scss:unknown-at-rules creates false positives for @if rules when conditions contains "&&"

[agabrys]

Hello, The plugin reported false positives for @if rules when conditions contains && characters.

Environment:

• CSS / SCSS / Less plugin: 3.1
• SonarQube: 5.6.7

Example file: https://github.com/agabrys/sonarqube-falsepositives/blob/master/src/main/sass/d20171218/unknown-at-rules.scss

Project: https://github.com/agabrys/sonarqube-falsepositives

Build: mvn clean package sonar

Best Regards

[racodond]

Hi @agabrys,

Could you please upgrade to the latest version (4.13) and tell me if you can reproduce?

Thank you

David

[agabrys]

Hi David, Thank you for fast answer.

Yes, I'll check it and let you know tomorrow. I didn't know that the latest version is 4.13 because the latest version available in SonarQube Update Center is 3.1.

[racodond]

OK. Thanks!

As a side note, all my plugins are no longer available in the Update Center: CSS, JSON, Cucumber Gherkin, Java Properties.

[agabrys]

I've tested the same code on:

• CSS / SCSS / Less plugin: 4.13
• SonarQube: 6.7

and I see the same problem:

[racodond]

Thanks! I'll investigate and keep you posted.

[racodond]

@agabrys: Indeed the && operator is not supported by the plugin. Only the and operator is. I can't find anything about an && operator at http://sass-lang.com/documentation/file.SASS_REFERENCE.html. Is it really supported by SCSS?

[agabrys]

I changed the example code to:

$variable1: true;
$variable2: false;

.queries {
    &-ok {
        @if $variable1 == $variable2 {
            color: blue;
        } @else {
            color: red;
        }
    }
    &-fp {
        @if $variable1 && $variable2 {
            color: blue;
        } @else {
            color: red;
        }
    }
    &-ok2 {
        @if $variable1 and $variable2 {
            color: blue;
        } @else {
            color: red;
        }
    }
}

SonarQube created issues only in &-fp block. I checked a generated file and it contains:

.queries-ok {
  color: red; }
.queries-fp {
  color: blue; }
.queries-ok2 {
  color: red; }

SCSS treats true && false as true. I checked also false && true and the result is the same: true. It means that the && operator is not supported but unfortunately it doesn't break the compilation process.

The rule works correctly but I think the message is misleading. Maybe this rule should ignore that code and the "SCSS parser failure" rule should raise an issue? What do you think?

[racodond]

Hi @agabrys,

Thanks for your feedback!

SCSS treats true && false as true. I checked also false && true and the result is the same: true. It means that the && operator is not supported but unfortunately it doesn't break the compilation process.

Indeed, it would be nicer if the SCSS parser failed while encountering this kind of syntax. This is something that you could make the SCSS team aware of.

The rule works correctly but I think the message is misleading.

It might be misleading but a prerequisite of the SonarQube plugin is that the files to parse should be syntactically valid.

Maybe this rule should ignore that code and the "SCSS parser failure" rule should raise an issue? What do you think?

@if $variable1 && $variable2 {
    color: blue;
}

Strictly speaking, this piece of code is valid CSS code. So, I don't really want to flag it as invalid syntax and raise an issue against "SCSS parser failure". Also, I don't really want to tweak the "scss:unknown-at-rules" rule to not raise this specific issue because strictly CSS-speaking this finding is kind of correct.

David

[agabrys]

I understand your point of view and I agree with it 👍 I'll let SCSS team know about this issue.

Thank you, I'm closing the ticket.

[agabrys]

I created an issue in the SASS tracking system: sass/sass#2429

[agabrys]

@racodond Natalie Weizenbaum from SASS team wrote::

This is valid syntax—true && false is parsed as a space-separated list containing true, &, &, and false. It may be worth adding a warning for &&, though, since it certainly looks confusing.

So the rule has an issue, but different than I thought at the beginning.

[racodond]

@agabrys: Thanks for the follow-up! It is really a corner case but I'll try to look at it as soon as I get some free time.