Get index of node within source string

[jeddy3]

Would it be possible to support getting the string-index of a node within its source string? Similar to how postcss-selector-parser does it.

The use-case is for stylelint, to feed these indexes into PostCSS 5's new warning index option, in order to get accurate positions for errors.

[TrySound]

@jeddy3 Do you mean position start?

[jeddy3]

Yep. I think so.

So given a value of:

red, rgba(0, 0, 0, 0);

You'll get something like

nodes: [
      type: 'function',
      value: 'rgba',
      before: "",
      after: "",
      sourceIndex: 5 // the start position of the `rgba` in the original value string

Although, now I'm wondering if I misunderstood our needs. I'm going to ask for @davidtheclark advice on this.

@davidtheclark I'm working on function-blacklist at the moment and I believe I need to get the start position of the offending function name to pass into report().

So given

a {
  background-image: 
    linear-gradient(45deg, blue, red),
    linear-gradient(90deg, black, white);
}

Am I right in thinking I'm after the start position of the two function names within the value string? And this is something we're hoping to get from the value parser?

[davidtheclark]

Yes, @jeddy3: That link to the postcss-selector-parser feature is dead on: that's what we need for a completely accurate position.

[jeddy3]

Thanks for the clarification @davidtheclark .

@TrySound Do you think it'll be possible?

[TrySound]

@jeddy3 Yep, I'll do that.

[jeddy3]

Thank you :)

[TrySound]

@jeddy3 @davidtheclark Do you expect sourceIndex in div since space before or div char?
word , -> sourceIndex: 5 without space
word , -> sourceIndex: 4 with space

[jeddy3]

That's a good question. I've not used the postcss-selector-parser API yet, so I'm going to ping @ben-eb to ask which of these two behaviours would be most consistent with what is in postcss-selector-parser?

[ben-eb]

Since sourceIndex is relative to the original string, whitespace must be taken into account for the position to be accurate. So for word , the comma would be sourceIndex: 5.

[TrySound]

@ben-eb You mean 4? :)

[ben-eb]

sourceIndex should be zero based. 😄

[TrySound]

@ben-eb posDiv - before.length ?

[ben-eb]

Whoops. I meant 4. 😄

[jeddy3]

@ben-eb Thanks for the popping by to answer the question :)

@TrySound Thanks again for getting this done, and so quickly too :)

[davidtheclark]

Wait ... in the word ,, the index of , should be 5 --- I don't understand why you both settled on 4. 'word ,'.indexOf(',') === 5

[davidtheclark]

Nevertheless ... I reviewed a bunch of the tests in the PR https://github.com/TrySound/postcss-value-parser/pull/17/files and they all made sense/seemed right to me, so maybe I'm just missing something here.

[TrySound]

@davidtheclark It takes also the space before which is start of the node

{ type: 'div', sourceIndex: 4, value: ',', before: ' ', after: '' }

[davidtheclark]

👍