fix: reject trailing newline in collection-name validation

Python's $ anchor matches just before a final newline, so a $-anchored
re.match(r'^[a-zA-Z0-9_-]{1,128}$', name) accepted "papers\n". In local
mode get_or_create_collection() then hit SQLite's CHECK via
INSERT OR IGNORE, silently created no row, and returned a Collection that
failed later on add(). Switch all three duplicated validators (local,
cloud, sqlite backends) to re.fullmatch() so the whole string must match.

pageindex/backend/cloud.py

@@ -89,7 +89,9 @@ class CloudBackend:
 
     @staticmethod
     def _validate_collection_name(name: str) -> None:
-        if not re.match(r'^[a-zA-Z0-9_-]{1,128}$', name):
+        # .fullmatch() (not .match()): a $-anchored .match() would accept a
+        # trailing newline ("papers\n") because $ matches just before a final \n.
+        if not re.fullmatch(r'[a-zA-Z0-9_-]{1,128}', name):
             raise PageIndexError(
                 f"Invalid collection name: {name!r}. "
                 "Must be 1-128 chars of [a-zA-Z0-9_-]."

pageindex/backend/local.py

@@ -17,7 +17,9 @@ from ..backend.protocol import AgentTools
 from ..errors import (FileTypeError, DocumentNotFoundError, CollectionNotFoundError,
                       IndexingError, PageIndexError)
 
-_COLLECTION_NAME_RE = re.compile(r'^[a-zA-Z0-9_-]{1,128}$')
+# Matched with .fullmatch() (not .match()): a $-anchored .match() would accept a
+# trailing newline ("papers\n") because $ matches just before a final \n.
+_COLLECTION_NAME_RE = re.compile(r'[a-zA-Z0-9_-]{1,128}')
 
 
 class LocalBackend:
@@ -45,7 +47,7 @@ class LocalBackend:
 
     # Collection management
     def _validate_collection_name(self, name: str) -> None:
-        if not _COLLECTION_NAME_RE.match(name):
+        if not _COLLECTION_NAME_RE.fullmatch(name):
             raise PageIndexError(f"Invalid collection name: {name!r}. Must be 1-128 chars of [a-zA-Z0-9_-].")
 
     def create_collection(self, name: str) -> None:

pageindex/storage/sqlite.py

@@ -10,11 +10,13 @@ from ..errors import CollectionAlreadyExistsError, PageIndexError
 # itself (not just relying on LocalBackend's pre-check or the schema's CHECK
 # constraint below) because it's a public StorageEngine that can be used
 # directly, bypassing LocalBackend entirely.
-_COLLECTION_NAME_RE = re.compile(r'^[a-zA-Z0-9_-]{1,128}$')
+# Matched with .fullmatch() (not .match()): a $-anchored .match() would accept a
+# trailing newline ("papers\n") because $ matches just before a final \n.
+_COLLECTION_NAME_RE = re.compile(r'[a-zA-Z0-9_-]{1,128}')
 
 
 def _validate_collection_name(name: str) -> None:
-    if not _COLLECTION_NAME_RE.match(name):
+    if not _COLLECTION_NAME_RE.fullmatch(name):
         raise PageIndexError(f"Invalid collection name: {name!r}. Must be 1-128 chars of [a-zA-Z0-9_-].")
 
 

tests/test_local_backend.py

@@ -188,6 +188,17 @@ def test_delete_collection_rejects_path_traversal(backend, tmp_path):
     assert canary.exists()
 
 
+@pytest.mark.parametrize("bad_name", ["papers\n", "\npapers", "papers\n\n"])
+def test_get_or_create_collection_rejects_trailing_newline(backend, bad_name):
+    # Regression: Python's $ matches just before a final \n, so a $-anchored
+    # .match() accepted "papers\n"; get_or_create_collection then hit the SQL
+    # CHECK via INSERT OR IGNORE, silently created no row, and handed back a
+    # Collection that failed later on add(). .fullmatch() rejects it up front.
+    from pageindex.errors import PageIndexError
+    with pytest.raises(PageIndexError, match="Invalid collection name"):
+        backend.get_or_create_collection(bad_name)
+
+
 def test_add_document_missing_file_raises_file_not_found(backend, tmp_path):
     backend.get_or_create_collection("papers")
     with pytest.raises(FileNotFoundError):

tests/test_sqlite_storage.py

@@ -36,6 +36,10 @@ def test_create_duplicate_collection_raises_pageindex_error(storage):
 @pytest.mark.parametrize("bad_name", [
     "a/../../etc/passwd", "/etc/passwd", "a$(whoami)", ".hidden",
     "a b", "válid", "", "a" * 129,
+    # A trailing newline must be rejected: Python's $ matches just before a
+    # final \n, so a $-anchored .match() would let "papers\n" slip through
+    # (then INSERT OR IGNORE silently no-ops on the SQL CHECK).
+    "papers\n", "\npapers", "papers\n\n",
 ])
 def test_create_collection_rejects_invalid_names_at_the_python_layer(storage, bad_name):
     """SQLiteStorage must validate collection names itself — it's a public