commit 22d9ee7e177bb0fcb514e898daaec98907ebde0c
author Benjamin Peterson <benjamin@python.org> Sat Dec 28 10:33:58 2013 -0600
committer Benjamin Peterson <benjamin@python.org> Sat Dec 28 10:33:58 2013 -0600
tree 054e2d5107f7c9a230c723ac130e32b5e1ed0602
parent 63cc99d9a6cf4751bd75b6bc32416fdb3a98440b

complain if the codec doesn't return unicode

Parser/tokenizer.c

diff --git a/Parser/tokenizer.c b/Parser/tokenizer.c
index 46cf9b2..249bb96 100644
--- a/Parser/tokenizer.c
+++ b/Parser/tokenizer.c
@@ -400,6 +400,12 @@
         buf = PyObject_CallObject(tok->decoding_readline, NULL);
         if (buf == NULL)
             return error_ret(tok);
+        if (!PyUnicode_Check(buf)) {
+            Py_DECREF(buf);
+            PyErr_SetString(PyExc_SyntaxError,
+                            "codec did not return a unicode object");
+            return error_ret(tok);
+        }
     } else {
         tok->decoding_buffer = NULL;
         if (PyString_CheckExact(buf))