Fix is_superselector_of function

[mgreter]

Fixes #823

[coveralls]

Coverage decreased (-0.02%) to 80.28% when pulling 625a2a0 on mgreter:bugfix/issue_823 into fd7f20f on sass:master.

[mgreter]

@xzyfer should we include this in next 3.2.0-beta.5 release?

[xzyfer]

Maybe this is best left for when selector functions lands.
On 3 Apr 2015 09:25, "Marcel Greter" [email protected] wrote:

@xzyfer https://github.com/xzyfer should we include this in next
3.2.0-beta.5 release?

—
Reply to this email directly or view it on GitHub
#1033 (comment).

[mgreter]

I actually thought it would make sense to have it merged before selector functions, since it could possibly have an impact there (they may suffer from the same bug). So I thought it would be good to already have it merged. Otherwise chances are that this bug will be fixed twice !?

[xzyfer]

Cool, go for it. I just thought it might be a risky change but it seems fine
On 3 Apr 2015 21:03, "Marcel Greter" [email protected] wrote:

I actually thought it would make sense to have it merged before selector
functions, since it could possibly have an impact there (they may suffer
from the same bug). So I thought it would be good to already have it
merged. Otherwise chances are that this bug will be fixed twice !?

—
Reply to this email directly or view it on GitHub
#1033 (comment).

[mgreter]

Pinged @onedayitwillmake so he maybe can verify this changes to be sane. I guess he should have a few additional spec tests by now, so maybe he can test if this brakes any existing behavior.

[coveralls]

Coverage decreased (-0.02%) to 80.7% when pulling ce01b5e on mgreter:bugfix/issue_823 into 70992f4 on sass:master.

[mgreter]

OK, just made some tests with the WIP branch and it seems to alter results for the new selector function tests. So I will leave this PR open for now for further feedback and references!

[mgreter]

OK, lets wait for selector-functions and the new spec tests which should cover this topic better.
IMO the selector functions cannot be merged without a proper superselector handling anyway!

[onedayitwillmake]

@mgreter I agree, i can merge that in. The selector function, is essentially parsing the selectors, returning the result of calling that method

[mgreter]

There is a very strange and IMO wrong logic at work with ruby sass and is_superselector. I fail too see why ruby sass produces different results for these two tests:

test-a {
  // false
  $a: "a + c + d";
  $b: "a + b + c + d";
  a: is_superselector($a, $b);
}

test-b {
  // true
  $a: "a + c + d + e";
  $b: "a + b + c + d + e";
  a: is_superselector($a, $b);
}

IMO I found the code part in ruby sass that is mainly responsible for this behavior:

# .foo > .baz is not a superselector of .foo > .bar > .baz or .foo >
# .bar .baz, despite the fact that .baz is a superselector of .bar >
# .baz and .bar .baz. Same goes for + and ~.
return if seq1.length == 3 && seq2.length > 3

With this info I might be able to re-produce the IMO "strange" behavior ...

[mgreter]

This is included in my WIP branch for 3.3!