HTMLSerializer and HTML entities

[danielknell]

>>> html5lib.serializer.serialize(html5lib.parse('<p>&nbsp;</p>'))
'<p>\xa0'

at the moment the parsing and serialising a document causes entities to be converted into special characters, including things like #00 and there is no way to pass additional entities to xml.sax.saxutils.escape.

I looked into subclassing the serialiser but the escaping happens in the middle of the serialize() method at:

https://github.com/html5lib/html5lib-python/blob/master/html5lib/serializer/htmlserializer.py#L223

perhaps the class should define an entities dict to pass through the standard html5 entities and special characters or do the escaping via a class method that can be overridden?

[gsnedders]

What's your use-case for wanting them escaped?

[gsnedders]

Basically, you might want to encode it as an ASCII string (which will obviously escape all non-ASCII characters), or is something like #38 more along the lines of what you want?

[danielknell]

I am building a html document using etree Element instances, and serialising with html5lib, special characters (especially things like the non breaking space) will be non-obvious to anyone viewing the rendered source.

[gsnedders]

Okay, that makes it sound like #38 is more or less what you want — readability of the source given "invisible" characters. Closing this as a dupe of that, then. (If you disagree, feel free to reopen this and comment.)