Selector Functions

[onedayitwillmake]

Please note: The code in this pull request was written by myself, and is does not represent my employer.

Redoing #1064.

This is a fresh checkout of master, and then manually went in and re-implemented the selector-functions.
This is a much simpler PR with less files modified, and only a single commit - hopefully that helps make everything more clear when merging

Adds the selector functions:

• selector-nest($selectors...)
• selector-append($selectors...)
• selector-extend($selector, $extendee, $extender)
• selector-replace($selector, $original, $replacement)
• selector-parse($selector)
• simple-selectors($selector)
• selector-unify($selector1, $selector2)

Regarding the sass-spec test, there are 2 issues when calling selector-unify and selector-nest and passing in a parent selector reference.

I believe these are related to a couple of other issues cropping up around & - so I think it's safe to merge this in and tackle those as a new issue.

Known Issues from sass-spec libsass issue_963 tests:

selector-unify

selector-unify(&, '.baz, .bang'); // never makes it passed the parser
Error: invalid selector after &
        on line 14 of test_selector_unify.scss
>>   content: selector-unify(&, '.baz, .bang');
   --------------------------^

selector-nest

selector-nest('.foo', '&.bar','baz &"); // '&' before .bar and after .baz are thrown out

[coveralls]

Coverage decreased (-1.43%) to 79.21% when pulling 0514d01 on onedayitwillmake:sel_functions into 25a1d79 on sass:master.

[coveralls]

Coverage decreased (-1.43%) to 79.21% when pulling 0514d01 on onedayitwillmake:sel_functions into 25a1d79 on sass:master.

[mgreter]

contextualize doesn't seem to be used?

[mgreter]

Hey @onedayitwillmake, already took a quick look yesterday and looks very nice and clean. Unfortunately appveyor doesn't yet pass, which is pretty much the only thing that I see as a blocker before merging this PR as is. @xzyfer I think at this point it will be best to merge this PR to master and I will then realign my WIP branch for the 3.3 release. I guess we need a few beta rounds anyway with 3.3.

Should we create a separate release coordination issue as I did for 3.2?

[onedayitwillmake]

@mgreter Thank you! Yes it seems to dislike the std::function I used as a custom comparator for creating a std::set< Complex_Selector*> - I assumed I could use C++11 feature.

Looking again, i see you already have a Complex_Selector_Pointer_Compare functor I can use instead!
I will switch to that and commit :)

[coveralls]

Coverage decreased (-1.41%) to 79.23% when pulling 60cc9fc on onedayitwillmake:sel_functions into 25a1d79 on sass:master.

[mgreter]

I have incorporated most of your changes in my latest WIP branch: #1249 - maybe check if I have missed anything! It seems to pass most of the tests and I tried to replace some functionality with more generic functions, which I added during the refactoring!
IMHO we now mostly need to tackle is_superselector in regard to :not (and other) pseudo selectors!
And once again, thanks for your contributions!

[mgreter]

I guess this can be closed since #1249 has been merged 🎉