PYTHON-5449 - Do not attach invalid document in exception message (#2539)

bson/__init__.py

@@ -1009,7 +1009,7 @@ def _dict_to_bson(
                 try:
                     elements.append(_element_to_bson(key, value, check_keys, opts))
                 except InvalidDocument as err:
-                    raise InvalidDocument(f"Invalid document {doc} | {err}") from err
+                    raise InvalidDocument(f"Invalid document: {err}", doc) from err
     except AttributeError:
         raise TypeError(f"encoder expected a mapping type but got: {doc!r}") from None
 

bson/_cbsonmodule.c

@@ -1645,11 +1645,11 @@ fail:
 }
 
 
-/* Update Invalid Document error message to include doc.
+/* Update Invalid Document error to include doc as a property.
  */
 void handle_invalid_doc_error(PyObject* dict) {
     PyObject *etype = NULL, *evalue = NULL, *etrace = NULL;
-    PyObject *msg = NULL, *dict_str = NULL, *new_msg = NULL;
+    PyObject *msg = NULL, *new_msg = NULL, *new_evalue = NULL;
     PyErr_Fetch(&etype, &evalue, &etrace);
     PyObject *InvalidDocument = _error("InvalidDocument");
     if (InvalidDocument == NULL) {
@@ -1659,26 +1659,22 @@ void handle_invalid_doc_error(PyObject* dict) {
     if (evalue && PyErr_GivenExceptionMatches(etype, InvalidDocument)) {
         PyObject *msg = PyObject_Str(evalue);
         if (msg) {
-            // Prepend doc to the existing message
-            PyObject *dict_str = PyObject_Str(dict);
-            if (dict_str == NULL) {
-                goto cleanup;
-            }
-            const char * dict_str_utf8 = PyUnicode_AsUTF8(dict_str);
-            if (dict_str_utf8 == NULL) {
-                goto cleanup;
-            }
             const char * msg_utf8 = PyUnicode_AsUTF8(msg);
             if (msg_utf8 == NULL) {
                 goto cleanup;
             }
-            PyObject *new_msg = PyUnicode_FromFormat("Invalid document %s | %s", dict_str_utf8, msg_utf8);
+            PyObject *new_msg = PyUnicode_FromFormat("Invalid document: %s", msg_utf8);
+            if (new_msg == NULL) {
+                goto cleanup;
+            }
+            // Add doc to the error instance as a property.
+            PyObject *new_evalue = PyObject_CallFunctionObjArgs(InvalidDocument, new_msg, dict, NULL);
             Py_DECREF(evalue);
             Py_DECREF(etype);
             etype = InvalidDocument;
             InvalidDocument = NULL;
-            if (new_msg) {
-                evalue = new_msg;
+            if (new_evalue) {
+                evalue = new_evalue;
             } else {
                 evalue = msg;
             }
@@ -1689,7 +1685,7 @@ cleanup:
     PyErr_Restore(etype, evalue, etrace);
     Py_XDECREF(msg);
     Py_XDECREF(InvalidDocument);
-    Py_XDECREF(dict_str);
+    Py_XDECREF(new_evalue);
     Py_XDECREF(new_msg);
 }
 

bson/errors.py

@@ -15,6 +15,8 @@
 """Exceptions raised by the BSON package."""
 from __future__ import annotations
 
+from typing import Any, Optional
+
 
 class BSONError(Exception):
     """Base class for all BSON exceptions."""
@@ -31,6 +33,17 @@ class InvalidStringData(BSONError):
 class InvalidDocument(BSONError):
     """Raised when trying to create a BSON object from an invalid document."""
 
+    def __init__(self, message: str, document: Optional[Any] = None) -> None:
+        super().__init__(message)
+        self._document = document
+
+    @property
+    def document(self) -> Any:
+        """The invalid document that caused the error.
+
+        ..versionadded:: 4.16"""
+        return self._document
+
 
 class InvalidId(BSONError):
     """Raised when trying to create an ObjectId from invalid data."""

doc/changelog.rst

@@ -1,6 +1,15 @@
 Changelog
 =========
 
+Changes in Version 4.16.0 (XXXX/XX/XX)
+--------------------------------------
+
+PyMongo 4.16 brings a number of changes including:
+
+- Removed invalid documents from :class:`bson.errors.InvalidDocument` error messages as
+  doing so may leak sensitive user data.
+  Instead, invalid documents are stored in :attr:`bson.errors.InvalidDocument.document`.
+
 Changes in Version 4.15.1 (2025/09/16)
 --------------------------------------
 

test/test_bson.py

@@ -1163,7 +1163,7 @@ class TestBSON(unittest.TestCase):
         ):
             encode({"t": Wrapper(1)})
 
-    def test_doc_in_invalid_document_error_message(self):
+    def test_doc_in_invalid_document_error_as_property(self):
         class Wrapper:
             def __init__(self, val):
                 self.val = val
@@ -1173,10 +1173,11 @@ class TestBSON(unittest.TestCase):
 
         self.assertEqual("1", repr(Wrapper(1)))
         doc = {"t": Wrapper(1)}
-        with self.assertRaisesRegex(InvalidDocument, f"Invalid document {doc}"):
+        with self.assertRaisesRegex(InvalidDocument, "Invalid document:") as cm:
             encode(doc)
+        self.assertEqual(cm.exception.document, doc)
 
-    def test_doc_in_invalid_document_error_message_mapping(self):
+    def test_doc_in_invalid_document_error_as_property_mapping(self):
         class MyMapping(abc.Mapping):
             def keys(self):
                 return ["t"]
@@ -1192,6 +1193,11 @@ class TestBSON(unittest.TestCase):
             def __iter__(self):
                 return iter(["t"])
 
+            def __eq__(self, other):
+                if isinstance(other, MyMapping):
+                    return True
+                return False
+
         class Wrapper:
             def __init__(self, val):
                 self.val = val
@@ -1201,8 +1207,9 @@ class TestBSON(unittest.TestCase):
 
         self.assertEqual("1", repr(Wrapper(1)))
         doc = MyMapping()
-        with self.assertRaisesRegex(InvalidDocument, f"Invalid document {doc}"):
+        with self.assertRaisesRegex(InvalidDocument, "Invalid document:") as cm:
             encode(doc)
+        self.assertEqual(cm.exception.document, doc)
 
 
 class TestCodecOptions(unittest.TestCase):