Error with & selector interpolation in @at-root blocks

[stol]

Hello,

I have a different behaviour with & parent selector, between sass and libsass

.component{
    color: red;
    @at-root{
        #{&}--foo#{&}--bar {
            color: blue;
        }
    }
}

.test{
        .selector#{&} {
            color: blue;
        }
}

// libsass 3.4.0-beta.4 outputs:

.component {
  color: red; }
  .component--foo.component--bar {
    color: blue; }

.test .selector.test {
  color: blue; }

// Ruby sass outputs:

.component {
  color: red; }
  .component .component--foo.component--bar {
    color: blue; }

.test .selector.test {
  color: blue; }

It works correctly when the interpolation is not inside @at-root blocks!

Thank you !

[ekskimn]

@xzyfer I don't think I have the rights to assign this issue to myself....would you mind assigning it to me?

[xzyfer]

@ekskimn looks like issues can only be assigned to core members

[ekskimn]

@xzyfer ah, got it =)

[mgreter]

@ekskimn you may add the WIP tag, so we know that someone may already working on it!

[xzyfer]

This still fails the spec.

[xzyfer]

Apologies, it looks like my local sass-spec is having issues.

[xzyfer]

This fix was a false positive due to sass/sass-spec#380. The spec still fails.

[tedw]

I noticed another issue with & interpolation in @at-root blocks:

.foo {
  @at-root button#{&} {
    color: red;
  }
}

// Generated CSS has an additional selector prepended
.foo button.foo {
  color: red;
}

This wasn't a problem in 3.2.2, only 3.2.3 and later.

I thought it made sense to include this here, but I'm happy to submit a separate issue if you prefer.

Thanks!

[ekskimn]

Hi all, i'm on this bug again...should have a fix for it shortly...I think I know what's going on with it.

[mgreter]

@ekskimn FYI I actually already have a WIP branch which passes this (an other at-root) issues ...

[ekskimn]

oh? is it pushed to the repo?

[mgreter]

#1249: Not sure if I already pushed all the latest work. But it's really in a WIP state, so no, it's not yet pushed and needs more refactoring from my side ...

[ekskimn]

Ah i see what you're doing....lol...you really didn't like the contextualize_eval, huh? =)

Do you happen to know when you might push it? My team and I are wanting to get that fix in to fix a particular issue we have.

Also do you happen to know if you're going to merge the selector-functions PR in as well?

[xzyfer]

The plan is to @mgreter PR and selector functions with 3.3. As for when that'll be it's hard to say.

[mgreter]

We now have a passing spec test since #1249 was merged 🎈

[esr360]

@tedw's comment several posts up still seems to be an issue? Is there a separate issue for this?

.foo {
    @at-root button#{&} {
        color: red;
    }
}

outputs:

.foo button.foo {
    color: red;
}

expected:

button.foo {
    color: red;
}

http://sassmeister.com/gist/d610ed626e510cea938c

EDIT: Sorry guys. I can see from this issue that it will be fixed (is fixed) in 3.3 #1491