changeset:   90156:6a347c0ffbfc
user:        Victor Stinner <victor.stinner@gmail.com>
date:        Sat Apr 05 11:56:37 2014 +0200
files:       Lib/test/test_unicode.py Objects/unicodeobject.c
description:
Issue #21118: Add unit test for invalid character replacement (code point higher than U+10ffff)


diff -r cca6b056236a -r 6a347c0ffbfc Lib/test/test_unicode.py
--- a/Lib/test/test_unicode.py	Sat Apr 05 11:44:04 2014 +0200
+++ b/Lib/test/test_unicode.py	Sat Apr 05 11:56:37 2014 +0200
@@ -280,6 +280,14 @@
         self.assertEqual("[\xe9]".translate(str.maketrans({'\xe9': None})),
                          "[]")
 
+        # invalid Unicode characters
+        invalid_char = 0x10ffff+1
+        for before in "a\xe9\u20ac\U0010ffff":
+            mapping = str.maketrans({before: invalid_char})
+            text = "[%s]" % before
+            self.assertRaises(ValueError, text.translate, mapping)
+
+        # errors
         self.assertRaises(TypeError, self.type2test.maketrans)
         self.assertRaises(ValueError, self.type2test.maketrans, 'abc', 'defg')
         self.assertRaises(TypeError, self.type2test.maketrans, 2, 'def')
diff -r cca6b056236a -r 6a347c0ffbfc Objects/unicodeobject.c
--- a/Objects/unicodeobject.c	Sat Apr 05 11:44:04 2014 +0200
+++ b/Objects/unicodeobject.c	Sat Apr 05 11:56:37 2014 +0200
@@ -8473,10 +8473,10 @@
     }
     else if (PyLong_Check(x)) {
         long value = PyLong_AS_LONG(x);
-        long max = PyUnicode_GetMax();
-        if (value < 0 || value > max) {
-            PyErr_Format(PyExc_TypeError,
-                         "character mapping must be in range(0x%x)", max+1);
+        if (value < 0 || value > MAX_UNICODE) {
+            PyErr_Format(PyExc_ValueError,
+                         "character mapping must be in range(0x%x)",
+                         MAX_UNICODE+1);
             Py_DECREF(x);
             return -1;
         }
@@ -8522,7 +8522,9 @@
     }
 
     if (PyLong_Check(item)) {
-        Py_UCS4 ch = (Py_UCS4)PyLong_AS_LONG(item);
+        long ch = (Py_UCS4)PyLong_AS_LONG(item);
+        /* PyLong_AS_LONG() cannot fail, charmaptranslate_lookup() already
+           used it */
         if (_PyUnicodeWriter_WriteCharInline(writer, ch) < 0) {
             Py_DECREF(item);
             return -1;
@@ -8570,11 +8572,9 @@
 
     if (PyLong_Check(item)) {
         long replace = (Py_UCS4)PyLong_AS_LONG(item);
-        if (replace == -1) {
-            Py_DECREF(item);
-            return -1;
-        }
-        if (replace < 0 || 127 < replace) {
+        /* PyLong_AS_LONG() cannot fail, charmaptranslate_lookup() already
+           used it */
+        if (127 < replace) {
             /* invalid character or character outside ASCII:
                skip the fast translate */
             goto exit;