fix: move start_coverage_collection after _initialized guard, reset _source_root on stop

- start_coverage_collection() was before the _initialized check, causing
  repeated initialize() calls to stop/restart coverage and lose data
- stop_coverage_collection() now resets _source_root for cleanup completeness

Author: sohil-kshirsagar

drift/core/coverage_server.py

@@ -91,7 +91,7 @@ def start_coverage_collection() -> bool:
 
 def stop_coverage_collection() -> None:
     """Stop coverage collection and clean up. Thread-safe."""
-    global _cov_instance, _branch_cache
+    global _cov_instance, _branch_cache, _source_root
     with _lock:
         if _cov_instance is not None:
             try:
@@ -100,6 +100,7 @@ def stop_coverage_collection() -> None:
                 pass
             _cov_instance = None
         _branch_cache = None
+        _source_root = None
 
 
 def take_coverage_snapshot(baseline: bool = False) -> dict:

drift/core/drift_sdk.py

@@ -179,17 +179,16 @@ def initialize(
             )
             env = env_from_var
 
-        # Start coverage collection early (before any SDK mode checks that might return early),
-        # but after the _initialized guard so we don't re-invoke on repeated initialize() calls.
-        # Coverage data is accessed via protobuf channel (communicator handles requests).
-        from .coverage_server import start_coverage_collection
-
-        start_coverage_collection()
-
         if cls._initialized:
             logger.debug("Already initialized, skipping...")
             return instance
 
+        # Start coverage collection after the _initialized guard so repeated
+        # initialize() calls don't stop/restart coverage and lose accumulated data.
+        from .coverage_server import start_coverage_collection
+
+        start_coverage_collection()
+
         file_config = instance.file_config
 
         if (