15 — Regression Tests v2 (Phase 16)

Object

v_rp_generator_parity_regression_v2 — 6 tests, all PASS, with teeth (tests that can actually fail were demonstrated to fail against the v1 generator).

Results

• RGT1 — AX-PROCESS membership complete in generated_v2 (22 == 22): PASS.
• RGT2 — TEETH: the v1 generator STILL misses the 3 verified workflows (gen v1 AX-PROCESS = 19, gap vs current = 3): PASS (the rule can fail; v1 demonstrates the failure).
• RGT3 — AX-PXT residual_reconcile corrected to live 2 in generated_v2 (gen2 = 2, current = 2): PASS.
• RGT4 — TEETH: the v1 generator STILL emits stale residual 8 (gen v1 = 8): PASS (the rule can fail; v1 demonstrates it).
• RGT5 — anti-false-green remains 6/6: PASS.
• RGT6 — v1 generated sentinel RETAINED (not dropped): PASS.

Coverage of the macro's required tests

• generator missing verified workflow fails: RGT2 (teeth via v1).
• generator residual 8 vs 2 fails: RGT4 (teeth via v1).
• generated parity must fail if any node missing: RGT2 shows the missing-node case still fails; validation v2 (doc 07) would report a gap if any node were missing.
• UI deploy guard fails if current alias points to stale v1: covered by the anti-false-green suite (T1 stale-v1 rejected) referenced by RGT5, and the deploy guard's stale_v1_binding_count gate (currently 0).
• anti-false-green remains pass: RGT5.
• v1 sentinel remains retained: RGT6.