Repository navigation
[codex] Ratchet call pattern lookup authority - #3106
Conversation
|
Verified the P5 caveat against the live PR body. It is already addressed by the |
# Conflicts: # docs/r3-program-plan.md
briansrls
left a comment
There was a problem hiding this comment.
Review metadata
- Provider / model:
codex/unknown - Commit:
995e4333· Trigger:schedule - Thinking:
380s wall
BLOCKING (1)
Root Cause
docs/r3-program-plan.mdGate #77 records the L-7 consumer assertion but not the SG-0 debt receipt → add one checkable receipt here or in the PR body: deleted scaffold path, SG-0 census shrink before/after, or explicit deferral naming the lane and concrete ROADMAP row.
ROADMAP — Incomplete
- e_p_call_pattern_lookup_authoritative: The authority assertion is verified against cost/complexity .dag and generated consumers, but the P5 hand-Rust receipt is incomplete.
| | 77 | `e_p_call_pattern_lookup_authoritative` | substrate-shape | T-E-P-Producer-Broadening | **CONSUMER_LANDED** — integration `e_p_call_pattern_lookup_authoritative` (`m2_substrate_inhabitance_test.rs`) pins `per_call_pattern_at` as the lens-facing query over the per-call evidence authority | CallPattern lookup authoritative | | ||
| | 76 | `e_p_per_call_descent_evidence_full_coverage` | substrate-shape | T-E-P-Producer-Broadening | **CONSUMER_LANDED** (PR #2147 carrier + PR #2190 consumer; refresh per cluster-analysis audit §1) | per-call DescentEvidence covers all live call sites | | ||
| | 77 | `e_p_call_pattern_lookup_authoritative` | substrate-shape | T-E-P-Producer-Broadening | **PASSING** — integrations `e_p_call_pattern_lookup_authoritative` + `e_p_call_pattern_lookup_authoritative_consumers_use_query_surface` (`m2_substrate_inhabitance_test.rs`) pin `per_call_pattern_at` as the lens-facing query over the per-call evidence authority and ratchet cost/complexity `.dag` + generated consumers against direct side-table or `lower_call_pattern` routing | CallPattern lookup authoritative | | ||
| | 78 | `e_p_sub_value_relation_per_call_landed` | substrate-shape | T-E-P-Producer-Broadening | **CONSUMER_LANDED** (T-E-P P1 slices 1-7 per cluster-analysis audit §1) | SubValueRelation per-call landing | |
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
|
Verified the optional P5 alignment note against the current branch. The gate #77 row now names the concrete ROADMAP authority: |
briansrls
left a comment
There was a problem hiding this comment.
Review metadata
- Provider / model:
codex/unknown - Commit:
edc87b96· Trigger:schedule - Thinking:
151s wall
Non-blocking — Strengths
src/v3/compiler/tests/integration/m2_substrate_inhabitance_test.rsThe added test stays implementation-layer and ratchets both.daglens sources and generated consumers against bypassingper_call_pattern_at.
ROADMAP — Verified
- e_p_call_pattern_lookup_authoritative: The authority assertion is now backed by the consumer-surface ratchet and the P5 hand-Rust deferral is complete.
✅ No blocking concerns.
|
Review metadata
1. Story of the diffThis PR turns gate #77 from “consumer landed” into “passing” by adding a ratchet that checks the CallPattern lookup authority is now the single lens-facing surface. The load-bearing mechanism is a new integration test, 2. Invariant categories
N/A — the diff edits a planning row and an integration test only; no
Compliant — P2 / L-7 single-authority is the point of the added ratchet: authored lenses are required to consume
Compliant — the new code stays as data plus a free test function: file contents are bound to local names, then checked through two explicit
Compliant — the test is behavior-shaped around one contract, “CallPattern consumers use the authoritative query surface,” and it avoids a full compile pipeline; it directly checks the relevant authored and generated consumer surfaces (
N/A — no locked design document is edited, and the diff does not change the meaning of a locked substrate or bootstrap decision.
Compliant — the added hand-Rust test surface is explicitly bounded as test-only SG-0 ratchet surface, says it introduces no substrate authority or runtime path, and names the dissolution trigger: porting this ratchet to 2.5. Top-down PM intent reviewCompliant — the PM-level intent here is to remove duplicate lookup authority, not to add a new model. The diff preserves that intent: the new ratchet requires the cost and complexity consumers to route through 3. VerdictAPPROVE. The PR is narrow and aligns the evidence with the gate promotion: one canonical query surface, direct-consumer routes rejected, and the temporary Rust receipt is bounded with a named dissolution path. I did not find a diff-cited invariant violation. |
Summary
.daglenses and their generated Rust consumers route CallPattern facts throughper_call_pattern_at.per_call_descent_evidenceside-table consumption and compiler-internallower_call_patternrouting in those consumer surfaces.e_p_call_pattern_lookup_authoritativeto PASSING with the new consumer-routing receipt.P5 receipt
This PR adds test-only hand-Rust under
src/v3/compiler/tests/integration/m2_substrate_inhabitance_test.rs. The new Rust does not introduce a substrate authority, producer, mirror, or runtime behavior path. It is a regression ratchet that checks the existing.daglens sources and generated Rust continue to route through the declaredper_call_pattern_atsubstrate accessor, and fail if consumers bypass it viaper_call_descent_evidenceorlower_call_pattern.Validation
cargo test -p v3-compiler --test integration e_p_call_pattern_lookup_authoritative -- --nocapturecargo fmt --all --checkNote: an initial focused test run used the wrong Cargo test target name (
m2_substrate_inhabitance_test) and failed before running tests; rerun above used the correctintegrationtarget and passed.