final_report_validator.py

@@ -20,6 +20,7 @@ from review_proofs import (
     assess_review_mutation_final_report,
     assess_validation_report,
 )
+from validation_status_vocabulary import assess_validation_status_vocabulary
 
 FINAL_REPORT_TASK_KINDS = frozenset({
     "review_pr",
@@ -598,6 +599,34 @@ def _rule_reviewer_main_checkout_baseline(report_text: str) -> list[dict[str, st
     ]
 
 
+def _rule_reviewer_validation_status_vocabulary(
+    report_text: str,
+    *,
+    action_log: list[dict] | None = None,
+) -> list[dict[str, str]]:
+    text = report_text or ""
+    if not re.search(
+        r"validation status|pr-head validation status|official validation status",
+        text,
+        re.IGNORECASE,
+    ):
+        return []
+    result = assess_validation_status_vocabulary(
+        text,
+        command_log=action_log,
+    )
+    if not result.get("block"):
+        return []
+    return _findings_from_reasons(
+        "reviewer.validation_status_vocabulary",
+        result.get("reasons") or [],
+        field="Validation status",
+        severity="block",
+        safe_next_action=result.get("safe_next_action")
+        or "use a validation status that matches the proof path executed",
+    )
+
+
 def _rule_reviewer_main_checkout_path(report_text: str) -> list[dict[str, str]]:
     text = report_text or ""
     if "baseline worktree path" not in text.lower():
@@ -902,6 +931,7 @@ _RULES_BY_TASK: dict[str, list[Callable[..., list[dict[str, str]]]]] = {
         _rule_reviewer_validation_structured,
         _rule_reviewer_linked_issue,
         _rule_reviewer_baseline_on_failure,
+        _rule_reviewer_validation_status_vocabulary,
         _rule_reviewer_main_checkout_baseline,
         _rule_reviewer_main_checkout_path,
         _rule_reviewer_already_landed_eligible,

skills/llm-project-workflow/workflows/review-merge-pr.md

@@ -568,6 +568,28 @@ If the cause is unknown, do not erase the earlier failure with plain
 `gitea_validate_review_final_report` rejects reports that omit known earlier
 validation failures when `validation_session.observed_failures` is supplied.
 
+## 21B. Validation status taxonomy (#406)
+
+When the final report summarizes how validation concluded, use one of these
+**validation status** labels (distinct from per-command pass/fail entries):
+
+* `passed` — raw PR-head validation passed on the unmodified head.
+* `failed` — raw PR-head validation failed and no allowed resolution path
+  was proven.
+* `baseline-equivalent failure accepted` — only when a clean baseline
+  worktree under `branches/` proves matching failure signatures on the target
+  branch (baseline path, target SHA, exact commands, failure lists, and
+  `failure signatures match: true`).
+* `raw-head failure resolved by merge simulation` — raw PR-head validation
+  failed, but merge simulation into the current target passed cleanly; report
+  merge simulation under `Worktree/index mutations` with full #317 proof.
+* `passed after transient failure investigation` — a later run passed after an
+  earlier failure in the same session; document the failure history (#396).
+
+Do not use `baseline-equivalent failure accepted` when only merge simulation
+resolved the failure. Do not use bare `passed` when raw PR-head validation
+failed unless one of the resolution statuses above applies.
+
 ## 22. Baseline validation rule
 
 Do not run tests in the main checkout.

tests/test_validation_status_vocabulary.py

@@ -0,0 +1,198 @@
+"""Tests for validation status vocabulary (#406)."""
+import sys
+import unittest
+from pathlib import Path
+
+sys.path.insert(0, str(Path(__file__).resolve().parent.parent))
+
+from final_report_validator import assess_final_report_validator  # noqa: E402
+from validation_status_vocabulary import (  # noqa: E402
+    STATUS_BASELINE_EQUIVALENT,
+    STATUS_FAILED,
+    STATUS_MERGE_SIM_RESOLVED,
+    STATUS_PASSED,
+    STATUS_TRANSIENT_PASS,
+    assess_validation_status_vocabulary,
+)
+
+
+def _handoff(**extra):
+    fields = {
+        "Task": "review PR #386",
+        "Validation status": STATUS_PASSED,
+        "Raw PR-head validation result": "passed",
+        "Merge simulation result": "not run",
+        "Baseline worktree used": "none",
+    }
+    fields.update(extra)
+    lines = ["## Controller Handoff", ""]
+    lines.extend(f"- {key}: {value}" for key, value in fields.items())
+    return "\n".join(lines)
+
+
+class TestValidationStatusVocabulary(unittest.TestCase):
+    def test_raw_head_pass_status(self):
+        report = _handoff()
+        result = assess_validation_status_vocabulary(report)
+        self.assertFalse(result["block"])
+        self.assertEqual(result["status_claimed"], STATUS_PASSED)
+
+    def test_raw_head_failure_with_baseline_match(self):
+        report = _handoff(
+            **{
+                "Validation status": STATUS_BASELINE_EQUIVALENT,
+                "Raw PR-head validation result": "failed",
+                "Baseline worktree used": "branches/baseline-master-pr386",
+                "Baseline target SHA": "a" * 40,
+                "Baseline failures": "test_foo failed",
+                "PR failures": "test_foo failed",
+                "Failure signatures match": "true",
+            }
+        )
+        result = assess_validation_status_vocabulary(report)
+        self.assertFalse(result["block"])
+        self.assertTrue(result["baseline_proof_complete"])
+
+    def test_baseline_equivalent_without_baseline_proof_blocked(self):
+        report = _handoff(
+            **{
+                "Validation status": STATUS_BASELINE_EQUIVALENT,
+                "Raw PR-head validation result": "failed",
+            }
+        )
+        result = assess_validation_status_vocabulary(report)
+        self.assertTrue(result["block"])
+        self.assertIn("baseline-equivalent", result["reasons"][0])
+
+    def test_merge_simulation_resolution_passes(self):
+        report = "\n".join([
+            _handoff(
+                **{
+                    "Validation status": STATUS_MERGE_SIM_RESOLVED,
+                    "Raw PR-head validation result": "failed",
+                    "Merge simulation result": "passed",
+                }
+            ),
+            "Worktree/index mutations: merge simulation in branches/review-pr386",
+            "Worktree path: branches/review-pr386",
+            "Pre-simulation clean status: clean",
+            "Merge result: clean merge",
+            "Abort command: git merge --abort",
+            "Post-abort clean status: clean",
+        ])
+        command_log = [
+            {"command": "git merge --no-commit prgs/master"},
+            {"command": "git merge --abort"},
+        ]
+        result = assess_validation_status_vocabulary(
+            report, command_log=command_log
+        )
+        self.assertFalse(result["block"])
+        self.assertTrue(result["merge_simulation_passed"])
+
+    def test_merge_simulation_failure_stays_failed(self):
+        report = _handoff(
+            **{
+                "Validation status": STATUS_FAILED,
+                "Raw PR-head validation result": "failed",
+                "Merge simulation result": "failed",
+            }
+        )
+        result = assess_validation_status_vocabulary(report)
+        self.assertFalse(result["block"])
+
+    def test_failed_status_with_passing_merge_sim_blocked(self):
+        report = "\n".join([
+            _handoff(
+                **{
+                    "Validation status": STATUS_FAILED,
+                    "Raw PR-head validation result": "failed",
+                    "Merge simulation result": "passed",
+                }
+            ),
+            "Worktree/index mutations: merge simulation",
+            "Worktree path: branches/review-pr386",
+            "Pre-simulation clean status: clean",
+            "Merge result: clean",
+            "Abort command: git merge --abort",
+            "Post-abort clean status: clean",
+        ])
+        result = assess_validation_status_vocabulary(
+            report,
+            command_log=[{"command": "git merge --no-commit prgs/master"}],
+        )
+        self.assertTrue(result["block"])
+
+    def test_transient_failure_then_pass(self):
+        report = _handoff(
+            **{
+                "Validation status": STATUS_TRANSIENT_PASS,
+                "Raw PR-head validation result": "passed",
+                "Transient validation failure history": (
+                    "first run failed with infra flake; rerun passed"
+                ),
+            }
+        )
+        result = assess_validation_status_vocabulary(report)
+        self.assertFalse(result["block"])
+
+    def test_transient_pass_without_history_blocked(self):
+        report = _handoff(
+            **{
+                "Validation status": STATUS_TRANSIENT_PASS,
+                "Raw PR-head validation result": "passed",
+            }
+        )
+        result = assess_validation_status_vocabulary(report)
+        self.assertTrue(result["block"])
+
+    def test_bare_passed_after_raw_failure_blocked(self):
+        report = _handoff(
+            **{
+                "Validation status": STATUS_PASSED,
+                "Raw PR-head validation result": "failed",
+            }
+        )
+        result = assess_validation_status_vocabulary(report)
+        self.assertTrue(result["block"])
+
+    def test_wrong_baseline_label_when_merge_sim_used_blocked(self):
+        report = "\n".join([
+            _handoff(
+                **{
+                    "Validation status": STATUS_BASELINE_EQUIVALENT,
+                    "Raw PR-head validation result": "failed",
+                    "Merge simulation result": "passed",
+                }
+            ),
+            "Worktree/index mutations: merge simulation",
+            "Worktree path: branches/review-pr386",
+            "Pre-simulation clean status: clean",
+            "Merge result: clean",
+            "Abort command: git merge --abort",
+            "Post-abort clean status: clean",
+        ])
+        result = assess_validation_status_vocabulary(
+            report,
+            command_log=[{"command": "git merge --no-commit prgs/master"}],
+        )
+        self.assertTrue(result["block"])
+        joined = " ".join(result["reasons"]).lower()
+        self.assertTrue(
+            "misleading" in joined or "baseline-equivalent" in joined
+        )
+
+    def test_final_report_validator_integration_blocks_misleading_label(self):
+        report = _handoff(
+            **{
+                "Validation status": STATUS_BASELINE_EQUIVALENT,
+                "Raw PR-head validation result": "failed",
+            }
+        )
+        result = assess_final_report_validator(report, "review_pr")
+        blocked_ids = {f["rule_id"] for f in result["findings"]}
+        self.assertIn("reviewer.validation_status_vocabulary", blocked_ids)
+
+
+if __name__ == "__main__":
+    unittest.main()

validation_status_vocabulary.py

@@ -0,0 +1,205 @@
+"""Precise validation-status vocabulary for reviewer final reports (#406)."""
+
+from __future__ import annotations
+
+import re
+from typing import Any
+
+from reviewer_merge_simulation import assess_merge_simulation_report
+
+STATUS_PASSED = "passed"
+STATUS_FAILED = "failed"
+STATUS_BASELINE_EQUIVALENT = "baseline-equivalent failure accepted"
+STATUS_MERGE_SIM_RESOLVED = "raw-head failure resolved by merge simulation"
+STATUS_TRANSIENT_PASS = "passed after transient failure investigation"
+
+ALLOWED_VALIDATION_STATUSES = frozenset({
+    STATUS_PASSED,
+    STATUS_FAILED,
+    STATUS_BASELINE_EQUIVALENT,
+    STATUS_MERGE_SIM_RESOLVED,
+    STATUS_TRANSIENT_PASS,
+})
+
+_STATUS_FIELD_RE = re.compile(
+    r"^\s*[-*]?\s*(?:validation status|pr-head validation status|"
+    r"official validation status)\s*:\s*(.+?)\s*$",
+    re.IGNORECASE | re.MULTILINE,
+)
+_RAW_HEAD_RESULT_RE = re.compile(
+    r"^\s*[-*]?\s*raw pr-head validation result\s*:\s*(.+?)\s*$",
+    re.IGNORECASE | re.MULTILINE,
+)
+_MERGE_SIM_RESULT_RE = re.compile(
+    r"^\s*[-*]?\s*merge simulation result\s*:\s*(.+?)\s*$",
+    re.IGNORECASE | re.MULTILINE,
+)
+_BASELINE_WORKTREE_USED_RE = re.compile(
+    r"^\s*[-*]?\s*baseline (?:validation )?worktree(?: used)?\s*:\s*(.+?)\s*$",
+    re.IGNORECASE | re.MULTILINE,
+)
+_BASELINE_TARGET_SHA_RE = re.compile(
+    r"^\s*[-*]?\s*baseline target sha\s*:\s*([0-9a-f]{7,40})\s*$",
+    re.IGNORECASE | re.MULTILINE,
+)
+_FAILURE_SIGNATURE_RE = re.compile(
+    r"failure signatures match\s*:\s*(true|yes)",
+    re.IGNORECASE,
+)
+_BASELINE_FAILURES_RE = re.compile(
+    r"baseline failures\s*:",
+    re.IGNORECASE,
+)
+_TRANSIENT_HISTORY_RE = re.compile(
+    r"(?:transient validation failure|earlier validation failure|"
+    r"prior failure|failure history|failed then passed)",
+    re.IGNORECASE,
+)
+_FULL_SHA = re.compile(r"^[0-9a-f]{40}$", re.IGNORECASE)
+
+
+def _first_match(pattern: re.Pattern[str], text: str) -> str:
+    match = pattern.search(text or "")
+    return (match.group(1).strip() if match else "")
+
+
+def _normalize_status_label(raw: str) -> str:
+    text = (raw or "").strip().lower()
+    for status in ALLOWED_VALIDATION_STATUSES:
+        if text == status.lower():
+            return status
+    return raw.strip()
+
+
+def _baseline_proof_complete(text: str, baseline_proof: dict | None) -> bool:
+    proof = baseline_proof or {}
+    worktree = (
+        (proof.get("worktree_path") or "").strip()
+        or _first_match(_BASELINE_WORKTREE_USED_RE, text)
+    ).lower()
+    if not worktree or worktree in {"none", "n/a", "not used", "not applicable"}:
+        return False
+    if "branches/" not in worktree and not worktree.startswith("branches/"):
+        return False
+    target_sha = (proof.get("baseline_target_sha") or "").strip()
+    if not target_sha:
+        target_sha = _first_match(_BASELINE_TARGET_SHA_RE, text)
+    if not _FULL_SHA.match(target_sha or ""):
+        return False
+    if proof.get("failure_signatures_match") is True:
+        return True
+    if _FAILURE_SIGNATURE_RE.search(text) and _BASELINE_FAILURES_RE.search(text):
+        return True
+    return False
+
+
+def _merge_simulation_passed(text: str, command_log: list | None) -> bool:
+    merge_result = _first_match(_MERGE_SIM_RESULT_RE, text).lower()
+    if merge_result in {"passed", "pass", "clean", "succeeded", "success"}:
+        sim = assess_merge_simulation_report(text, command_log=command_log)
+        return sim.get("proven") and not sim.get("block")
+    if "pass" in merge_result and "fail" not in merge_result:
+        sim = assess_merge_simulation_report(text, command_log=command_log)
+        return sim.get("proven") and not sim.get("block")
+    return False
+
+
+def _raw_head_failed(text: str) -> bool:
+    raw = _first_match(_RAW_HEAD_RESULT_RE, text).lower()
+    if raw in {"failed", "fail", "failure"}:
+        return True
+    if "fail" in raw and "pass" not in raw:
+        return True
+    return bool(re.search(r"\bfailed\b.*pr-head validation", text, re.IGNORECASE))
+
+
+def _raw_head_passed(text: str) -> bool:
+    raw = _first_match(_RAW_HEAD_RESULT_RE, text).lower()
+    return raw in {"passed", "pass", "success"}
+
+
+def assess_validation_status_vocabulary(
+    report_text: str,
+    *,
+    command_log: list | None = None,
+    baseline_proof: dict | None = None,
+) -> dict[str, Any]:
+    """Bind validation-status labels to the proof path that actually ran (#406)."""
+    text = report_text or ""
+    reasons: list[str] = []
+    status_raw = _first_match(_STATUS_FIELD_RE, text)
+    status = _normalize_status_label(status_raw) if status_raw else ""
+
+    if status_raw and status not in ALLOWED_VALIDATION_STATUSES:
+        reasons.append(
+            f"unknown validation status {status_raw!r}; use one of "
+            f"{sorted(ALLOWED_VALIDATION_STATUSES)}"
+        )
+
+    if status == STATUS_BASELINE_EQUIVALENT:
+        if not _baseline_proof_complete(text, baseline_proof):
+            reasons.append(
+                "baseline-equivalent failure accepted requires baseline "
+                "worktree path, baseline target SHA, and matching failure "
+                "signatures (#406)"
+            )
+
+    if status == STATUS_MERGE_SIM_RESOLVED:
+        if not _raw_head_failed(text):
+            reasons.append(
+                "raw-head failure resolved by merge simulation requires "
+                "raw PR-head validation result: failed (#406)"
+            )
+        if not _merge_simulation_passed(text, command_log):
+            reasons.append(
+                "raw-head failure resolved by merge simulation requires "
+                "passing merge simulation with worktree/index mutation proof "
+                "(#317/#406)"
+            )
+
+    if status == STATUS_TRANSIENT_PASS:
+        if not _TRANSIENT_HISTORY_RE.search(text):
+            reasons.append(
+                "passed after transient failure investigation requires "
+                "documented earlier validation failure history (#396/#406)"
+            )
+
+    if status == STATUS_PASSED and _raw_head_failed(text):
+        reasons.append(
+            "validation status passed contradicts raw PR-head validation "
+            "failure; use a precise status (#406)"
+        )
+
+    if status == STATUS_BASELINE_EQUIVALENT and _merge_simulation_passed(
+        text, command_log
+    ) and not _baseline_proof_complete(text, baseline_proof):
+        reasons.append(
+            "baseline-equivalent failure accepted is misleading when only "
+            "merge simulation resolved the failure; use "
+            "'raw-head failure resolved by merge simulation' (#406)"
+        )
+
+    if status == STATUS_FAILED and _merge_simulation_passed(text, command_log):
+        reasons.append(
+            "validation status failed contradicts passing merge simulation; "
+            "report the precise resolution status (#406)"
+        )
+
+    block = bool(reasons)
+    return {
+        "block": block,
+        "proven": not block,
+        "status_claimed": status or None,
+        "raw_status_label": status_raw or None,
+        "raw_head_failed": _raw_head_failed(text),
+        "raw_head_passed": _raw_head_passed(text),
+        "merge_simulation_passed": _merge_simulation_passed(text, command_log),
+        "baseline_proof_complete": _baseline_proof_complete(text, baseline_proof),
+        "reasons": reasons,
+        "safe_next_action": (
+            "use a validation status that matches the proof path executed "
+            "(baseline worktree, merge simulation, or transient history)"
+            if reasons
+            else "proceed"
+        ),
+    }