fix(preprocess): harden crash recovery integrity

This commit is contained in:
2026-07-12 04:49:15 +02:00
parent 1d5f8c76a7
commit c964920f16
7 changed files with 440 additions and 37 deletions
+24 -3
View File
@@ -46,6 +46,7 @@ class CorpusStore:
self.root = Path(root)
self.active_path = self.root / "ACTIVE"
self._replace = os.replace
self._fsync_directory = self._sync_root
self._ensure_root()
def _ensure_root(self) -> None:
@@ -110,15 +111,35 @@ class CorpusStore:
manifest = self.manifest(generation)
if manifest.manifest_id != generation:
raise UnsafeCorpusPath("manifest generation mismatch")
if self.active_generation() == generation:
return generation
previous = self.active_generation()
temporary = self.active_path.with_name(f".ACTIVE.{uuid.uuid4().hex}.tmp")
_atomic_write(temporary, (generation + "\n").encode())
self._replace(temporary, self.active_path)
replaced = False
try:
_atomic_write(temporary, (generation + "\n").encode())
self._replace(temporary, self.active_path)
replaced = True
self._fsync_directory()
except BaseException:
temporary.unlink(missing_ok=True)
if replaced:
if previous is None:
self.active_path.unlink(missing_ok=True)
else:
rollback = self.active_path.with_name(f".ACTIVE.rollback.{uuid.uuid4().hex}.tmp")
_atomic_write(rollback, (previous + "\n").encode())
self._replace(rollback, self.active_path)
self._sync_root()
raise
return generation
def _sync_root(self) -> None:
directory = os.open(self.root, os.O_RDONLY | os.O_DIRECTORY | os.O_NOFOLLOW)
try:
os.fsync(directory)
finally:
os.close(directory)
return generation
def active_generation(self) -> str | None:
try: