Commit id-generator construction in the multi-writer id-gen tests

Constructing a `MultiWriterIdGenerator` prunes stale `stream_positions` rows for
writers no longer in the config. The test harness built each generator inside a
`runWithConnection` callback and never committed, so that cleanup only persisted
because adbapi keeps one connection per thread and leaves its transaction open
across calls — a later generator on the same thread then saw the uncommitted
delete.

The native Rust pool hands out a fresh connection per call and rolls back any
left mid-transaction, so the cleanup was lost and test_writer_config_change read
a stale writer position (persisted-upto 3 instead of 6). Commit the connection
after constructing the generator, so the cleanup persists regardless of pool
semantics; the delete is legitimate work that should be committed anyway.

Fixes test_writer_config_change on the Rust backend; psycopg2 and sqlite still
pass (16/16 on both).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W3G4M92AmwSSZCbmtMJU3d

tests/storage/test_id_generators.py

@@ -78,7 +78,7 @@ class MultiWriterIdGeneratorBase(HomeserverTestCase):
         writers: list[str] | None = None,
     ) -> MultiWriterIdGenerator:
         def _create(conn: LoggingDatabaseConnection) -> MultiWriterIdGenerator:
-            return MultiWriterIdGenerator(
+            id_gen = MultiWriterIdGenerator(
                 db_conn=conn,
                 db=self.db_pool,
                 notifier=self.hs.get_replication_notifier(),
@@ -90,6 +90,15 @@ class MultiWriterIdGeneratorBase(HomeserverTestCase):
                 writers=writers or ["master"],
                 positive=self.positive,
             )
+            # Constructing the generator prunes stale `stream_positions` rows
+            # (writers no longer in the config); commit so that persists for the
+            # next generator we create. Without this the test relies on the
+            # connection pool keeping the uncommitted transaction open and
+            # visible across `runWithConnection` calls, which adbapi happens to
+            # do (one connection per thread) but the native pool does not (a
+            # fresh connection each call, rolled back if left mid-transaction).
+            conn.commit()
+            return id_gen
 
         self.instances[instance_name] = self.get_success_or_raise(
             self.db_pool.runWithConnection(_create)