diff -r 85c04fdaa404 Lib/test/test_pep263.py
--- a/Lib/test/test_pep263.py	Wed May 22 15:28:30 2013 +0300
+++ b/Lib/test/test_pep263.py	Fri May 24 12:54:04 2013 +0300
@@ -41,6 +41,27 @@
         # two bytes in common with the UTF-8 BOM
         self.assertRaises(SyntaxError, eval, '\xef\xbb\x20')
 
+    def test_error_message(self):
+        compile('# -*- coding: iso-8859-15 -*-\n', 'dummy', 'exec')
+        compile('\xef\xbb\xbf\n', 'dummy', 'exec')
+        compile('\xef\xbb\xbf# -*- coding: utf-8 -*-\n', 'dummy', 'exec')
+        self.assertRaisesRegexp(SyntaxError, 'fake', compile,
+                                '# -*- coding: fake -*-\n',
+                                'dummy', 'exec')
+        self.assertRaisesRegexp(SyntaxError, 'iso-8859-15', compile,
+                                '\xef\xbb\xbf# -*- coding: iso-8859-15 -*-\n',
+                                'dummy', 'exec')
+        self.assertRaisesRegexp(SyntaxError, 'BOM', compile,
+                                '\xef\xbb\xbf# -*- coding: iso-8859-15 -*-\n',
+                                'dummy', 'exec')
+        self.assertRaisesRegexp(SyntaxError, 'fake', compile,
+                                '\xef\xbb\xbf# -*- coding: fake -*-\n',
+                                'dummy', 'exec')
+        self.assertRaisesRegexp(SyntaxError, 'BOM', compile,
+                                '\xef\xbb\xbf# -*- coding: fake -*-\n',
+                                'dummy', 'exec')
+
+
 def test_main():
     test_support.run_unittest(PEP263Test)
 
diff -r 85c04fdaa404 Parser/tokenizer.c
--- a/Parser/tokenizer.c	Wed May 22 15:28:30 2013 +0300
+++ b/Parser/tokenizer.c	Fri May 24 12:54:04 2013 +0300
@@ -277,8 +277,11 @@
                     tok->encoding = cs;
                     tok->decoding_state = -1;
                 }
-                else
+                else {
+                    PyErr_Format(PyExc_SyntaxError,
+                                 "encoding problem: %s", cs);
                     PyMem_FREE(cs);
+                }
 #else
                 /* Without Unicode support, we cannot
                    process the coding spec. Since there
@@ -289,15 +292,12 @@
             }
         } else {                /* then, compare cs with BOM */
             r = (strcmp(tok->encoding, cs) == 0);
+            if (!r)
+                PyErr_Format(PyExc_SyntaxError,
+                             "encoding problem: %s with BOM", cs);
             PyMem_FREE(cs);
         }
     }
-    if (!r) {
-        cs = tok->encoding;
-        if (!cs)
-            cs = "with BOM";
-        PyErr_Format(PyExc_SyntaxError, "encoding problem: %s", cs);
-    }
     return r;
 }