fix(evidence): lock GC and preflight search ownership

harness/tht/cli/preprocess_cmd.py

@@ -114,8 +114,7 @@ def gc_from_config(config: Path, *, dry_run: bool = False):
         retain_published_generations=cfg.vector.retain_published_generations,
     )
     pipeline.workspace_id = config.stem.lower().replace(".", "-").replace("_", "-")
-    with pipeline.store.writer_lock():
-        return pipeline.gc(workspace_root=corpus_root.parent, dry_run=dry_run)
+    return pipeline.gc(workspace_root=corpus_root.parent, dry_run=dry_run)
 
 
 @preprocess_app.command("evidence")

harness/tht/cli/search_cmd.py

@@ -57,13 +57,17 @@ def search_cmd(
     from tht.search import combined_search
 
     cfg = _load_config_or_exit(config)
+    from tht.search.evidence import validate_corpus_workspace
+
+    workspace_id = config.stem.lower().replace(".", "-").replace("_", "-")
+    validate_corpus_workspace(cfg, workspace_id)
     dwh_snapshot = _leased_dwh_snapshot(cfg, ctx)
     require_vector_cfg(cfg)
     from tht.search.evidence import active_searcher
 
     runtime_searcher = active_searcher(
         cfg, open_searcher(cfg),
-        workspace_id=config.stem.lower().replace(".", "-").replace("_", "-"),
+        workspace_id=workspace_id,
     )
     if kind is not None and kind not in KIND_MAP:
         typer.secho(
@@ -256,6 +260,10 @@ def pack_cmd(
     from tht.vectorstore.rest_client import VectorRestError
 
     cfg = _load_config_or_exit(config)
+    from tht.search.evidence import validate_corpus_workspace
+
+    workspace_id = config.stem.lower().replace(".", "-").replace("_", "-")
+    validate_corpus_workspace(cfg, workspace_id)
     dwh_snapshot = _leased_dwh_snapshot(cfg, ctx)
     require_vector_cfg(cfg)
 
@@ -272,7 +280,7 @@ def pack_cmd(
 
         searcher = active_searcher(
             cfg, open_searcher(cfg),
-            workspace_id=config.stem.lower().replace(".", "-").replace("_", "-"),
+            workspace_id=workspace_id,
         )
         embedder = make_embedder(cfg.embeddings)
         vec = embedder.embed_query(question)

harness/tht/corpus/pipeline.py

@@ -124,6 +124,10 @@ class CorpusPipeline:
         return protected
 
     def gc(self, *, workspace_root: Path, dry_run: bool = False) -> dict:
+        with self.store.writer_lock():
+            return self._gc(workspace_root=workspace_root, dry_run=dry_run)
+
+    def _gc(self, *, workspace_root: Path, dry_run: bool = False) -> dict:
         self._assert_workspace_binding()
         published = self.store.published_generations()
         list_vectors = getattr(self.vector_store, "list_evidence_generations", None)

harness/tht/corpus/store.py

@@ -10,6 +10,7 @@ import stat
 import shutil
 import uuid
 import hashlib
+import threading
 from datetime import UTC, datetime
 from pathlib import Path
 from contextlib import contextmanager
@@ -49,6 +50,7 @@ class CorpusStore:
         self.active_path = self.root / "ACTIVE"
         self._replace = os.replace
         self._fsync_directory = self._sync_root
+        self._lock_state = threading.local()
         self._ensure_root()
 
     def _ensure_root(self) -> None:
@@ -61,6 +63,14 @@ class CorpusStore:
 
     @contextmanager
     def writer_lock(self):
+        depth = getattr(self._lock_state, "depth", 0)
+        if depth:
+            self._lock_state.depth = depth + 1
+            try:
+                yield
+            finally:
+                self._lock_state.depth -= 1
+            return
         lock_path = self.root / ".writer.lock"
         fd = os.open(lock_path, os.O_RDWR | os.O_CREAT | os.O_NOFOLLOW | os.O_CLOEXEC, 0o600)
         try:
@@ -68,8 +78,10 @@ class CorpusStore:
             if not stat.S_ISREG(info.st_mode) or info.st_uid != os.getuid() or info.st_nlink != 1:
                 raise UnsafeCorpusPath("corpus writer lock is unsafe")
             fcntl.flock(fd, fcntl.LOCK_EX)
+            self._lock_state.depth = 1
             yield
         finally:
+            self._lock_state.depth = 0
             fcntl.flock(fd, fcntl.LOCK_UN)
             os.close(fd)
 

harness/tht/search/evidence.py

@@ -76,6 +76,26 @@ def active_searcher(cfg, delegate, *, workspace_id: str | None = None):
     return ActiveEvidenceSearcher(CorpusStore(corpus_root), delegate, workspace_id)
 
 
+def validate_corpus_workspace(cfg, workspace_id: str) -> None:
+    """Fail before downstream retrieval setup when configured corpus ownership differs."""
+    corpus = CorpusStore(cfg.paths.artifacts.parent / "corpus")
+    with corpus.writer_lock():
+        manifest = corpus.active_manifest()
+        if manifest is None:
+            return
+        persisted = manifest.metadata.get("workspace_id")
+        if not isinstance(persisted, str) or re.fullmatch(
+            r"[a-z][a-z0-9_-]{0,63}", persisted
+        ) is None:
+            raise CorpusWorkspaceMismatchError(
+                "corpus workspace ownership is missing or invalid; use a new corpus root or rebuild"
+            )
+        if persisted != workspace_id:
+            raise CorpusWorkspaceMismatchError(
+                "corpus belongs to a different workspace; use a new corpus root or rebuild"
+            )
+
+
 def resolve_evidence_file(
     store: CorpusStore, evidence_id: str, *, materialized_root=None,
 ) -> str: