Fix: Correct Content-Length for StringIO with multi-byte characters

src/requests/utils.py

@@ -142,6 +142,11 @@ def super_len(o):
         # of latin-1 (iso-8859-1) like http.client.
         o = o.encode("utf-8")
 
+    if isinstance(o, io.StringIO):
+        total_length = len(o.getvalue().encode("utf-8"))
+        current_position = len(o.getvalue()[: o.tell()].encode("utf-8"))
+        return max(0, total_length - current_position)
+
     if hasattr(o, "__len__"):
         total_length = len(o)
 

tests/test_utils.py

@@ -142,6 +142,18 @@ def test_super_len_with_tell(self):
         foo.read(2)
         assert super_len(foo) == 3
 
+    def test_super_len_stringio_multibyte(self):
+        # UTF-8: "👍" is 4 bytes
+        data = "👍"
+        s = StringIO.StringIO(data)
+        assert super_len(s) == 4
+
+        # Test partial read
+        data_mixed = "abc👍"
+        s = StringIO.StringIO(data_mixed)
+        s.read(3)
+        assert super_len(s) == 4
+
     def test_super_len_with_fileno(self):
         with open(__file__, "rb") as f:
             length = super_len(f)