From e349bf23584eef20e0d1e1b2989d9b1430f15507 Mon Sep 17 00:00:00 2001
From: Zackery Spytz <zspytz@gmail.com>
Date: Sat, 18 Aug 2018 22:43:38 -0600
Subject: bpo-22602: Raise an exception in the UTF-7 decoder for ill-formed
 sequences starting with "+". (GH-8741)

The UTF-7 decoder now raises UnicodeDecodeError for ill-formed
sequences starting with "+" (as specified in RFC 2152).
---
 Objects/unicodeobject.c | 5 +++++
 1 file changed, 5 insertions(+)

(limited to 'Objects/unicodeobject.c')

diff --git a/Objects/unicodeobject.c b/Objects/unicodeobject.c
index 04fd6d03b4..0460d18493 100644
--- a/Objects/unicodeobject.c
+++ b/Objects/unicodeobject.c
@@ -4479,6 +4479,11 @@ PyUnicode_DecodeUTF7Stateful(const char *s,
                 if (_PyUnicodeWriter_WriteCharInline(&writer, '+') < 0)
                     goto onError;
             }
+            else if (s < e && !IS_BASE64(*s)) {
+                s++;
+                errmsg = "ill-formed sequence";
+                goto utf7Error;
+            }
             else { /* begin base64-encoded section */
                 inShift = 1;
                 surrogate = 0;
-- 
cgit v1.2.1