Add text-decoration-skip default value to UA stylesheet

[drott]

Following the discussions in [1] we propose to add the value "objects
ink" to the body, html elements for the text-decoration-skip property. I
interpret the thread to have reached consensus on the aspect of making
ink skipping a default behavior by means of a UA stylesheet. We believe
it brings better typography to the web.

[1] w3c/csswg-drafts#727 (comment)

Tests: w3c/web-platform-tests#7605

[drott]

CC @kojiishi @domenic @foolip

[drott]

CC @litherum

[annevk]

In the "Intent to Ship" thread @zcorpan asked why this wasn't applied to *|*:root instead. That feedback is yet to be addressed. (It also would seem sufficient to apply this just to html and not body, if we don't decide to go with *|*:root.)

cc @fantasai @tabatkins

[drott]

In the "Intent to Ship" thread @zcorpan asked why this wasn't applied to |:root instead. That feedback is yet to be addressed. (It also would seem sufficient to apply this just to html and not body, if we don't decide to go with |:root.)

I am certainly happy to update the CL with *|*:root, would you like me to update the PR accordingly?

[litherum]

Why should this be in the UA stylesheet and not the initial value of the property?

[tabatkins]

Phew, that original thread sure meandered. But yeah, it seems to have settled on "ink" being part of the initial value; there's no reason to put it in the UA stylesheet. This shouldn't be merged; we should poke the CSSWG thread and have the initial value changed. (Either to objects ink or to auto.)

[domenic]

Why should this be in the UA stylesheet and not the initial value of the property?

This is a great question. I hadn't noticed before, but that's apparently how the spec does some things regarding text:

The initial value for the 'color' property is expected to be black. The initial value for the 'background-color' property is expected to be 'transparent'.

@litherum, @tabatkins, do you know what the differences are between these approaches, and why the spec would choose one versus the other?

@tabatkins, you seem to be saying the CSS spec usually governs initial values. Should HTML remove the above sentence about color/background-color?

[domenic]

Also, the test at https://github.com/w3c/web-platform-tests/pull/7605/files will still be correct for the initial value version, right?

[tabatkins]

'color' and 'background-color' were made UA-defined for accessibility reasons, I believe. (That said, the current spec confusingly just says "the initial value is 'black'" in the prose; I'll raise an issue about that.)

We don't do that for anything else. We only use UA-stylesheet rules when a particular value isn't good globally. That isn't the case here; if we're agreeing that ink-skipping is a good default behavior, it looks like it's good for everyone - that seems to be what the thread has settled on.

[drott]

Also, the test at https://github.com/w3c/web-platform-tests/pull/7605/files will still be correct for the initial value version, right?

@domenic: Yes, I believe so.

This shouldn't be merged; we should poke the CSSWG thread and have the initial value changed. (Either to objects ink or to auto.)

@tabatkins: Thanks, I am attempting to revive the discussion in w3c/csswg-drafts#727 (comment) - perhaps you could comment there, thank you.

[domenic]

It looks like this was superceded by w3c/csswg-drafts@70e40c9 . However, please merge web-platform-tests/wpt#7605 still!