feat: Added cycle detection (getsentry#236)

mitsuhiko · untitaker · commit fe9fa1d69945 · 2019-01-17T11:03:22.000+01:00
* feat: Added cycle detection

* fix: Bugs in cycle detector

* test: Simple cycle test

* test: Better test

* fix: Implement cycle detector for frame vars

* fix: Linting
diff --git a/sentry_sdk/client.py b/sentry_sdk/client.py
@@ -12,6 +12,7 @@
     get_type_name,
     capture_internal_exceptions,
     current_stacktrace,
+    break_cycles,
     logger,
 )
 from sentry_sdk.transport import make_transport
@@ -118,21 +119,23 @@ def _prepare_event(self, event, hint, scope):
             event, self.options["in_app_exclude"], self.options["in_app_include"]
         )
 
+        # Postprocess the event here so that annotated types do
+        # generally not surface in before_send
+        if event is not None:
+            event = break_cycles(event)
+            strip_event_mut(event)
+            event = flatten_metadata(event)
+            event = convert_types(event)
+
         before_send = self.options["before_send"]
         if before_send is not None:
+            new_event = None
             with capture_internal_exceptions():
                 new_event = before_send(event, hint)
             if new_event is None:
                 logger.info("before send dropped event (%s)", event)
             event = new_event
 
-        # Postprocess the event in the very end so that annotated types do
-        # generally not surface in before_send
-        if event is not None:
-            strip_event_mut(event)
-            event = flatten_metadata(event)
-            event = convert_types(event)
-
         return event
 
     def _is_ignored_error(self, event, hint=None):
diff --git a/sentry_sdk/utils.py b/sentry_sdk/utils.py
@@ -31,6 +31,8 @@
 # The logger is created here but initialized in the debug support module
 logger = logging.getLogger("sentry_sdk.errors")
 
+CYCLE_MARKER = object()
+
 
 def _get_debug_hub():
     # This function is replaced by debug.py
@@ -312,9 +314,13 @@ def _walk(obj, depth):
                 return [_walk(x, depth + 1) for x in obj]
             if isinstance(obj, Mapping):
                 return {safe_str(k): _walk(v, depth + 1) for k, v in obj.items()}
+
+        if obj is CYCLE_MARKER:
+            return obj
+
         return safe_repr(obj)
 
-    return _walk(obj, 0)
+    return _walk(break_cycles(obj), 0)
 
 
 def extract_locals(frame):
@@ -615,7 +621,26 @@ def strip_frame_mut(frame):
         frame["vars"] = strip_databag(frame["vars"])
 
 
+def break_cycles(obj, memo=None):
+    if memo is None:
+        memo = {}
+    if id(obj) in memo:
+        return CYCLE_MARKER
+    memo[id(obj)] = obj
+
+    try:
+        if isinstance(obj, Mapping):
+            return {k: break_cycles(v, memo) for k, v in obj.items()}
+        if isinstance(obj, Sequence) and not isinstance(obj, (text_type, bytes)):
+            return [break_cycles(v, memo) for v in obj]
+        return obj
+    finally:
+        del memo[id(obj)]
+
+
 def convert_types(obj):
+    if obj is CYCLE_MARKER:
+        return u"<cyclic>"
     if isinstance(obj, datetime):
         return obj.strftime("%Y-%m-%dT%H:%M:%SZ")
     if isinstance(obj, Mapping):
diff --git a/tests/test_client.py b/tests/test_client.py
@@ -313,3 +313,55 @@ def test_nan(sentry_init, capture_events):
     frames = event["exception"]["values"][0]["stacktrace"]["frames"]
     frame, = frames
     assert frame["vars"]["nan"] == "nan"
+
+
+def test_cyclic_frame_vars(sentry_init, capture_events):
+    sentry_init()
+    events = capture_events()
+
+    try:
+        a = {}
+        a["a"] = a
+        1 / 0
+    except Exception:
+        capture_exception()
+
+    event, = events
+    assert event["exception"]["values"][0]["stacktrace"]["frames"][0]["vars"]["a"] == {
+        "a": "<cyclic>"
+    }
+
+
+def test_cyclic_data(sentry_init, capture_events):
+    sentry_init()
+    events = capture_events()
+
+    with configure_scope() as scope:
+        data = {}
+        data["is_cyclic"] = data
+
+        other_data = ""
+        data["not_cyclic"] = other_data
+        data["not_cyclic2"] = other_data
+        scope.set_extra("foo", data)
+
+    capture_message("hi")
+    event, = events
+
+    data = event["extra"]["foo"]
+    assert data == {"not_cyclic2": "", "not_cyclic": "", "is_cyclic": "<cyclic>"}
+
+
+def test_databag_stripping(sentry_init, capture_events):
+    sentry_init()
+    events = capture_events()
+
+    try:
+        a = "A" * 16000  # noqa
+        1 / 0
+    except Exception:
+        capture_exception()
+
+    event, = events
+
+    assert len(json.dumps(event)) < 10000
