Remove `-webkit-background-clip` from compat spec.

[karlcow]

https://drafts.csswg.org/css-backgrounds-4/#background-clip

The current values as defined in

<bg-clip> = box | border | text

and as noted by @gsnedders on w3c/csswg-drafts#5604

the -webkit-background-clip and background-clip should have the same values per discussion on w3c/csswg-drafts#4839

What would need to be checked?

• the differences in values between compat, CSS background clip, CSS-4 background-clip
• the state of implementations. chrome for example is the only browser not having an unprefixed version of -webkit-background-clip: text
• the state of the Web on usage (optional)
• eventually create a PR

[karlcow]

spec	values	initial
compat	border-box, padding-box, content-box, text	none
css background	border-box, padding-box, content-box	border-box
css 4 background	<box>, border, text	border-box

<box> here means border-box, padding-box, content-box

[karlcow]

testing the values for background-clip in WebKit, Gecko and Blink

[karlcow]

So basically I think it probably means that can be removed from the compat spec. The only thing is that browsers can't possibly get rid of the -webkit- values, because many sites rely on them.

As of today

• background-clip: 36% usage
• -webkit-background-clip: 8% ❗and climbing.

[karlcow]

@foolip @lilles @miketaylr I'm happy to create a PR for it. The compat spec prose doesn't seem to be necessary here anymore.

[miketaylr]

@karlcow just so I understand, all the engines support unprefixed background-clip, so you propose to remove the prefixed one from the compat standard? If more than one UA still supports a prefixed alias, I would prefer to keep it in the spec. Once everyone is committed to unship the webkit-prefixed one, it seems good to remove.

Or do I misunderstand?

[lilles]

The prefixed version should not be removed from the compat-spec, but the "text" value does not need to have special handling anymore since that value is now part of the background-clip spec as well. Can probably make the prefixed property an alias.

[karlcow]

what @lilles said, there is no magic anymore or special case.

1. everyone support the same thing (but chrome which doesn't have yet the unprefixed version for the text value).
2. the only requirement for -webkit-background-clip is that if it's remove from browsers it will break sites, so there's a need for it to be kept as an alias, but I'm not sure this needs to be in the compat spec, more than the CSS spec?

Or is there a specific section needed about all the prefixed properties, we can't remove even if well implemented because the Web relies on them? :)

[karlcow]

@miketaylr would that work? #196

[miketaylr]

Or is there a specific section needed about all the prefixed properties, we can't remove even if well implemented because the Web relies on them? :)

Yep! That's kind of the whole point for the spec. If a different spec wants to spec the prefixed variant, we can remove them then: an example, https://www.w3.org/TR/geometry-1/#webkitcssmatrix, which "started" here). But we want them to be defined somewhere.