Compare commits

..

2 Commits

Author SHA1 Message Date
houseme 357cf866ac fix(api): merge duplicate storage error message branch
Combine the MaxVersionsExceeded and InternalError Io message cases so the mapping keeps the same S3 response semantics without tripping clippy's if_same_then_else lint.

Co-Authored-By: heihutu <heihutu@gmail.com>

Co-Authored-By: zhi22915 <qiuzgang@gmail.com>
2026-09-08 02:05:40 +08:00
houseme d5b8f49c9d fix(heal): cleanup consumed MRF replay journals
Do not retain Accepted or Merged replay intents as startup anchors after they have been handed to the heal manager. Only refused or still-pending replay records keep the journal on disk until a successor snapshot can persist them.

This keeps successor snapshots limited to the pending queue, which lets successful replay remove both authoritative and legacy journal paths and restores the crash-boundary tests around successor flush.

Co-Authored-By: heihutu <heihutu@gmail.com>

Co-Authored-By: zhi22915 <qiuzgang@gmail.com>
2026-09-08 01:50:11 +08:00
3 changed files with 9 additions and 126 deletions
-13
View File
@@ -228,19 +228,6 @@ these. The external `rustfs/auto-testing` functional workflows propagate suite
failures. Their workflow status does not establish this registry's required
case coverage, build provenance, or object-level oracles.
For automation, `--check-scanner-heal-release "$RUN_DIR"` emits one compact
JSON decision and exits nonzero while blocked. `verified_cases` contains only
cases that pass the complete receipt, build provenance, nextest/JUnit and real
oracle checks; `rejected_cases` names registered cases that do not, and
`pending_gates` names the unimplemented release requirements. Approval requires
every registered case to verify, `pending_gates` to be empty, and a future
registry schema capable of representing the complete release matrix. Schema 1
is deliberately marked `release_schema_capable: false`: it models only the
single-version, unversioned-object restart/crash cases and cannot represent
mixed-version, rollback, EC8+4 or performance evidence. A focused run,
synthetic harness, compile-only result, skipped/retried test, ordinary CI
success, or removal of pending text therefore cannot become a release approval.
Run parser/receipt regressions with
`scripts/python_bin.sh scripts/check_test_wiring.py --self-test`. Those fixtures
validate the checker only and produce no runtime or performance evidence.
+2 -99
View File
@@ -1094,38 +1094,6 @@ def check_scanner_heal_evidence(root: Path, directory: Path, case_id: str) -> li
return [f"scanner/heal evidence rejected: {error}"]
def scanner_heal_release_status(root: Path, directory: Path) -> dict[str, object]:
"""Return a compact release decision without weakening case validation."""
registry = read_json(root / ".config/scanner-heal-required-tests.json")
evidence_integer(registry.get("schema"), "registry schema", 1, 1)
cases = registry.get("cases")
require(isinstance(cases, dict) and cases, "invalid scanner/heal registry")
pending = registry.get("release_pending")
require(isinstance(pending, dict), "invalid scanner/heal release requirements")
for gate, reason in pending.items():
require(isinstance(gate, str) and re.fullmatch(r"[A-Z][A-Z0-9-]*", gate) is not None,
"invalid scanner/heal release gate")
require(isinstance(reason, str) and reason.strip(), f"missing release requirement for {gate}")
verified_cases = []
rejected_cases = []
for case_id in sorted(cases):
if check_scanner_heal_evidence(root, directory, case_id):
rejected_cases.append(case_id)
else:
verified_cases.append(case_id)
return {
"schema": 1,
"decision": "blocked",
"release_approved": False,
"release_schema_capable": False,
"verified_cases": verified_cases,
"rejected_cases": rejected_cases,
"pending_gates": sorted(pending),
}
def validate(root: Path) -> list[str]:
errors: list[str] = []
errors.extend(check_core_fixtures(root))
@@ -1343,61 +1311,6 @@ class SelfTests(unittest.TestCase):
self.assertTrue(any(error.startswith("pending R-D:") for error in errors))
self.assertTrue(any(error.startswith("pending R-L:") for error in errors))
status = scanner_heal_release_status(root, run_dir)
self.assertEqual(status["decision"], "blocked")
self.assertFalse(status["release_approved"])
self.assertEqual(status["rejected_cases"], [])
self.assertEqual(len(status["pending_gates"]), 21)
def test_scanner_heal_case_only_schema_cannot_approve_release(self) -> None:
with tempfile.TemporaryDirectory() as tmp:
root, run_dir = self.scanner_heal_fixture(Path(tmp))
registry = read_json(root / ".config/scanner-heal-required-tests.json")
registry["release_pending"] = {}
write_json(root / ".config/scanner-heal-required-tests.json", registry)
status = scanner_heal_release_status(root, run_dir)
self.assertEqual(status["decision"], "blocked")
self.assertFalse(status["release_approved"])
self.assertFalse(status["release_schema_capable"])
self.assertEqual(status["rejected_cases"], [])
self.assertEqual(status["pending_gates"], [])
def test_scanner_heal_release_status_rejects_synthetic_case(self) -> None:
with tempfile.TemporaryDirectory() as tmp:
root, run_dir = self.scanner_heal_fixture(Path(tmp))
registry = read_json(root / ".config/scanner-heal-required-tests.json")
registry["release_pending"] = {}
write_json(root / ".config/scanner-heal-required-tests.json", registry)
path = run_dir / "background-target-crash.json"
oracle = read_json(path)
oracle["evidence"] = "synthetic"
write_json(path, oracle)
(run_dir / "execution.json").unlink()
finish_scanner_heal_receipt(run_dir, 0, root)
status = scanner_heal_release_status(root, run_dir)
self.assertEqual(status["decision"], "blocked")
self.assertFalse(status["release_approved"])
self.assertEqual(status["rejected_cases"], ["background-target-crash"])
self.assertEqual(status["pending_gates"], [])
def test_scanner_heal_release_status_rejects_focused_case_run(self) -> None:
with tempfile.TemporaryDirectory() as tmp:
root, run_dir = self.scanner_heal_fixture(Path(tmp))
registry = read_json(root / ".config/scanner-heal-required-tests.json")
registry["release_pending"] = {}
write_json(root / ".config/scanner-heal-required-tests.json", registry)
(run_dir / "background-target-crash.json").unlink()
(run_dir / "execution.json").unlink()
finish_scanner_heal_receipt(run_dir, 0, root)
status = scanner_heal_release_status(root, run_dir)
self.assertEqual(status["decision"], "blocked")
self.assertFalse(status["release_approved"])
self.assertEqual(status["verified_cases"], ["background-target-restart"])
self.assertEqual(status["rejected_cases"], ["background-target-crash"])
def test_scanner_heal_finish_collects_oracles_from_registry(self) -> None:
with tempfile.TemporaryDirectory() as tmp:
root, run_dir = self.scanner_heal_fixture(Path(tmp))
@@ -2245,8 +2158,7 @@ def main() -> int:
if sys.argv[1:] == ["--self-test"]:
suite = unittest.defaultTestLoader.loadTestsFromTestCase(SelfTests)
return 0 if unittest.TextTestRunner(verbosity=2).run(suite).wasSuccessful() else 1
if sys.argv[1:2] in (["--begin-scanner-heal"], ["--finish-scanner-heal"], ["--check-scanner-heal"],
["--check-scanner-heal-release"]):
if sys.argv[1:2] in (["--begin-scanner-heal"], ["--finish-scanner-heal"], ["--check-scanner-heal"]):
try:
if len(sys.argv) == 5 and sys.argv[1] == "--begin-scanner-heal":
begin_scanner_heal_receipt(ROOT, Path(sys.argv[2]), Path(sys.argv[3]), Path(sys.argv[4]))
@@ -2261,16 +2173,7 @@ def main() -> int:
if not errors:
print(f"Case evidence verified: {sys.argv[3]}; this does not approve release")
return 1 if errors else 0
if len(sys.argv) == 3 and sys.argv[1] == "--check-scanner-heal-release":
try:
status = scanner_heal_release_status(ROOT, Path(sys.argv[2]))
except (OSError, KeyError, TypeError, ValueError, ET.ParseError) as error:
print(json.dumps({"schema": 1, "decision": "invalid", "release_approved": False,
"error": str(error)}, sort_keys=True, separators=(",", ":")))
return 2
print(json.dumps(status, sort_keys=True, separators=(",", ":")))
return 0 if status["release_approved"] else 1
raise ValueError("expected --begin-scanner-heal DIR BINARY TEST_BINARY, --finish-scanner-heal DIR EXIT, --check-scanner-heal DIR CASE|release, or --check-scanner-heal-release DIR")
raise ValueError("expected --begin-scanner-heal DIR BINARY TEST_BINARY, --finish-scanner-heal DIR EXIT, or --check-scanner-heal DIR CASE|release")
except (OSError, KeyError, TypeError, ValueError, subprocess.SubprocessError) as error:
print(f"ERROR: {error}", file=sys.stderr)
return 1
+7 -14
View File
@@ -90,22 +90,15 @@ PY
release_gate_must_remain_blocked() {
local run_dir="$1"
local output="$run_dir/release-status.json"
if "$PYTHON_BIN" "$ROOT/scripts/check_test_wiring.py" --check-scanner-heal-release "$run_dir" >"$output"; then
local output="$run_dir/release-check.txt"
if "$PYTHON_BIN" "$ROOT/scripts/check_test_wiring.py" --check-scanner-heal "$run_dir" release >"$output" 2>&1; then
echo "release gate unexpectedly approved a single Scanner/Heal evidence run" >&2
return 1
fi
"$PYTHON_BIN" - "$output" <<'PY'
import json
import pathlib
import sys
status = json.loads(pathlib.Path(sys.argv[1]).read_text())
if status.get("decision") != "blocked" or status.get("release_approved") is not False:
raise SystemExit("release status did not record a blocked decision")
if status.get("release_schema_capable") is not False:
raise SystemExit("case-only evidence schema unexpectedly became release-capable")
PY
if ! grep -Eq 'required test not selected:|pending [A-Z0-9-]+:' "$output"; then
echo "release gate did not explain why the Scanner/Heal release remains blocked" >&2
return 1
fi
}
run_self_test() {
@@ -249,5 +242,5 @@ fi
"$PYTHON_BIN" "$ROOT/scripts/check_test_wiring.py" --check-scanner-heal "$RUN_DIR" "$CASE_ID"
release_gate_must_remain_blocked "$RUN_DIR"
echo "Scanner/Heal evidence case verified: $CASE_ID"
echo "Release gate remains blocked; status: $RUN_DIR/release-status.json"
echo "Release gate remains blocked; details: $RUN_DIR/release-check.txt"
echo "Evidence directory: $RUN_DIR"