Serialize vars early to avoid living references

sentry_sdk/client.py

@@ -5,12 +5,12 @@
 from collections.abc import Mapping
 from datetime import datetime, timezone
 from importlib import import_module
+from typing import cast
 
 from sentry_sdk._compat import PY37, check_uwsgi_thread_support
 from sentry_sdk.utils import (
     capture_internal_exceptions,
     current_stacktrace,
-    disable_capture_event,
     format_timestamp,
     get_sdk_name,
     get_type_name,
@@ -525,10 +525,13 @@ def _prepare_event(
         # Postprocess the event here so that annotated types do
         # generally not surface in before_send
         if event is not None:
-            event = serialize(
-                event,
-                max_request_body_size=self.options.get("max_request_body_size"),
-                max_value_length=self.options.get("max_value_length"),
+            event = cast(
+                "Event",
+                serialize(
+                    cast("Dict[str, Any]", event),
+                    max_request_body_size=self.options.get("max_request_body_size"),
+                    max_value_length=self.options.get("max_value_length"),
+                ),
             )
 
         before_send = self.options["before_send"]
@@ -726,9 +729,6 @@ def capture_event(
 
         :returns: An event ID. May be `None` if there is no DSN set or of if the SDK decided to discard the event for other reasons. In such situations setting `debug=True` on `init()` may help.
         """
-        if disable_capture_event.get(False):
-            return None
-
         if hint is None:
             hint = {}
         event_id = event.get("event_id")

sentry_sdk/integrations/pure_eval.py

@@ -131,7 +131,8 @@ def start(n):
     atok = source.asttokens()
 
     expressions.sort(key=closeness, reverse=True)
-    return {
+    vars = {
         atok.get_text(nodes[0]): value
         for nodes, value in expressions[: serializer.MAX_DATABAG_BREADTH]
     }
+    return serializer.serialize(vars, is_vars=True)

sentry_sdk/scope.py

@@ -31,6 +31,7 @@
     capture_internal_exception,
     capture_internal_exceptions,
     ContextVar,
+    disable_capture_event,
     event_from_exception,
     exc_info_from_error,
     logger,
@@ -1130,6 +1131,9 @@ def capture_event(self, event, hint=None, scope=None, **scope_kwargs):
 
         :returns: An `event_id` if the SDK decided to send the event (see :py:meth:`sentry_sdk.client._Client.capture_event`).
         """
+        if disable_capture_event.get(False):
+            return None
+
         scope = self._merge_scopes(scope, scope_kwargs)
 
         event_id = self.get_client().capture_event(event=event, hint=hint, scope=scope)
@@ -1157,6 +1161,9 @@ def capture_message(self, message, level=None, scope=None, **scope_kwargs):
 
         :returns: An `event_id` if the SDK decided to send the event (see :py:meth:`sentry_sdk.client._Client.capture_event`).
         """
+        if disable_capture_event.get(False):
+            return None
+
         if level is None:
             level = "info"
 
@@ -1182,6 +1189,9 @@ def capture_exception(self, error=None, scope=None, **scope_kwargs):
 
         :returns: An `event_id` if the SDK decided to send the event (see :py:meth:`sentry_sdk.client._Client.capture_event`).
         """
+        if disable_capture_event.get(False):
+            return None
+
         if error is not None:
             exc_info = exc_info_from_error(error)
         else:

sentry_sdk/serializer.py

@@ -25,7 +25,7 @@
     from typing import Type
     from typing import Union
 
-    from sentry_sdk._types import NotImplementedType, Event
+    from sentry_sdk._types import NotImplementedType
 
     Span = Dict[str, Any]
 
@@ -95,7 +95,7 @@ def __exit__(
 
 
 def serialize(event, **kwargs):
-    # type: (Event, **Any) -> Event
+    # type: (Dict[str, Any], **Any) -> Dict[str, Any]
     memo = Memo()
     path = []  # type: List[Segment]
     meta_stack = []  # type: List[Dict[str, Any]]
@@ -104,6 +104,7 @@ def serialize(event, **kwargs):
         kwargs.pop("max_request_body_size", None) == "always"
     )  # type: bool
     max_value_length = kwargs.pop("max_value_length", None)  # type: Optional[int]
+    is_vars = kwargs.pop("is_vars", False)
 
     def _annotate(**meta):
         # type: (**Any) -> None
@@ -138,23 +139,7 @@ def _should_repr_strings():
             are not (yet) in frame variables, but might encounter them when
             recursing (e.g.  we're in `event.exception`)
         """
-        try:
-            p0 = path[0]
-            if p0 == "stacktrace" and path[1] == "frames" and path[3] == "vars":
-                return True
-
-            if (
-                p0 in ("threads", "exception")
-                and path[1] == "values"
-                and path[3] == "stacktrace"
-                and path[4] == "frames"
-                and path[6] == "vars"
-            ):
-                return True
-        except IndexError:
-            return None
-
-        return False
+        return is_vars
 
     def _is_databag():
         # type: () -> Optional[bool]
@@ -387,7 +372,7 @@ def _serialize_node_impl(
     disable_capture_event.set(True)
     try:
         serialized_event = _serialize_node(event, **kwargs)
-        if meta_stack and isinstance(serialized_event, dict):
+        if not is_vars and meta_stack and isinstance(serialized_event, dict):
             serialized_event["_meta"] = meta_stack[0]
 
         return serialized_event

sentry_sdk/utils.py

@@ -616,7 +616,9 @@ def serialize_frame(
         )
 
     if include_local_variables:
-        rv["vars"] = frame.f_locals.copy()
+        from sentry_sdk.serializer import serialize
+
+        rv["vars"] = serialize(dict(frame.f_locals), is_vars=True)
 
     return rv
 

tests/test_scrubber.py

@@ -187,3 +187,20 @@ def test_recursive_event_scrubber(sentry_init, capture_events):
 
     (event,) = events
     assert event["extra"]["deep"]["deeper"][0]["deepest"]["password"] == "'[Filtered]'"
+
+
+def test_recursive_scrubber_does_not_override_original(sentry_init, capture_events):
+    sentry_init(event_scrubber=EventScrubber(recursive=True))
+    events = capture_events()
+
+    data = {"csrf": "secret"}
+    try:
+        raise RuntimeError("An error")
+    except Exception:
+        capture_exception()
+
+    (event,) = events
+    frames = event["exception"]["values"][0]["stacktrace"]["frames"]
+    (frame,) = frames
+    assert data["csrf"] == "secret"
+    assert frame["vars"]["data"]["csrf"] == "[Filtered]"