From 2c626a37d1481035019162e951c748cb854696c7 Mon Sep 17 00:00:00 2001
From: Serhiy Storchaka <storchaka@gmail.com>
Date: Thu, 15 May 2014 14:37:42 +0300
Subject: Issue #13916: Disallowed the surrogatepass error handler for non
 UTF-* encodings.

---
 Python/codecs.c | 23 +++++++++++++++++++----
 1 file changed, 19 insertions(+), 4 deletions(-)

(limited to 'Python/codecs.c')

diff --git a/Python/codecs.c b/Python/codecs.c
index e06d6e0922..700313633e 100644
--- a/Python/codecs.c
+++ b/Python/codecs.c
@@ -901,6 +901,7 @@ PyObject *PyCodec_BackslashReplaceErrors(PyObject *exc)
     }
 }
 
+#define ENC_UNKNOWN     -1
 #define ENC_UTF8        0
 #define ENC_UTF16BE     1
 #define ENC_UTF16LE     2
@@ -916,7 +917,11 @@ get_standard_encoding(const char *encoding, int *bytelength)
         encoding += 3;
         if (*encoding == '-' || *encoding == '_' )
             encoding++;
-        if (encoding[0] == '1' && encoding[1] == '6') {
+        if (encoding[0] == '8' && encoding[1] == '\0') {
+            *bytelength = 3;
+            return ENC_UTF8;
+        }
+        else if (encoding[0] == '1' && encoding[1] == '6') {
             encoding += 2;
             *bytelength = 2;
             if (*encoding == '\0') {
@@ -955,9 +960,7 @@ get_standard_encoding(const char *encoding, int *bytelength)
             }
         }
     }
-    /* utf-8 */
-    *bytelength = 3;
-    return ENC_UTF8;
+    return ENC_UNKNOWN;
 }
 
 /* This handler is declared static until someone demonstrates
@@ -994,6 +997,12 @@ PyCodec_SurrogatePassErrors(PyObject *exc)
         }
         code = get_standard_encoding(encoding, &bytelength);
         Py_DECREF(encode);
+        if (code == ENC_UNKNOWN) {
+            /* Not supported, fail with original exception */
+            PyErr_SetObject(PyExceptionInstance_Class(exc), exc);
+            Py_DECREF(object);
+            return NULL;
+        }
 
         res = PyBytes_FromStringAndSize(NULL, bytelength*(end-start));
         if (!res) {
@@ -1068,6 +1077,12 @@ PyCodec_SurrogatePassErrors(PyObject *exc)
         }
         code = get_standard_encoding(encoding, &bytelength);
         Py_DECREF(encode);
+        if (code == ENC_UNKNOWN) {
+            /* Not supported, fail with original exception */
+            PyErr_SetObject(PyExceptionInstance_Class(exc), exc);
+            Py_DECREF(object);
+            return NULL;
+        }
 
         /* Try decoding a single surrogate character. If
            there are more, let the codec call us again. */
-- 
cgit v1.2.1


From 569a55317259d71ebde1c85ef48874eba3824a5e Mon Sep 17 00:00:00 2001
From: Victor Stinner <victor.stinner@gmail.com>
Date: Fri, 16 May 2014 14:46:20 +0200
Subject: Issue #13916: Fix surrogatepass error handler on Windows

---
 Python/codecs.c | 4 ++++
 1 file changed, 4 insertions(+)

(limited to 'Python/codecs.c')

diff --git a/Python/codecs.c b/Python/codecs.c
index 700313633e..4c2ae381b3 100644
--- a/Python/codecs.c
+++ b/Python/codecs.c
@@ -960,6 +960,10 @@ get_standard_encoding(const char *encoding, int *bytelength)
             }
         }
     }
+    else if (strcmp(encoding, "CP_UTF8") == 0) {
+        *bytelength = 3;
+        return ENC_UTF8;
+    }
     return ENC_UNKNOWN;
 }
 
-- 
cgit v1.2.1