Release 1.0

[kaj]

Tracking issue for rsass 1.0 release criteria.

• Basic sass constructs.
• Sass interpolation (#{"foo" + $bar}) in values (pr Feature/value interpolation #2).
  • Sass interpolation in selectors.
  • Sass interpolation in at-rules.
  • Sass interpolation in property names.
• A way to define functions in Rust and call them from sass (pr Support user-defined functions in rust #4).
• Correct string handling / quoting / unquoting.
• Module support:
  • Support @use, @import and @foward (issues Support the @use directive #60)
  • Support most of the builtin modules
  • Support the sass:meta module #129
  • Support the sass:selector module. #130
• Support the @extend directive #65
• A reasonable amount of success on the sass-spec.
• Of course, the release candidate should build successfully on CI (Linux, Windows and Mac)

[Boddlnagg]

Will widespread CSS/SASS frameworks like Bourbon or ZURB Foundation work with 1.0? Currently these two don't work.

[kaj]

Will widespread CSS/SASS frameworks like Bourbon or ZURB Foundation work with 1.0? Currently these two don't work.

I don't want to add beeing able to transpile specific frameworks to the release criteria, but it may well be that the missing interpolation is the thing that is lacking for them, and if so, it will work.

One thing that I have not added to the release criteria, but that I want to support as soon as possible, is the selector-likes stop markers that is used for css animations (actually, using keywords from and to already works, but using percentages does not).

[Boddlnagg]

Okay, I looked at the errors that I get when I try to use Bourbon and Foundation with rsass, and the current blocker actually seems to be the missing interpolation (interpolation in property names for Bourbon and interpolation in media queries for Foundation), though this does not say whether everything will work once that is implemented.

Anyway, thanks for your work on this project!

[kaj]

I added a bullet point for string quoting / unqouting. I'm considering requiring support for placeholder selectors and @extend directive.

[Geobert]

Hi, don't forget windows support! Thanks for your initiative!

• build on appveyor

[kaj]

Yes, it "should" build on windows, and I hope there won't be any big problems with that in itself (I think rsass works on windows as-is), but actually having continous verification of that would be a good thing. I'll try to look into it in a week or so. Hints welcome ... :-)

[Geobert]

I can give you a config file for appveyor :)

[Geobert]

Here: https://gist.github.com/Geobert/c3a3ed1c62d4d2f5294a093b36fcd175

[kaj]

Appveyor build added and successfull. Thanks for the config file, @Geobert .

[naturallymitchell]

A reasonable amount of success on the sass-spec.

Would you please break this down further so we can gauge ongoing progress and work remaining?

[kaj]

A reasonable amount of success on the sass-spec.

Would you please break this down further so we can gauge ongoing progress and work remaining?

I extracted Support placeholder selectors and the @extend directive as a separate item. Other than that, I don't really know. Also, I think at least issues #41 , #42 and #46 should be resolved.

[naturallymitchell]

how do the tests look now? (since 0597e3c)

[kaj]

how do the tests look now? (since 0597e3c)

Pretty similar to how they looked before, only some tests that contained multiple properties testing different aspects of something has been broken up into a set of tests testing one aspect each, as for example 0597e3c#diff-b97f6e13978eef129ce0222c336c6d02

[Keats]

@kaj How close would you say it is to usable? I'm looking at alternatives to sass-rs for Zola and there is this library and grass which seem to be at roughly the same stage.

I've tried running the cmd line on the new bootstrap 5 scss files but I get:

Error: $weight: Expected -20 to be within 0 and 100.
    ,
156 |   @return mix(white, $color, $weight);
    |           ^^^^^^^^^^^^^^^^^^^^^^^^^^^
    '
    ../../Downloads/bootstrap-5.0.0/scss/_functions.scss 156:11  import
    ../../Downloads/bootstrap-5.0.0/scss/bootstrap-grid.scss 10:9  root stylesheet
    ,
166 |   @return if($weight > 0, shade-color($color, $weight), tint-color($color, -$weight));
    |                                                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    '
    ../../Downloads/bootstrap-5.0.0/scss/_functions.scss 166:57  import
    ../../Downloads/bootstrap-5.0.0/scss/bootstrap-grid.scss 10:9  root stylesheet
    ,
291 | $link-hover-color:                        shift-color($link-color, $link-shade-percentage) !default;
    |                                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    '
    ../../Downloads/bootstrap-5.0.0/scss/_variables.scss 291:43  import
    ../../Downloads/bootstrap-5.0.0/scss/bootstrap-grid.scss 11:9  root stylesheet

The JS implementation (using dart-sass) from their package.json seem to compile it without errors. Now according to their docs mix should be a value in between 0-100 so the error should be legit.
It would useful to vendor a few common Sass libraries and compile them as additional tests. It also gives an idea of which missing feature is commonly used.

[kaj]

After looking at this for a while I realized what happens here: Since the shift-color function uses the if function rather than the @if directive, both its arguments are evaluated. For normal functions, evaluating the arguments before entering the function body is expected, but as the documentation actually mentions the if function is special in that regard. So it needs a special-case implementation, which is a bit ugly but probably not that hard to implement.

[kaj]

As for the more general "use bootstrap and other libraries to find out what to implement next", I have only done so to some extent, but I don't think it's relevant to include that as an actual test case, at least not until it suceeds once manually. :-)