Fix #1885 -- always return unicode in slugify

Signed-off-by: Chris Warrick <kwpolska@gmail.com>
getnikola · Jul 11, 2015 · 50f92ad · 50f92ad
1 parent c8d97e2
commit 50f92ad
Show file tree

Hide file tree

Showing 3 changed files with 40 additions and 3 deletions.
diff --git a/CHANGES.txt b/CHANGES.txt
@@ -4,6 +4,7 @@ New in master
 Features
 --------
 
+* Always return unicode in slugify (Issue #1885)
 * Remove logging handlers (Issue #1797)
 * Add ``-d``, ``--detach`` option to ``nikola serve`` (Issue #1871)
 * Use provided teaser format (``*_READ_MORE_LINK``) with custom teaser text

diff --git a/nikola/utils.py b/nikola/utils.py
@@ -728,9 +728,9 @@ def slugify(value, force=False):
     if USE_SLUGIFY or force:
         # This is the standard state of slugify, which actually does some work.
         # It is the preferred style, especially for Western languages.
-        value = unidecode(value)
-        value = str(_slugify_strip_re.sub('', value).strip().lower())
-        return _slugify_hyphenate_re.sub('-', value)
+        value = unicode_str(unidecode(value))
+        value = _slugify_strip_re.sub('', value, re.UNICODE).strip().lower()
+        return _slugify_hyphenate_re.sub('-', value, re.UNICODE)
     else:
         # This is the “disarmed” state of slugify, which lets the user
         # have any character they please (be it regular ASCII with spaces,

diff --git a/tests/test_slugify.py b/tests/test_slugify.py
@@ -0,0 +1,36 @@
+# -*- coding: utf-8 -*-
+import nikola.utils
+
+def test_ascii():
+    o = nikola.utils.slugify(u'abcdef')
+    assert o == u'abcdef'
+    assert isinstance(o, nikola.utils.unicode_str)
+
+def test_ascii_dash():
+    o = nikola.utils.slugify(u'abc-def')
+    assert o == u'abc-def'
+    assert isinstance(o, nikola.utils.unicode_str)
+
+def test_pl():
+    o = nikola.utils.slugify(u'ąbćdef')
+    assert o == u'abcdef'
+    assert isinstance(o, nikola.utils.unicode_str)
+
+def test_pl_dash():
+    o = nikola.utils.slugify(u'ąbć-def')
+    assert o == u'abc-def'
+    assert isinstance(o, nikola.utils.unicode_str)
+
+def test_disarmed():
+    nikola.utils.USE_SLUGIFY = False
+    o = nikola.utils.slugify(u'ąbć-def')
+    assert o == u'ąbć-def'
+    assert isinstance(o, nikola.utils.unicode_str)
+    nikola.utils.USE_SLUGIFY = True
+
+def test_disarmed_weird():
+    nikola.utils.USE_SLUGIFY = False
+    o = nikola.utils.slugify(u'ąbć-def "Hello World"?#H<e>l/l\\o:W\'o\rr*l\td|!\n')
+    assert o == u'ąbć-def -Hello World---H-e-l-l-o-W-o-r-l-d-!-'
+    assert isinstance(o, nikola.utils.unicode_str)
+    nikola.utils.USE_SLUGIFY = True