Merge branch 'mina86-master'

htmlmin/escape.py

@@ -25,6 +25,8 @@
 SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 """
 
+import re
+
 try:
   from html import escape
 except ImportError:
@@ -45,7 +47,9 @@
 ZERO = ord('0')
 NINE = ord('9')
 
-chars_to_escape_re = re.compile(r'[=><`]')
+
+# https://www.w3.org/TR/html5/syntax.html#attributes-0
+CHARS_TO_QUOTE_RE = re.compile(u'[\x20\x09\x0a\x0c\x0d=><`]')
 
 def escape_tag(val):
   return escape(val)
@@ -58,16 +62,20 @@ def escape_attr_value(val, double_quote=False):
   has_html_tag = '<' in val or '>' in val
   if double_quote:
     return (val.replace('"', '&#34;'), DOUBLE_QUOTE)
-  if '"' in val or has_html_tag:
-    if "'" in val or has_html_tag:
-      return (val.replace('"', '&#34;'), DOUBLE_QUOTE)
-    else:
-      return (val, SINGLE_QUOTE)
-  elif "'" in val:
-    return (val, DOUBLE_QUOTE)
 
-  if (not val or any((c.isspace() for c in val)) or 
-      chars_to_escape_re.search(val)):
+  double_quote_count = 0
+  single_quote_count = 0
+  for ch in val:
+    if ch == '"':
+      double_quote_count += 1
+    elif ch == "'":
+      single_quote_count += 1
+  if double_quote_count > single_quote_count:
+    return (val.replace("'", '&#39;'), SINGLE_QUOTE)
+  elif single_quote_count:
+    return (val.replace('"', '&#34;'), DOUBLE_QUOTE)
+
+  if not val or CHARS_TO_QUOTE_RE.search(val):
     return (val, DOUBLE_QUOTE)
   return (val, NO_QUOTES)
 

htmlmin/main.py

@@ -38,7 +38,8 @@ def minify(input,
            remove_optional_attribute_quotes=True,
            keep_pre=False,
            pre_tags=parser.PRE_TAGS,
-           pre_attr='pre'):
+           pre_attr='pre',
+           cls=parser.HTMLMinParser):
   """Minifies HTML in one shot.
 
   :param input: A string containing the HTML to be minified.
@@ -84,7 +85,7 @@ def minify(input,
   If you are going to be minifying multiple HTML documents, each with the same
   settings, consider using :class:`.Minifier`.
   """
-  minifier = parser.HTMLMinParser(
+  minifier = cls(
       remove_comments=remove_comments,
       remove_empty_space=remove_empty_space,
       remove_all_empty_space=remove_all_empty_space,
@@ -118,12 +119,13 @@ def __init__(self,
                remove_optional_attribute_quotes=True,
                keep_pre=False,
                pre_tags=parser.PRE_TAGS,
-               pre_attr='pre'):
+               pre_attr='pre',
+               cls=parser.HTMLMinParser):
     """Initialize the Minifier.
 
     See :class:`htmlmin.minify` for an explanation of options.
     """
-    self._parser = parser.HTMLMinParser(
+    self._parser = cls(
       remove_comments=remove_comments,
       remove_empty_space=remove_empty_space,
       remove_all_empty_space=remove_all_empty_space,
@@ -180,4 +182,3 @@ def finalize(self):
     result = self._parser.result
     self._parser.reset()
     return result
-