From: Benjamin Peterson <benjamin@python.org>
Date: Tue, 19 Jan 2016 05:17:54 +0000 (-0800)
Subject: set tp_new from the class in the hierarchy that actually owns the descriptor (closes... 
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=a4acf1ff847f072b2035bbb02c9af5496e59d6d6;p=python

set tp_new from the class in the hierarchy that actually owns the descriptor (closes #25731)

Debugging by Eryk Sun.
---

diff --git a/Objects/typeobject.c b/Objects/typeobject.c
index 8a5623f879..baee1f1ad2 100644
--- a/Objects/typeobject.c
+++ b/Objects/typeobject.c
@@ -6239,7 +6239,7 @@ update_one_slot(PyTypeObject *type, slotdef *p)
                sanity checks and constructing a new argument
                list.  Cut all that nonsense short -- this speeds
                up instance creation tremendously. */
-            specific = (void *)type->tp_new;
+            specific = (void *)((PyTypeObject *)PyCFunction_GET_SELF(descr))->tp_new;
             /* XXX I'm not 100% sure that there isn't a hole
                in this reasoning that requires additional
                sanity checks.  I'll buy the first person to