Questions: implementation + html parser adapter about litehtml HOT 4 OPEN

1. am I right in thinking that gumbo is only used for html parsing, and this lib leverages it by in turn taking care of css (computing rendering positions/size/clipping/etc.)?

Yes, gumbo is used to parse html only. CSS are parsed by litehtml itself.

1. would you think it would be feasible to adapt the library to use something like BeautifulSoup/JSoup for the html parsing instead of gumbo?

Hmm... this is interesting. I guess this is not a big problem. It is simple to build elements tree from other parser. I'll add this into todo-list 👍

1. would you mind if I (try to) port the lib to another language (obviously crediting you and including your license)?

I'll be glad to see ports to another languages.

from litehtml.

Great! 👍
And nice to hear it sounds interesting to you too.

from litehtml.

And please post here if you make any progress in that direction.

from litehtml.

Hi there, and sorry for bumping this up. Just a couple more questions/observations (which really are feature requests :D).

• I can't stress enough how much I like the render-agnostic property of this lib, so what I'd hope for is to have two more interfaces (like litehtml::document_container): one for handling html (say parseHtml(string)), and one for parsing/applying css (say computeStyle(style, doc))
• I get it could be cumbersome to do (as this may also need an intermediate representation), but I still think that separating the processes could make things easier to work with (and also make it possible to switch libraries for the specific needs)

I know I'm speaking out of my direct needs/hopes, but do these requests seem sensible to you and addressable in the nearest future?

PS: maybe this needs to be a separate issue?!
PPS: main reason for asking this is that there are only a few libs out there (in the opensource zone), that return layout props from css, although many offer to parse html and/or css (with no layout). So mixing&matching them might be very interesting (if only for benchmarking purposes and a common task).

Thanks.

from litehtml.