2 dimensional sugar

[d4tocchini]

#box1[size] == #box2[size]

should compile to 2 constraints:

...
constraints: [
  ['eq', ['get', '#box1[width]'], ['get', '#box2[width]']]
  ['eq', ['get', '#box1[height]'], ['get', '#box2[height]']]
]

[bergie]

:+1:

[paulyoung]

I started looking into this, went to write a spec first, and saw this comment:

This should probably be handled with a preparser or optimizer, not the main PEG grammar

[paulyoung]

I think we're saying that a preparser would transform:

#box1[size] == #box2[size]

into:

#box1[width] == #box2[width]
#box1[height] == #box2[height]

We'd need a good strategy for not duplicating a lot of code between the two grammars since they'd both need the same rules. I've asked about this here: dmajda/pegjs#268

[d4tocchini]

That comment now seems too certain. Your point about duplicating parser rules hold weight. Probably best this is done within the CCSS grammar anyway so VFL & VGL can use 2-D properties and just send them down the stack.

There are some small details to consider, should the below throw a warning?

#box1[size] == #box2[width];

I say go on with your spec, this would be a super helpful addition!

Would be cool to support things like:

#box1[top-left] == #box2[center];

[paulyoung]

For reference, this exists if we ever wanted to use it: https://github.com/andreineculau/pegjs-override-action

[paulyoung]

Any requests for 2D mappings? So far I've configured:

{
  "bottom-left": ["left", "bottom"],
  "bottom-right": ["right", "bottom"],
  center: ["center-x", "center-y"],
  "intrinsic-size": ["intrinsic-width", "intrinsic-height"],
  position: ["x", "y"],
  size: ["width", "height"],
  "top-left": ["left", "top"],
  "top-right": ["right", "top"]
}

[paulyoung]

I have 2D sugar working for stays now too.

[paulyoung]

What's the status of measure? I see an old spec hinting that 2D should work for this too.

Was it deprecated?

[paulyoung]

There are some small details to consider, should the below throw a warning?

#box1[size] == #box2[width];

I need some input on this. Currently it will only create a constraint for the width but also doesn't issue any warnings.

[paulyoung]

Would something like this be preferred?

#box1[width] == #box2[width] !strong;
#box1[height] == #box2[width];

[d4tocchini]

measure was replaced with intrinsics, good catch.

as for #box1[size] == #box2[width];, it should probably constrain just the width so that 2-d props can be used in things like VFL helpers where the dimension would be inferred.