Support parsing noscript content in script-enabled mode

[stevecheckoway]

This change adds support for parsing noscript content with the scripting flag enabled. In this case, Gumbo will treat the content of the noscript element as raw text, exactly as the browser would.

What problem is this PR intended to solve?

Closes #3178

Have you included adequate test coverage?

Yes.

Does this change affect the behavior of either the C or the Java implementations?

It adds support for this parsing option to libgumbo only.

[stevecheckoway]

Suggestions on a better option name are especially welcome.

I didn't want to use something like scripting_enabled since that strikes me as misleading.

[stevecheckoway]

I just realized that I haven't had the html5lib tests running so this is really not ready for review yet. Gotta figure out what I did wrong here first.

[stevecheckoway]

It seems that we're now passing the scripting enabled html5lib tree construction tests.

[flavorjones]

This PR should also update the docstring in lib/nokogiri/html5.rb.

Let me know if you'd like me to put in any of this work? I'm happy to do as much or as little as you like. ♥

[flavorjones]

Suggestions on a better option name are especially welcome

The name parse_noscript_as_text is not awful, and does what it says. If I was going to suggest anything, it would be a vaguer description like emulate_browser or pretend_scripting_is_enabled.

Or if we want to be funny, wwtbd ("What Would The Browser Do?")

[stevecheckoway]

This PR should also update the docstring in lib/nokogiri/html5.rb.

I did that as well. But I left the existing option name. I'll defer to your experience in API design.

[flavorjones]

LGTM! I added a commit to fix up some of the rdoc formatting (which is famously challenging to write correctly without looking at the generated docs in a browser) and a CHANGELOG entry.

[flavorjones]

The changelog entry isn't worded perfectly ... I'll fix it in postproduction