Read MAX_STRING_LENGTH from client options (#2121)

Author: Peter Uittenbroek

sentry_sdk/client.py

@@ -286,7 +286,10 @@ def _prepare_event(
                     "values": [
                         {
                             "stacktrace": current_stacktrace(
-                                self.options["include_local_variables"]
+                                self.options["max_string_length"],
+                                include_local_variables=self.options[
+                                    "include_local_variables"
+                                ],
                             ),
                             "crashed": False,
                             "current": True,

sentry_sdk/consts.py

@@ -197,6 +197,7 @@ def __init__(
         ],  # type: Optional[Sequence[str]]
         functions_to_trace=[],  # type: Sequence[Dict[str, str]]  # noqa: B006
         event_scrubber=None,  # type: Optional[sentry_sdk.scrubber.EventScrubber]
+        max_string_length=1024,  # type: int
     ):
         # type: (...) -> None
         pass

sentry_sdk/integrations/logging.py

@@ -205,7 +205,10 @@ def _emit(self, record):
                     "values": [
                         {
                             "stacktrace": current_stacktrace(
-                                client_options["include_local_variables"]
+                                client_options["max_string_length"],
+                                include_local_variables=client_options[
+                                    "include_local_variables"
+                                ],
                             ),
                             "crashed": False,
                             "current": True,

sentry_sdk/serializer.py

@@ -121,6 +121,7 @@ def serialize(event, **kwargs):
     meta_stack = []  # type: List[Dict[str, Any]]
 
     keep_request_bodies = kwargs.pop("request_bodies", None) == "always"  # type: bool
+    max_string_length = kwargs.pop("max_string_length", None)  # type: Optional[int]
 
     def _annotate(**meta):
         # type: (**Any) -> None
@@ -293,7 +294,9 @@ def _serialize_node_impl(
         if remaining_depth is not None and remaining_depth <= 0:
             _annotate(rem=[["!limit", "x"]])
             if is_databag:
-                return _flatten_annotated(strip_string(safe_repr(obj)))
+                return _flatten_annotated(
+                    strip_string(safe_repr(obj), max_length=max_string_length)
+                )
             return None
 
         if is_databag and global_repr_processors:
@@ -394,7 +397,7 @@ def _serialize_node_impl(
         if is_span_description:
             return obj
 
-        return _flatten_annotated(strip_string(obj))
+        return _flatten_annotated(strip_string(obj, max_string_length))
 
     #
     # Start of serialize() function

sentry_sdk/utils.py

@@ -468,6 +468,7 @@ def iter_stacks(tb):
 def get_lines_from_file(
     filename,  # type: str
     lineno,  # type: int
+    max_length,  # type: int
     loader=None,  # type: Optional[Any]
     module=None,  # type: Optional[str]
 ):
@@ -496,11 +497,12 @@ def get_lines_from_file(
 
     try:
         pre_context = [
-            strip_string(line.strip("\r\n")) for line in source[lower_bound:lineno]
+            strip_string(line.strip("\r\n"), max_length=max_length)
+            for line in source[lower_bound:lineno]
         ]
-        context_line = strip_string(source[lineno].strip("\r\n"))
+        context_line = strip_string(source[lineno].strip("\r\n"), max_length=max_length)
         post_context = [
-            strip_string(line.strip("\r\n"))
+            strip_string(line.strip("\r\n"), max_length=max_length)
             for line in source[(lineno + 1) : upper_bound]
         ]
         return pre_context, context_line, post_context
@@ -512,6 +514,7 @@ def get_lines_from_file(
 def get_source_context(
     frame,  # type: FrameType
     tb_lineno,  # type: int
+    max_string_length,  # type: int
 ):
     # type: (...) -> Tuple[List[Annotated[str]], Optional[Annotated[str]], List[Annotated[str]]]
     try:
@@ -528,7 +531,9 @@ def get_source_context(
         loader = None
     lineno = tb_lineno - 1
     if lineno is not None and abs_path:
-        return get_lines_from_file(abs_path, lineno, loader, module)
+        return get_lines_from_file(
+            abs_path, lineno, max_string_length, loader=loader, module=module
+        )
     return [], None, []
 
 
@@ -602,7 +607,11 @@ def filename_for_module(module, abs_path):
 
 
 def serialize_frame(
-    frame, tb_lineno=None, include_local_variables=True, include_source_context=True
+    frame,
+    tb_lineno=None,
+    include_local_variables=True,
+    include_source_context=True,
+    max_string_length=None,
 ):
     # type: (FrameType, Optional[int], bool, bool) -> Dict[str, Any]
     f_code = getattr(frame, "f_code", None)
@@ -630,7 +639,7 @@ def serialize_frame(
 
     if include_source_context:
         rv["pre_context"], rv["context_line"], rv["post_context"] = get_source_context(
-            frame, tb_lineno
+            frame, tb_lineno, max_string_length
         )
 
     if include_local_variables:
@@ -639,7 +648,9 @@ def serialize_frame(
     return rv
 
 
-def current_stacktrace(include_local_variables=True, include_source_context=True):
+def current_stacktrace(
+    max_string_length, include_local_variables=True, include_source_context=True
+):
     # type: (bool, bool) -> Any
     __tracebackhide__ = True
     frames = []
@@ -652,6 +663,7 @@ def current_stacktrace(include_local_variables=True, include_source_context=True
                     f,
                     include_local_variables=include_local_variables,
                     include_source_context=include_source_context,
+                    max_string_length=max_string_length,
                 )
             )
         f = f.f_back
@@ -724,16 +736,19 @@ def single_exception_from_error_tuple(
     if client_options is None:
         include_local_variables = True
         include_source_context = True
+        max_string_length = MAX_STRING_LENGTH  # fallback
     else:
         include_local_variables = client_options["include_local_variables"]
         include_source_context = client_options["include_source_context"]
+        max_string_length = client_options["max_string_length"]
 
     frames = [
         serialize_frame(
             tb.tb_frame,
             tb_lineno=tb.tb_lineno,
             include_local_variables=include_local_variables,
             include_source_context=include_source_context,
+            max_string_length=max_string_length,
         )
         for tb in iter_stacks(tb)
     ]
@@ -819,9 +834,7 @@ def exceptions_from_error(
     parent_id = exception_id
     exception_id += 1
 
-    should_supress_context = (
-        hasattr(exc_value, "__suppress_context__") and exc_value.__suppress_context__  # type: ignore
-    )
+    should_supress_context = hasattr(exc_value, "__suppress_context__") and exc_value.__suppress_context__  # type: ignore
     if should_supress_context:
         # Add direct cause.
         # The field `__cause__` is set when raised with the exception (using the `from` keyword).
@@ -1082,7 +1095,6 @@ def _is_in_project_root(abs_path, project_root):
 
 def strip_string(value, max_length=None):
     # type: (str, Optional[int]) -> Union[AnnotatedValue, str]
-    # TODO: read max_length from config
     if not value:
         return value
 

tests/test_serializer.py

@@ -144,3 +144,20 @@ def test_no_trimming_if_request_bodies_is_always(body_normalizer):
     result = body_normalizer(data, request_bodies="always")
 
     assert result == data
+
+
+def test_max_string_length_default(body_normalizer):
+    data = {"key": "a" * 2000}
+
+    result = body_normalizer(data)
+
+    assert len(result["key"]) == 1024  # fallback max length
+
+
+def test_max_string_length(body_normalizer):
+    data = {"key": "a" * 2000}
+
+    max_string_length = 2000
+    result = body_normalizer(data, max_string_length=max_string_length)
+
+    assert len(result["key"]) == max_string_length