Singularity fails to compile with error: "function output-isolation has no parameter named $span". (libSass 3.2.0)

[OnkelTem]

I hope I selected the correct issue queue for this report.
On the latest libSass (3.2.0) SGS (1.6.2) fails to compile with this error:

Error: libraries/singularity/stylesheets/singularitygs/_api.scss
  59:87  function output-isolation has no parameter named $span

The next code is responsible for this (line marked with *):

  @if $Direction == 'both' or $From == 'left' or ($Direction == 'rtl' and $From == 'opposite') {
      $Left: call('output-#{$output-style}', map-merge($Span-Map, ('direction': left))); /***/
    }

    @if $Direction == 'both' or $From == 'right' or ($Direction == 'ltr' and $From == 'opposite') {
      $Right: call('output-#{$output-style}', map-merge($Span-Map, ('direction': right)));
    }

The function output-isolation is defined like this:

@function output-isolation($Span-Map) {
  ...

Ideas what is going wrong and how to workaround it?

[xzyfer]

Hi @OnkelTem thanks for the report.

Unfortunately we're not able to progress on this issue in it's current state. Noone on the Libsass team is familiar with Signularity, nor do we have time to debug their entire code base.

Please supply a copy-pastable snippet of Sass code that produces the error you're encounter. Only then can we begin to understand the actual problem you're encountering.

[OnkelTem]

Ok, working on this

[xzyfer]

Appreciate it @OnkelTem

[Snugug]

I happen to know quite a bit about Singularity, and I'm not entirely sure what's going on. I believe the problem is related to argument handling for dynamic call function.

The file in question is here, but I believe the operative lines are L35-47 and L59 where the actual call takes place.

If I were a guessing man, I think the problem is we're passing a map in to the arguments for a dynamic call function and that map is being converted to arguments instead of being passed as a single argument that is a map.

Here is a working SassMeister using Ruby Sass, it'll fail when using the current beta of Libsass (and interestingly works with Libsass 3.1.0) so this should be a regression.

[xzyfer]

You're absolutely correct @Snugug. I've opened a new issue to better track this #1133

[xzyfer]

This was fixed in #1135 and will be in the next release.