Make pseudo-elements lower-case in the ASCII range.

See http://www.w3.org/TR/selectors/#casesens

Pseudo-classes were already case-insensitive, but the
lower-casing was moved to the parser.

cssselect/parser.py

@@ -25,6 +25,11 @@
     _unichr = chr
 
 
+def ascii_lower(string):
+    """Lower-case, but only in the ASCII range."""
+    return string.encode('utf8').lower().decode('utf8')
+
+
 class SelectorError(Exception):
     """Common parent for :class:`SelectorSyntaxError` and
     :class:`ExpressionError`.
@@ -52,6 +57,8 @@ class Selector(object):
     """
     def __init__(self, tree, pseudo_element=None):
         self.parsed_tree = tree
+        if pseudo_element is not None:
+            pseudo_element = ascii_lower(pseudo_element)
         #: The identifier for the pseudo-element as a string, or ``None``.
         #:
         #: +-------------------------+----------------+----------------+
@@ -114,7 +121,7 @@ class Function(object):
     """
     def __init__(self, selector, name, arguments):
         self.selector = selector
-        self.name = name
+        self.name = ascii_lower(name)
         self.arguments = arguments
 
     def __repr__(self):
@@ -137,7 +144,7 @@ class Pseudo(object):
     """
     def __init__(self, selector, ident):
         self.selector = selector
-        self.ident = ident
+        self.ident = ascii_lower(ident)
 
     def __repr__(self):
         return '%s[%r:%s]' % (
@@ -393,7 +400,8 @@ def parse_simple_selector(stream, inside_negation=False):
                 pseudo_element = stream.next_ident()
                 continue
             ident = stream.next_ident()
-            if ident in ('first-line', 'first-letter', 'before', 'after'):
+            if ident.lower() in ('first-line', 'first-letter',
+                                 'before', 'after'):
                 # Special case: CSS 2.1 pseudo-elements can have a single ':'
                 # Any new pseudo-element must have two.
                 pseudo_element = _unicode(ident)

cssselect/tests.py

@@ -167,17 +167,17 @@ def parse_one(css):
         assert parse_one(':empty') == ('Pseudo[Element[*]:empty]', None)
 
         # Special cases for CSS 2.1 pseudo-elements
-        assert parse_one(':before') == ('Element[*]', 'before')
-        assert parse_one(':after') == ('Element[*]', 'after')
-        assert parse_one(':first-line') == ('Element[*]', 'first-line')
-        assert parse_one(':first-letter') == ('Element[*]', 'first-letter')
+        assert parse_one(':BEfore') == ('Element[*]', 'before')
+        assert parse_one(':aftER') == ('Element[*]', 'after')
+        assert parse_one(':First-Line') == ('Element[*]', 'first-line')
+        assert parse_one(':First-Letter') == ('Element[*]', 'first-letter')
 
-        assert parse_one('::before') == ('Element[*]', 'before')
-        assert parse_one('::after') == ('Element[*]', 'after')
-        assert parse_one('::first-line') == ('Element[*]', 'first-line')
-        assert parse_one('::first-letter') == ('Element[*]', 'first-letter')
+        assert parse_one('::befoRE') == ('Element[*]', 'before')
+        assert parse_one('::AFter') == ('Element[*]', 'after')
+        assert parse_one('::firsT-linE') == ('Element[*]', 'first-line')
+        assert parse_one('::firsT-letteR') == ('Element[*]', 'first-letter')
 
-        assert parse_one('::selection') == ('Element[*]', 'selection')
+        assert parse_one('::Selection') == ('Element[*]', 'selection')
         assert parse_one('foo:after') == ('Element[foo]', 'after')
         assert parse_one('foo::selection') == ('Element[foo]', 'selection')
         assert parse_one('lorem#ipsum ~ a#b.c[href]:empty::selection') == (
@@ -346,13 +346,15 @@ def xpath(css):
             "e[last() = 1]")
         assert xpath('e:empty') == (
             "e[not(*) and not(normalize-space())]")
+        assert xpath('e:EmPTY') == (
+            "e[not(*) and not(normalize-space())]")
         assert xpath('e:root') == (
             "e[not(parent::*)]")
         assert xpath('e:hover') == (
             "e[0]")  # never matches
         assert xpath('e:contains("foo")') == (
             "e[contains(string(.), 'foo')]")
-        assert xpath('e:contains(foo)') == (
+        assert xpath('e:ConTains(foo)') == (
             "e[contains(string(.), 'foo')]")
         assert xpath('e.warning') == (
             "e[@class and contains("
@@ -361,7 +363,7 @@ def xpath(css):
             "e[@id = 'myid']")
         assert xpath('e:not(:nth-child(odd))') == (
             "e[not((position() -1) mod 2 = 0 and position() >= 1)]")
-        assert xpath('e:not(*)') == (
+        assert xpath('e:nOT(*)') == (
             "e[0]")  # never matches
         assert xpath('e f') == (
             "e/descendant-or-self::*/f")

cssselect/xpath.py

@@ -234,7 +234,7 @@ def xpath_negation(self, negation):
 
     def xpath_function(self, function):
         """Translate a functional pseudo-class."""
-        method = 'xpath_%s_function' % function.name.replace('-', '_').lower()
+        method = 'xpath_%s_function' % function.name.replace('-', '_')
         method = getattr(self, method, None)
         if not method:
             raise ExpressionError(
@@ -243,7 +243,7 @@ def xpath_function(self, function):
 
     def xpath_pseudo(self, pseudo):
         """Translate a pseudo-class."""
-        method = 'xpath_%s_pseudo' % pseudo.ident.replace('-', '_').lower()
+        method = 'xpath_%s_pseudo' % pseudo.ident.replace('-', '_')
         method = getattr(self, method, None)
         if not method:
             # TODO: better error message for pseudo-elements?