Merge pull request #1886 from getnikola/fix-1885

Kwpolska · Kwpolska · commit f390f364a104 · 2015-07-11T18:46:54.000+02:00
Fix #1885 -- always return unicode in slugify
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
@@ -729,9 +729,9 @@ 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,
@@ -741,7 +741,7 @@ def slugify(value, force=False):
         # We still replace some characters, though.  In particular, we need
         # to replace ? and #, which should not appear in URLs, and some
         # Windows-unsafe characters.  This list might be even longer.
-        rc = '/\\?#"\'\r\n\t*:<>|"'
+        rc = '/\\?#"\'\r\n\t*:<>|'
 
         for c in rc:
             value = value.replace(c, '-')
diff --git a/tests/test_slugify.py b/tests/test_slugify.py
@@ -0,0 +1,65 @@
+# -*- coding: utf-8 -*-
+
+u"""Test slugify."""
+
+from __future__ import unicode_literals
+import nikola.utils
+
+
+def test_ascii():
+    """Test an ASCII-only string."""
+    o = nikola.utils.slugify(u'hello')
+    assert o == u'hello'
+    assert isinstance(o, nikola.utils.unicode_str)
+
+
+def test_ascii_dash():
+    """Test an ASCII string, with dashes."""
+    o = nikola.utils.slugify(u'hello-world')
+    assert o == u'hello-world'
+    assert isinstance(o, nikola.utils.unicode_str)
+
+
+def test_ascii_fancy():
+    """Test an ASCII string, with fancy characters."""
+    o = nikola.utils.slugify(u'The quick brown fox jumps over the lazy dog!-123.456')
+    assert o == u'the-quick-brown-fox-jumps-over-the-lazy-dog-123456'
+    assert isinstance(o, nikola.utils.unicode_str)
+
+
+def test_pl():
+    """Test a string with Polish diacritical characters."""
+    o = nikola.utils.slugify(u'zażółćgęśląjaźń')
+    assert o == u'zazolcgeslajazn'
+    assert isinstance(o, nikola.utils.unicode_str)
+
+
+def test_pl_dash():
+    """Test a string with Polish diacritical characters and dashes."""
+    o = nikola.utils.slugify(u'zażółć-gęślą-jaźń')
+    assert o == u'zazolc-gesla-jazn'
+
+
+def test_pl_fancy():
+    """Test a string with Polish diacritical characters and fancy characters."""
+    o = nikola.utils.slugify(u'Zażółć gęślą jaźń!-123.456')
+    assert o == u'zazolc-gesla-jazn-123456'
+    assert isinstance(o, nikola.utils.unicode_str)
+
+
+def test_disarmed():
+    """Test disarmed slugify."""
+    nikola.utils.USE_SLUGIFY = False
+    o = nikola.utils.slugify(u'Zażółć gęślą jaźń!-123.456')
+    assert o == u'Zażółć gęślą jaźń!-123.456'
+    assert isinstance(o, nikola.utils.unicode_str)
+    nikola.utils.USE_SLUGIFY = True
+
+
+def test_disarmed_weird():
+    """Test disarmed slugify with banned characters."""
+    nikola.utils.USE_SLUGIFY = False
+    o = nikola.utils.slugify(u'Zażółć gęślą jaźń!-123.456 "Hello World"?#H<e>l/l\\o:W\'o\rr*l\td|!\n')
+    assert o == u'Zażółć gęślą jaźń!-123.456 -Hello World---H-e-l-l-o-W-o-r-l-d-!-'
+    assert isinstance(o, nikola.utils.unicode_str)
+    nikola.utils.USE_SLUGIFY = True
