Merge pull request #264 from PyCQA/feature/unicode-bug-fix-258

Fix bug where snowballstemmer chokes on Unicode word

docs/release_notes.rst

@@ -17,6 +17,9 @@ Bug Fixes
 * Fixed an issue where the ``--source`` flag would result in improperly
   spaced output (#256, #257, #260).
+* Fixed an issue where if a first word in a docstring had Unicode characters
+  and the docstring was not a unicode string, an exception would be raised
+  (#258, #264).
 2.0.0 - April 18th, 2017

setup.py

 from __future__ import with_statement
-import os
 from setuptools import setup
 # Do not update the version manually - it is managed by `bumpversion`.

src/pydocstyle/checker.py

@@ -372,7 +372,14 @@ class ConventionChecker(object):
                 if check_word in IMPERATIVE_BLACKLIST:
                     return violations.D401b(first_word)
-                correct_form = IMPERATIVE_VERBS.get(stem(check_word))
+                try:
+                    correct_form = IMPERATIVE_VERBS.get(stem(check_word))
+                except UnicodeDecodeError:
+                    # This is raised when the docstring contains unicode
+                    # characters in the first word, but is not a unicode
+                    # string. In which case D302 will be reported. Ignoring.
+                    return
+
                 if correct_form and correct_form != check_word:
                     return violations.D401(
                         correct_form.capitalize(),

src/tests/test_cases/test.py

@@ -271,6 +271,10 @@ if sys.version_info[0] <= 2:
     def unicode_unmarked():
         """Юникод."""
+    @expect('D302: Use u""" for Unicode docstrings')
+    def first_word_has_unicode_byte():
+        """あy."""
+
 @expect("D400: First line should end with a period (not 'y')")
 def lwnlkjl():
@@ -376,5 +380,6 @@ def bad_decorated_function():
     """Bad (E501) but decorated"""
     pass
+
 expect(os.path.normcase(__file__ if __file__[-1] != 'c' else __file__[:-1]),
        'D100: Missing docstring in public module')

tox.ini

@@ -11,7 +11,7 @@ envlist = py27, py33, py34, py35, pypy, docs
 setenv =
     LANG=C
     LC_ALL=C
-commands = py.test --pep8 --cache-clear -v src/tests
+commands = py.test --pep8 --cache-clear -vv src/tests
 deps =
     -rrequirements/runtime.txt
     -rrequirements/tests.txt