Expose custom_repr function that precedes safe_repr invocation in serializer (#3438)

closes #3427

Author: sl0thentr0py

sentry_sdk/client.py

@@ -531,6 +531,7 @@ def _prepare_event(
                     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"),
+                    custom_repr=self.options.get("custom_repr"),
                 ),
             )
 

sentry_sdk/consts.py

@@ -539,6 +539,7 @@ def __init__(
         spotlight=None,  # type: Optional[Union[bool, str]]
         cert_file=None,  # type: Optional[str]
         key_file=None,  # type: Optional[str]
+        custom_repr=None,  # type: Optional[Callable[..., Optional[str]]]
     ):
         # type: (...) -> None
         pass

sentry_sdk/serializer.py

@@ -112,6 +112,7 @@ def serialize(event, **kwargs):
     :param max_request_body_size: If set to "always", will never trim request bodies.
     :param max_value_length: The max length to strip strings to, defaults to sentry_sdk.consts.DEFAULT_MAX_VALUE_LENGTH
     :param is_vars: If we're serializing vars early, we want to repr() things that are JSON-serializable to make their type more apparent. For example, it's useful to see the difference between a unicode-string and a bytestring when viewing a stacktrace.
+    :param custom_repr: A custom repr function that runs before safe_repr on the object to be serialized. If it returns None or throws internally, we will fallback to safe_repr.
 
     """
     memo = Memo()
@@ -123,6 +124,17 @@ def serialize(event, **kwargs):
     )  # type: bool
     max_value_length = kwargs.pop("max_value_length", None)  # type: Optional[int]
     is_vars = kwargs.pop("is_vars", False)
+    custom_repr = kwargs.pop("custom_repr", None)  # type: Callable[..., Optional[str]]
+
+    def _safe_repr_wrapper(value):
+        # type: (Any) -> str
+        try:
+            repr_value = None
+            if custom_repr is not None:
+                repr_value = custom_repr(value)
+            return repr_value or safe_repr(value)
+        except Exception:
+            return safe_repr(value)
 
     def _annotate(**meta):
         # type: (**Any) -> None
@@ -257,7 +269,7 @@ def _serialize_node_impl(
             _annotate(rem=[["!limit", "x"]])
             if is_databag:
                 return _flatten_annotated(
-                    strip_string(safe_repr(obj), max_length=max_value_length)
+                    strip_string(_safe_repr_wrapper(obj), max_length=max_value_length)
                 )
             return None
 
@@ -274,7 +286,7 @@ def _serialize_node_impl(
             if should_repr_strings or (
                 isinstance(obj, float) and (math.isinf(obj) or math.isnan(obj))
             ):
-                return safe_repr(obj)
+                return _safe_repr_wrapper(obj)
             else:
                 return obj
 
@@ -285,7 +297,7 @@ def _serialize_node_impl(
             return (
                 str(format_timestamp(obj))
                 if not should_repr_strings
-                else safe_repr(obj)
+                else _safe_repr_wrapper(obj)
             )
 
         elif isinstance(obj, Mapping):
@@ -345,13 +357,13 @@ def _serialize_node_impl(
             return rv_list
 
         if should_repr_strings:
-            obj = safe_repr(obj)
+            obj = _safe_repr_wrapper(obj)
         else:
             if isinstance(obj, bytes) or isinstance(obj, bytearray):
                 obj = obj.decode("utf-8", "replace")
 
             if not isinstance(obj, str):
-                obj = safe_repr(obj)
+                obj = _safe_repr_wrapper(obj)
 
         is_span_description = (
             len(path) == 3 and path[0] == "spans" and path[-1] == "description"

sentry_sdk/utils.py

@@ -585,8 +585,9 @@ def serialize_frame(
     include_local_variables=True,
     include_source_context=True,
     max_value_length=None,
+    custom_repr=None,
 ):
-    # type: (FrameType, Optional[int], bool, bool, Optional[int]) -> Dict[str, Any]
+    # type: (FrameType, Optional[int], bool, bool, Optional[int], Optional[Callable[..., Optional[str]]]) -> Dict[str, Any]
     f_code = getattr(frame, "f_code", None)
     if not f_code:
         abs_path = None
@@ -618,7 +619,9 @@ def serialize_frame(
     if include_local_variables:
         from sentry_sdk.serializer import serialize
 
-        rv["vars"] = serialize(dict(frame.f_locals), is_vars=True)
+        rv["vars"] = serialize(
+            dict(frame.f_locals), is_vars=True, custom_repr=custom_repr
+        )
 
     return rv
 
@@ -723,10 +726,12 @@ def single_exception_from_error_tuple(
         include_local_variables = True
         include_source_context = True
         max_value_length = DEFAULT_MAX_VALUE_LENGTH  # fallback
+        custom_repr = None
     else:
         include_local_variables = client_options["include_local_variables"]
         include_source_context = client_options["include_source_context"]
         max_value_length = client_options["max_value_length"]
+        custom_repr = client_options.get("custom_repr")
 
     frames = [
         serialize_frame(
@@ -735,6 +740,7 @@ def single_exception_from_error_tuple(
             include_local_variables=include_local_variables,
             include_source_context=include_source_context,
             max_value_length=max_value_length,
+            custom_repr=custom_repr,
         )
         for tb in iter_stacks(tb)
     ]

tests/test_client.py

@@ -944,6 +944,39 @@ def __repr__(self):
     assert frame["vars"]["environ"] == {"a": "<This is me>"}
 
 
+def test_custom_repr_on_vars(sentry_init, capture_events):
+    class Foo:
+        pass
+
+    class Fail:
+        pass
+
+    def custom_repr(value):
+        if isinstance(value, Foo):
+            return "custom repr"
+        elif isinstance(value, Fail):
+            raise ValueError("oops")
+        else:
+            return None
+
+    sentry_init(custom_repr=custom_repr)
+    events = capture_events()
+
+    try:
+        my_vars = {"foo": Foo(), "fail": Fail(), "normal": 42}
+        1 / 0
+    except ZeroDivisionError:
+        capture_exception()
+
+    (event,) = events
+    (exception,) = event["exception"]["values"]
+    (frame,) = exception["stacktrace"]["frames"]
+    my_vars = frame["vars"]["my_vars"]
+    assert my_vars["foo"] == "custom repr"
+    assert my_vars["normal"] == "42"
+    assert "Fail object" in my_vars["fail"]
+
+
 @pytest.mark.parametrize(
     "dsn",
     [

tests/test_serializer.py

@@ -114,6 +114,31 @@ def test_custom_mapping_doesnt_mess_with_mock(extra_normalizer):
     assert len(m.mock_calls) == 0
 
 
+def test_custom_repr(extra_normalizer):
+    class Foo:
+        pass
+
+    def custom_repr(value):
+        if isinstance(value, Foo):
+            return "custom"
+        else:
+            return value
+
+    result = extra_normalizer({"foo": Foo(), "string": "abc"}, custom_repr=custom_repr)
+    assert result == {"foo": "custom", "string": "abc"}
+
+
+def test_custom_repr_graceful_fallback_to_safe_repr(extra_normalizer):
+    class Foo:
+        pass
+
+    def custom_repr(value):
+        raise ValueError("oops")
+
+    result = extra_normalizer({"foo": Foo()}, custom_repr=custom_repr)
+    assert "Foo object" in result["foo"]
+
+
 def test_trim_databag_breadth(body_normalizer):
     data = {
         "key{}".format(i): "value{}".format(i) for i in range(MAX_DATABAG_BREADTH + 10)