Class C: decompose wave1_gate1_d_census onto falsifier_substrate_long_lane

[briansrls]

Auto-opened by session-dashboard for session merry-ibex-306.
Pushing to session/merry-ibex-306 advances this PR.

Summary

Class C decomposition for wave1_gate1_d_census (docs/plans/test-decomposition-wcf-cut.md): the row's cost is genuine corpus breadth (10 ingested fixtures through a real tokenize→parse→normalize→resolve pass each — 134526ms measured witness wall vs 32034ms one-time resolve), not a fixture cost-shape defect, so it splits rather than reshapes.

• Per-PR: the mechanism-only sibling (self_host/wave1_gate1_d_mvp_subsumption_witness_test.dag, 5 table_fixture_* check_fns, measured eval_self=0.000ms) is already covered by the source-root *_test.dag discovery walk — no explicit CommitCheckEnrollment needed, so the redundant explicit roster row for it was deleted rather than added.
• Corpus grain: wave1_gate1_d_census_subsumption_witness_holds (long/wave1_gate1_d_mvp_subsumption_witness_test.dag) is re-homed onto the falsifier's 4-hourly cadence as one more row on the existing falsifier_substrate_long_lane_entries (landed by PR Class C: provenance producer — verify fixture walls per-PR + live grain nightly enrollment #7230, session/proud-bat-741-provenance-class-c) — no new batch/type minted. A WitnessExclusionRow (pattern long/wave1_gate1_d_mvp_subsumption_witness_test.dag, classification FalsifierSubstrateLongLane) carves the file out of per-PR discovery so the routing is honestly classified, not defaulted.
• Branch merges PR Class C: provenance producer — verify fixture walls per-PR + live grain nightly enrollment #7230 in directly (not yet on main) to reuse its landed FalsifierSubstrateLongLane cadence/exclusion/admission machinery rather than duplicating it (DESIGN §3).

Test plan (green-by-execution, claim_batch --source-root dag --source-root src/v2)

• src/v2/test/claim/ci_floor_plan_witness_test.dag ci_floor_plan_witnesses, falsifier_plan_structure_holds — PASS (batch-6 enrollment + expected_batches count correct with the census row folded into the existing lane)
• src/v2/test/claim/witness_admission_test.dag witness_admission_invariant_holds — PASS
• src/v2/test/claim/witness_exclusion_reconciliation_test.dag witness_exclusion_frontier_reconciliation_holds, witness_exclusion_explicit_roster_rows_consistent_holds, witness_exclusion_red_control_unbacked_substrate_long_lane_row_reds — PASS (including the discriminating RED control)

Worker attestation

• Title describes the change (not the session id or branch).
• PR body summarises what and why.
• Tests run: see Test plan above.
• No commits on this branch are surprises.
• No secrets / credentials / large binaries staged.

[gunbai-bot]

Addressed review 42719 (both cited defects — the missing import and the expected_batches gap) by converging onto PR #7230's already-landed FalsifierSubstrateLongLane mechanism via a real merge, rather than patching my duplicate mint in place: the import is now inherited correctly from #7230, and expected_batches/falsifier_plan_structure_holds was already correct in #7230's landed code (verified green-by-execution on the merged commit). Review 42719 was posted against the pre-fix sha 8099d21; current head is 6d99f10.

— sent from merry-ibex-306