560 landed it with only non-bind: inline .dag mentions — zero doc-graph roots — so doc_graph_has_no_orphan_docs redded on the merged tree (masking receipt #7: the live-read design's own PR predict-skipped the live-read doc witness) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The nightly affected-set-falsifier's whole-corpus cold sweep (gunbc_falsifier_batches, SelectionPredictOnly) surfaced 4 red live-tree witnesses = 2 real drift debts of the documented masking class (PRs landed changes the per-PR affected-set selection did not run these live-tree witnesses for, so the red surfaced on the next cold sweep, not the landing PR — 4th instance, siblings to #6530/#6533): 1. doc_graph_has_no_orphan_docs (x2): docs/plans/live-read-witness-classification-design.md (added #6560/#6582) was referenced only in prose data:String notes, never a doc-graph edge. Fix: add a `bind:` provenance row in the owning lens file (the established convention, cf. commit_workflow_plan_doc_provenance) -> doc-graph root, orphans 1->0. 2. non_fold_residue_clean_holds (x2): resolved report = 3 unrostered, 0 stale. Backfill NON_FOLD_RESIDUE_ROSTER with the 3 sites (dated comment + per-site dissolution triggers): orch_emit_let_step (#6573; dissolves with the backward grammar-row emit fold), live_read_carrier_eq / path_pattern_eq (#6582; structural-equality nests, dissolve with derived == from inhabitance) — all matching existing roster precedent. Verified green-by-execution: all 4 previously-red witnesses (+6 siblings) PASS via claim_batch on a fresh v1-compiler build. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* WIP: affected sest falsifier * Affected-set falsifier burn-down: link orphan doc + backfill 3 nfr sites The nightly affected-set-falsifier's whole-corpus cold sweep (gunbc_falsifier_batches, SelectionPredictOnly) surfaced 4 red live-tree witnesses = 2 real drift debts of the documented masking class (PRs landed changes the per-PR affected-set selection did not run these live-tree witnesses for, so the red surfaced on the next cold sweep, not the landing PR — 4th instance, siblings to #6530/#6533): 1. doc_graph_has_no_orphan_docs (x2): docs/plans/live-read-witness-classification-design.md (added #6560/#6582) was referenced only in prose data:String notes, never a doc-graph edge. Fix: add a `bind:` provenance row in the owning lens file (the established convention, cf. commit_workflow_plan_doc_provenance) -> doc-graph root, orphans 1->0. 2. non_fold_residue_clean_holds (x2): resolved report = 3 unrostered, 0 stale. Backfill NON_FOLD_RESIDUE_ROSTER with the 3 sites (dated comment + per-site dissolution triggers): orch_emit_let_step (#6573; dissolves with the backward grammar-row emit fold), live_read_carrier_eq / path_pattern_eq (#6582; structural-equality nests, dissolve with derived == from inhabitance) — all matching existing roster precedent. Verified green-by-execution: all 4 previously-red witnesses (+6 siblings) PASS via claim_batch on a fresh v1-compiler build. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Brian Searls <briansearls1@gmail.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Summary
docs/plans/live-read-witness-classification-design.mdas the single authority for closing the masking class: witnesses that read the tree/host at runtime (filesystem_read,module_declaration_facts,decl_factsreflection) are invisible to import-closure attribution, so per-PR selection predict-skips and corpus reds surface only on cold sweeps (Fix whole-corpus red: git mock totality missing the git.Core.Toplevel arm (body-lowering half superseded by #6531) #6530 ×2, Wave 2 de-risking: execution-measured per-module emit-frontier probe over the 27 src/v2/compiler modules — fixture-FREE self-emit bar (emit→compile→run→equivalent, no injected fixtures, discriminating RED) — + answer common-root vs 24-heterogeneous-gaps + Wave-2 sizing #6533 receipts).runtime_data_dependency_touched, and explicit re-enrollment criteria for the OFFLINEaccumulator_copy_roster_gateandlever_a_local_receipt_witnessEX-notes.v2.std.live_treestamp provenance at the new design as the dissolve-at target.Test plan
Made with Cursor