Felix Boehm
Felix Boehm
The reasoning behind this is that tags can be changed at any moment, to any commit. So right now there is an attack vector where an action is taken over,...
That's almost the answer: The reason why `decodeEntities` doesn't work is that it is an option for htmlparser2, which isn't the default parser. Have a look at https://github.com/cheeriojs/cheerio#using-htmlparser2 for more...
Closing this as I still don't understand when this is needed. Happy to revisit with some more context!
Duplicate of https://github.com/fb55/htmlparser2/issues/1237. See https://github.com/fb55/htmlparser2/issues/1237#issuecomment-1182522861 for some suggestions for how to work around this.
> We should probably just update this test to follow the Location type. Though maybe put something in it to identify it actually called our test callback rather than the...
Assuming this is resolved.
Interesting issue, and not easy to fix. Lowercasing all attribute names before the lookup is a valid workaround for now. Based on how the HTML tokenizer works, attribute names will...
It is a bit odd that this was already added to `html5lib-tests`. I will hold out on merging until the spec change has been merged.
@Kaleidea Thanks a lot for the context!
`InsertionMode` is actually exported: https://github.com/inikulin/parse5/blob/master/packages/parse5/lib/index.ts We mark it as internal, to avoid clutter in the documentation. But there is nothing stopping you from importing it.