Merge commit from fork

* PYTHON-5996 Harden bson buffer size guard against signed integer overflow

* PYTHON-5996 Add changelog entry

* remove changelog entry for now

* Update doc/changelog.rst

* Update doc/changelog.rst

* Update doc/changelog.rst

* Update doc/changelog.rst

Author: blink1073

bson/buffer.c

@@ -18,6 +18,7 @@
 #define PY_SSIZE_T_CLEAN
 #include "Python.h"
 
+#include <limits.h>
 #include <stdlib.h>
 #include <string.h>
 
@@ -105,18 +106,25 @@ static int buffer_grow(buffer_t buffer, int min_length) {
  * Return non-zero and sets MemoryError on allocation failure.
  * Return non-zero and sets ValueError if `size` would exceed 2GiB. */
 static int buffer_assure_space(buffer_t buffer, int size) {
-    int new_size = buffer->position + size;
-    /* Check for overflow. */
-    if (new_size < buffer->position) {
+    long long new_size;
+    if (size < 0) {
         PyErr_SetString(PyExc_ValueError,
                         "Document would overflow BSON size limit");
         return 1;
     }
 
-    if (new_size <= buffer->size) {
+    /* Compute in a wider type so the addition cannot overflow `int`. */
+    new_size = (long long)buffer->position + (long long)size;
+    if (new_size > INT_MAX) {
+        PyErr_SetString(PyExc_ValueError,
+                        "Document would overflow BSON size limit");
+        return 1;
+    }
+
+    if ((int)new_size <= buffer->size) {
         return 0;
     }
-    return buffer_grow(buffer, new_size);
+    return buffer_grow(buffer, (int)new_size);
 }
 
 /* Save `size` bytes from the current position in `buffer` (and grow if needed).

test/test_bson.py

@@ -691,6 +691,14 @@ def test_overflow(self):
         self.assertTrue(encode({"x": -9223372036854775808}))
         self.assertRaises(OverflowError, encode, {"x": -9223372036854775809})
 
+    @unittest.skipUnless(bson.has_c(), "This test requires the C extension")
+    def test_encode_size_limit(self):
+        # PYTHON-5996: encoding must raise when a document's encoded size
+        # exceeds the BSON size limit.
+        big_value = "a" * (1 << 30)
+        with self.assertRaises(ValueError):
+            encode({"a": big_value, "b": big_value, "c": big_value})
+
     def test_small_long_encode_decode(self):
         encoded1 = encode({"x": 256})
         decoded1 = decode(encoded1)["x"]