parse5 serialization on text nodes and removal of some htmlEntities conversion #107
Conversation
|
|
||
| var dummyNode = parser.parse(' '); | ||
| dummyNode.childNodes = [dom]; | ||
| var contents = serializer.serialize(dummyNode); |
There was a problem hiding this comment.
This isn't using the new parse5 serialization option to not escape the string as I expected. Wouldn't that remove the need for your string replacement below?
|
The option is enabled by default, and with regards to the string replacements, for single quotes, and double quotes, it's somewhat impossible for anyone really to know if they should be converted. Since the parser of parse5 converts escape sequences to what they are meant to represent, that means that: ', ', and &(r/l)squote; all map to ' in text This means, that if we see ' in text, we have no idea which it could be, and what it should be, since all 3 forms are valid (additionally, what if the users didn't want a fancy single quote) There is no spec for whether they must be escaped, so, the serializer doesn't convert them. Thus, if we want to escape them, we must do it ourselves, if you are saying that we shouldn't escape them; by all means, I see no problem with this, but we can't read the user's mind so long as the parser is doing what it wants to the document |
closes #100