commit c7797dc7482035ee166ca2e941b623382b92e1fc
author Serhiy Storchaka <storchaka@gmail.com> Sun May 31 20:21:00 2015 +0300
committer Serhiy Storchaka <storchaka@gmail.com> Sun May 31 20:21:00 2015 +0300
tree 526e26fa4dac506f02859fdbe946d33ed4165f5e
parent cfb7028df4bdf12325786e48ebef3b4982efa119

Issue #19543: Emit deprecation warning for known non-text encodings.

Backported issues #19619: encode() and decode() methods and constructors
of str, unicode and bytearray classes now emit deprecation warning for known
non-text encodings when Python is ran with the -3 option.

Backported issues #20404: io.TextIOWrapper (and hence io.open()) now uses the
internal codec marking system added to emit deprecation warning for known non-text
encodings at stream construction time when Python is ran with the -3 option.

Modules/_io/textio.c

diff --git a/Modules/_io/textio.c b/Modules/_io/textio.c
index 8ac8a4a..9981d4c 100644
--- a/Modules/_io/textio.c
+++ b/Modules/_io/textio.c
@@ -826,7 +826,7 @@
     char *kwlist[] = {"buffer", "encoding", "errors",
                       "newline", "line_buffering",
                       NULL};
-    PyObject *buffer, *raw;
+    PyObject *buffer, *raw, *codec_info = NULL;
     char *encoding = NULL;
     char *errors = NULL;
     char *newline = NULL;
@@ -909,6 +909,17 @@
                         "could not determine default encoding");
     }
 
+    /* Check we have been asked for a real text encoding */
+    codec_info = _PyCodec_LookupTextEncoding(encoding, "codecs.open()");
+    if (codec_info == NULL) {
+        Py_CLEAR(self->encoding);
+        goto error;
+    }
+
+    /* XXX: Failures beyond this point have the potential to leak elements
+     * of the partially constructed object (like self->encoding)
+     */
+
     if (errors == NULL)
         errors = "strict";
     self->errors = PyBytes_FromString(errors);
@@ -922,7 +933,7 @@
     if (newline) {
         self->readnl = PyString_FromString(newline);
         if (self->readnl == NULL)
-            return -1;
+            goto error;
     }
     self->writetranslate = (newline == NULL || newline[0] != '\0');
     if (!self->readuniversal && self->writetranslate) {
@@ -944,8 +955,8 @@
     if (r == -1)
         goto error;
     if (r == 1) {
-        self->decoder = PyCodec_IncrementalDecoder(
-            encoding, errors);
+        self->decoder = _PyCodecInfo_GetIncrementalDecoder(codec_info,
+                                                           errors);
         if (self->decoder == NULL)
             goto error;
 
@@ -969,17 +980,12 @@
     if (r == -1)
         goto error;
     if (r == 1) {
-        PyObject *ci;
-        self->encoder = PyCodec_IncrementalEncoder(
-            encoding, errors);
+        self->encoder = _PyCodecInfo_GetIncrementalEncoder(codec_info,
+                                                           errors);
         if (self->encoder == NULL)
             goto error;
         /* Get the normalized named of the codec */
-        ci = _PyCodec_Lookup(encoding);
-        if (ci == NULL)
-            goto error;
-        res = PyObject_GetAttrString(ci, "name");
-        Py_DECREF(ci);
+        res = PyObject_GetAttrString(codec_info, "name");
         if (res == NULL) {
             if (PyErr_ExceptionMatches(PyExc_AttributeError))
                 PyErr_Clear();
@@ -999,6 +1005,9 @@
         Py_XDECREF(res);
     }
 
+    /* Finished sorting out the codec details */
+    Py_DECREF(codec_info);
+
     self->buffer = buffer;
     Py_INCREF(buffer);
 
@@ -1059,6 +1068,7 @@
     return 0;
 
   error:
+    Py_XDECREF(codec_info);
     return -1;
 }