html2haml won't work with the new elements in html5 because hpricot is very conservative about content models. hpricot is in maintenance mode and they may not accept a patch to make it more liberal. In any case, since hpricot development and suport is limited at best moving to something under active development/support like nokogiri would be nice.
Even if the patch goes into hpricot, it'd be a good idea to move to (something like) nokogiri.
I don't have time to do this right now so I'm posting more for posterity and so others will find something when they have problems with html2haml on html5.
+1 to this.
I agree we should migrate away from Hpricot. This will be on the roadmap for 3.2; though I'm not 100% sure we should use Nokogiri. Html2haml doesn't necessarily need to be fast, and we may want to use an XML library with no native dependencies to facilitate usage on other platforms. I'm aware that Nokogiri has been ported to JRuby and that binary gems for Windows are available, but I think it's worth looking into avoiding a native dependency if possible so that we can reliably depend on it in the gemspec and make html2haml work out of the box.
Nokogiri isn't good just because of the speed, but libxml is one of the more forgiving and standard ways to interpret possibly flawed XML data. I use it extensively and would recommend it.
In my opinion, html2haml should probably be its own project.
@hcatlin yeah, splitting it off into its own gem would be helpful. I'll see if I can do that today.
@norman thanks for your work! :)
@norman if we do that, we can include anything we want... dependencies be DAMNED!
In other news, just got libsass building on Travis-CI! Woot!
Ok, I've begun extracting html2haml to its own gem. I've opened an issue in haml/html2haml#1 to replace this one, so I'll close it here now.