Fix overrides.extends order when including same rules

[kuoruan]

What steps are needed to reproduce the bug?

#6380 changes the behaviour of overrides.extends @jasikpark

In following config file, ./a is the last item, it should in the last of the merge array.

But in the pull request

stylelint/lib/augmentConfig.js

Lines 303 to 305 in 4daef24

	if (b.extends) {
	extendsMerger.extends = [...new Set(extendsMerger.extends.concat(b.extends))];
	}

./a in first extends is used.

What Stylelint configuration is needed to reproduce the bug?

modules.export = {
  extends: ["./a", "./b"],
  overrides: [
    {
      files: ["*.module.css"],
      extends: ["./c", "./b", "./a"]
    }
  ]
}

How did you run Stylelint?

none

Which version of Stylelint are you using?

15.1.0

What did you expect to happen?

overrides.extends work as expect

What actually happened?

see previous

Does the bug relate to non-standard syntax?

No response

Proposal to fix the bug

No response

[ybiquitous]

@kuoruan Thanks for the report with the template. As you said, it seems overrides.extends should precede over `extends.

But, at the same time, this seems an edge case. Could you tell us your specific situation when you found a behavior, please? If possible, it would be so appreciated if you could provide a minimal repository for the reproduction.

[kuoruan]

Hi, @ybiquitous

here is a example:

https://github.com/kuoruan/stylelint-extends-example/blob/master/.stylelintrc.js
https://codesandbox.io/s/competent-wu-e8xor5?file=/.stylelintrc.js

[ybiquitous]

@kuoruan Thanks for sharing the demo. I understand your pain.

We expect selector-class-pattern should have a custom message defined in ./rules/base.js, but actually it has not:

$ npx stylelint --print-config .stylelintrc.js | grep -A3 'selector-class-pattern'
    "selector-class-pattern": [
      "^([a-z][a-z0-9]*)((-|--|__)[a-z0-9]+)*$",
      {}
    ],

$ npx stylelint src/main.scss

src/main.scss
 1:1  ✖  Expected class selector ".c__a-b" to be kebab-case  selector-class-pattern

1 problem (1 error, 0 warnings)

I've labeled the issue as ready to implement. Please consider contributing if you have time.

There are steps on how to fix a bug in a rule in the Developer guide.