commit e822b034e766e03cd8fbe7ab52fbc2d46fff6d33
author Serhiy Storchaka <storchaka@gmail.com> Tue Aug 06 16:56:26 2013 +0300
committer Serhiy Storchaka <storchaka@gmail.com> Tue Aug 06 16:56:26 2013 +0300
tree a2e79671034ae47c4a0d6d0d030aecda007c8123
parent 5ad3514822a80c094f1cfe47ae59450001043482

Issue #15866: The xmlcharrefreplace error handler no more produces two XML
entities for a non-BMP character on narrow build.

Lib/test/test_codeccallbacks.py

diff --git a/Lib/test/test_codeccallbacks.py b/Lib/test/test_codeccallbacks.py
index 61c2df2..ecaf997 100644
--- a/Lib/test/test_codeccallbacks.py
+++ b/Lib/test/test_codeccallbacks.py
@@ -66,15 +66,34 @@
         # replace unencodable characters which numeric character entities.
         # For ascii, latin-1 and charmaps this is completely implemented
         # in C and should be reasonably fast.
-        s = u"\u30b9\u30d1\u30e2 \xe4nd eggs"
+        s = u"\u30b9\u30d1\u30e2 \xe4nd egg\u0161"
         self.assertEqual(
             s.encode("ascii", "xmlcharrefreplace"),
-            "スパモ änd eggs"
+            "スパモ änd eggš"
         )
         self.assertEqual(
             s.encode("latin-1", "xmlcharrefreplace"),
-            "スパモ \xe4nd eggs"
+            "スパモ \xe4nd eggš"
         )
+        self.assertEqual(
+            s.encode("iso-8859-15", "xmlcharrefreplace"),
+            "スパモ \xe4nd egg\xa8"
+        )
+
+    def test_xmlcharrefreplace_with_surrogates(self):
+        tests = [(u'\U0001f49d', '💝'),
+                 (u'\ud83d', '�'),
+                 (u'\udc9d', '�'),
+                 (u'\ud83d\udc9d', '💝' if len(u'\U0001f49d') > 1 else
+                                   '��'),
+                ]
+        for encoding in ['ascii', 'latin1', 'iso-8859-15']:
+            for s, exp in tests:
+                self.assertEqual(s.encode(encoding, 'xmlcharrefreplace'),
+                                 exp, msg='%r.encode(%r)' % (s, encoding))
+                self.assertEqual((s+'X').encode(encoding, 'xmlcharrefreplace'),
+                                 exp+'X',
+                                 msg='%r.encode(%r)' % (s + 'X', encoding))
 
     def test_xmlcharnamereplace(self):
         # This time use a named character entity for unencodable

Lib/test/test_unicode.py

diff --git a/Lib/test/test_unicode.py b/Lib/test/test_unicode.py
index e44fe03..666cab8 100644
--- a/Lib/test/test_unicode.py
+++ b/Lib/test/test_unicode.py
@@ -1658,6 +1658,18 @@
         self.assertEqual(unicode_encodedecimal(u"123\u20ac\u0660", "replace"),
                          b'123?0')
 
+    def test_encode_decimal_with_surrogates(self):
+        from _testcapi import unicode_encodedecimal
+        tests = [(u'\U0001f49d', '💝'),
+                 (u'\ud83d', '�'),
+                 (u'\udc9d', '�'),
+                 (u'\ud83d\udc9d', '💝' if len(u'\U0001f49d') > 1 else
+                                  '��'),
+                ]
+        for s, exp in tests:
+            self.assertEqual(
+                    unicode_encodedecimal(u"123" + s, "xmlcharrefreplace"),
+                    '123' + exp)
 
 def test_main():
     test_support.run_unittest(__name__)