Merge pull request #274 from imankulov/fix-encoding-detection

Fix UnicodeEncodeError on file encoding detection

Author: akx

babel/util.py

@@ -65,7 +65,7 @@ def parse_encoding(fp):
             try:
                 import parser
                 parser.suite(line1.decode('latin-1'))
-            except (ImportError, SyntaxError):
+            except (ImportError, SyntaxError, UnicodeEncodeError):
                 # Either it's a real syntax error, in which case the source is
                 # not valid python source, or line2 is a continuation of line1,
                 # in which case we don't want to scan line2 for a magic

tests/test_util.py

@@ -14,6 +14,7 @@
 import unittest
 
 from babel import util
+from babel._compat import BytesIO
 
 
 def test_distinct():
@@ -52,3 +53,17 @@ def test_zone_zero_offset(self):
     def test_zone_positive_offset(self):
         self.assertEqual('Etc/GMT+330', util.FixedOffsetTimezone(330).zone)
 
+
+parse_encoding = lambda s: util.parse_encoding(BytesIO(s.encode('utf-8')))
+
+
+def test_parse_encoding_defined():
+    assert parse_encoding(u'# coding: utf-8') == 'utf-8'
+
+
+def test_parse_encoding_undefined():
+    assert parse_encoding(u'') is None
+
+
+def test_parse_encoding_non_ascii():
+    assert parse_encoding(u'K\xf6ln') is None