Menu

Search

Documentation Index

Theme: Symblix Light

Chapter 11 — gates transcripts (captured)

Last updated:

# Chapter 11 — gates transcripts (captured)

Real output from the verification machine (same environment as the earlier transcripts). The Chapter 11 overlay (ch11-gates) adds: the critic’s conformance runners + fixtures (testsuite/conformance/), the regenerated paradigm_matrix.yaml, the streamed-context matrix entry, the legacy paradigm parity vectors (named/euclo/surface/paradigm_test.go), the coverage override for the runner main (tooling/coverage/overrides.yaml), and the two embedded-template sources (the recipe and the prompt — ADR-002). It also carries the D-2 fix to the agent itself (cognitionzoo/critic/critic.go — the critique phase now renders the compiled slice, mirroring reflection).

# 1. The gauntlet, all green 

$ ./verify.sh gates ch11        # build, vet, 34 critic tests, then:
lint-arch:                  [PASS]
check-gates-honest:         [PASS]
domain-check:               [PASS]
grep-architecture-gates:    [PASS]
test-euclo-golden:          [PASS]
test-conformance:           [PASS]

$ make test-coverage
  [PASS] codeburg.org/lexbit/relurpify/cognitionzoo/critic           78.28% (floor  40; 209/267 stmts)
  [PASS] codeburg.org/lexbit/relurpify/cognitionzoo/critic/cmd/critic-runner   0.00% (floor   0; 0/54 stmts)
[PASS] coverage: all 151 packages at or above their floors (default floor 40%)

# 2. The failures the gauntlet taught (all real, all fixed honestly)

# Conformance: declared cases with no runners

The harness is registry-driven; the critic’s four declared cases failed on arrival:

--- FAIL: TestParadigmContractConformance
    conformance case "critic/critique_writes_scratch" has no runner
    (contract declared, runtime drift)

Fixed by the runners file (critic_conformance_test.go, registering four runners that mirror reflection’s line for line) plus two fixtures (critic_critique.erpe, critic_revise.erpe) and — after a real failure — a third (critic_pass.erpe, no capture line): the pass-case first failed because the shared fixture’s capture revisions bound the result field revisions: 0 (present). Reflection ships separate pass/revise fixtures for exactly this reason; the mirror now does too.

# The matrix, generated not hand-written 

--- FAIL: TestParadigmMatrixUpToDate
    paradigm matrix drifted from the contract registry:
    (set UPDATE_MATRIX=1 to regenerate)
--- FAIL: TestParadigmMatrixInvariants
    declared case critic/critic/critique_writes_scratch has no status annotation

UPDATE_MATRIX=1 regenerates the committed matrix from the registry (same discipline as the goldens); the status annotations (implemented) seed the generator’s table. The regenerated matrix’s critic rows:

    - paradigm: critic
      directive: critique
      status: implemented
      cases:
        - critic/critique_writes_scratch
        - critic/critique_pass_ends_loop
    - paradigm: critic
      directive: revise
      status: implemented
      cases:
        - critic/revise_fires_on_predicate
        - critic/revision_cap_bounded

# The stream matrix caught a real integration gap 

--- FAIL: TestParadigmStreamedContextMatrix
    paradigm "critic" is registered but has no streamed-context matrix
    runner (paradigm_not_integrated)

Running the sentinel proved the deeper fact: the Chapter 8 critic never rendered the compiled slice into its critique prompt — reflection does (paradigm.StreamedSection, the D-2 seam). The fix is in the agent, not the test: runCritique now renders the streamed section into its system slot exactly as reflection’s review does, and the matrix runner (streamMatrixCritic) pins the sentinel. All 34 package tests still pass with the integration.

# The surface parity vectors 

--- FAIL: TestAllParadigms_MatchesLegacyValidateList
    AllParadigms() returned 10 paradigms; want 9
--- FAIL: TestAgentParadigms_MatchesLegacyAgentList
    AgentParadigms() returned 9 paradigms; want 8

The surface vocabulary (Chapter 9’s ParadigmCritic) grew the canonical lists; the legacy parity vectors grow with them — critic joins both, deliberately, in the forked test file.

# The config-tree drift gate (ADR-002) 

--- FAIL: TestDriftGatePassesWhenInSync
    Only in relurpify_cfg/euclo: critic_review.erpe
    Only in relurpify_cfg/prompts/agents: agent.critic.default.prompt

The checked-in relurpify_cfg/ must be the materialization of the embedded template bundle. Fix per ADR-002: add the recipe and prompt as template sources (userconfig/templates/embedfs/…) and regenerate:

$ make generate-config
$ make check-config-tree-drift
[PASS] relurpify_cfg/ matches embedded templates

git status relurpify_cfg/ after regeneration shows exactly the two new files — nothing else drifted.

# The coverage floor (D9, verbatim)

The runner main arrived at 0.0%:

	codeburg.org/lexbit/relurpify/cognitionzoo/critic/cmd/critic-runner
		coverage: 0.0% of statements

The house pattern for mains is an explicit floor entry with a reason (floor: 0; reason: …; ratchets up as tests land), so the overlay’s overrides.yaml fork carries one for critic-runner. The package itself — written test-first since Chapter 2 — lands at 78.28% against the enforced floor of 40 (the Makefile’s global default; overrides.yaml is the per-package ratchet) and above the project’s 70% aspiration.

On this page