Use dict on Py >= 3.7 for attributes now it is ordered

gsnedders · gsnedders · commit 0db23a35d842 · 2020-06-17T20:18:04.000+01:00
diff --git a/html5lib/html5parser.py b/html5lib/html5parser.py
@@ -2,7 +2,9 @@
 from six import with_metaclass, viewkeys
 
 import types
+
 from collections import OrderedDict
+from sys import version_info
 
 from . import _inputstream
 from . import _tokenizer
@@ -24,6 +26,12 @@
 )
 
 
+if version_info >= (3, 7):
+    attributeMap = dict
+else:
+    attributeMap = OrderedDict
+
+
 def parse(doc, treebuilder="etree", namespaceHTMLElements=True, **kwargs):
     """Parse an HTML document as a string or file-like object into a tree
 
@@ -329,10 +337,11 @@ def normalizeToken(self, token):
         # HTML5 specific normalizations to the token stream
         if token["type"] == tokenTypes["StartTag"]:
             raw = token["data"]
-            token["data"] = OrderedDict(raw)
-            if len(raw) > len(token["data"]):
+            data = attributeMap(raw)
+            if len(raw) > len(data):
                 # we had some duplicated attribute, fix so first wins
-                token["data"].update(raw[::-1])
+                data.update(raw[::-1])
+            token["data"] = data
 
         return token
 
@@ -2770,8 +2779,8 @@ def processEndTag(self, token):
 def adjust_attributes(token, replacements):
     needs_adjustment = viewkeys(token['data']) & viewkeys(replacements)
     if needs_adjustment:
-        token['data'] = OrderedDict((replacements.get(k, k), v)
-                                    for k, v in token['data'].items())
+        token['data'] = attributeMap((replacements.get(k, k), v)
+                                     for k, v in token['data'].items())
 
 
 def impliedTagToken(name, type="EndTag", attributes=None,
