[css-align][css-multicol] Make "column-gap: normal" to be 1em in multi-column per spec

[mrego]

Specs:

• https://drafts.csswg.org/css-align/#valdef-row-gap-normal

  The normal represents a UA-dependent length on multi-column elements, and computes to 0px in all other contexts.

• https://drafts.csswg.org/css-multicol-1/#valdef-column-gap-normal

  Identical to , but with a UA-specified length. A value of 1em is suggested.

There's an issue in css-align about this:

ISSUE 4: Should normal just compute to 1em in multicol?

I've realized there's a WPT test checking exactly that: http://w3c-test.org/css/css-multicol/multicol-gap-001.xht
Right now is marked as optional, but it passes in all the browsers I checked: Chromium, Safari, Firefox and Edge.

Maybe it's possible to update the specs and make this test mandatory. WDYT?

[rachelandrew]

Do we know what non-web browser UAs are doing here? I imagine they may be using something different.

[frivoal]

At least Vivliostyle, Prince, and Pdfreactor use 1em as well. I do not have an Antenna House install to test, but i'd be surprised if it differed.

[frivoal]

@MurakamiShinyu Do you know what the behavior of AH formatter is for this? Is column-gap normal 1em, or some other value?

[MurakamiShinyu]

AH formatter uses 1em as well.

[frivoal]

In that case, I think we should just resolve that it's 1em, and remove the "UA defined" bit, as that doesn't seem to help anyone.

[rachelandrew]

I agree, it seems that there is interop on this, so there seems no sense in leaving room for ambiguity.

[css-meeting-bot]

The Working Group just discussed Make "column-gap: normal" to be 1em in multi-column per spec, and agreed to the following resolutions:

• RESOLVED: remove the UI defined bit and go with 1em for computing normal in multi-col

The full IRC log of that discussion

<dael> Topic: Make "column-gap: normal" to be 1em in multi-column per spec
<dael> AmeliaBR: In the multi-col spec there's a suggested value of 1em
<dael> AmeliaBR: Everyone is doing 1em so we suggest it's mandatory.
<AmeliaBR> s/AmeliaBR/rachelandrew /
<dael> florian: And this interop includes print. Let's go for it.
<dael> github: https://github.com//issues/2145#issuecomment-377445142
<fantasai> s/AmeliaBR/rachelandrew/
<dael> astearns: proposed is remove the UI defined bit and go with 1em for computing normal in multi-col
<dael> RESOLVED: remove the UI defined bit and go with 1em for computing normal in multi-col

[rachelandrew]

failed to reference the issue, but commit is here for multicol bf38268