Enhance <input type=color> with alpha and colorspace=display-p3

[annevk]

The new alpha and colorspace attributes give it parity with the 2D canvas API.

As part of this change we do away with the "simple color" concept as that has been replaced by CSS colors now.

Fixes #3400.

---

WebKit can contribute tests as part of implementing this feature.

It would be great to get some help from @svgeesus @tabatkins and @LeaVerou on resolving #8917 so HTML can clearly link a way to serialize CSS color values (for the Hexadecimal state we can continue to maintain that ourselves). I think we have a plan there, but let me know if there's anything still unclear (prolly best in that issue).

• At least two implementers are interested (and none opposed):
  • WebKit
  • Maybe Chromium given Tab's participation in the issue and positive comments on WHATNOT?
• Tests are written and can be reviewed and commented upon at:
  • …
• Implementation bugs are filed:
  • Chromium: https://issues.chromium.org/issues/368059226
  • Gecko: https://bugzilla.mozilla.org/show_bug.cgi?id=1919718
  • WebKit: TBA
• MDN issue is filed: Document <input type=color alpha colorspace=display-p3> mdn/content#35953
• The top of this comment includes a clear commit message to use.

(See WHATWG Working Mode: Changes for more details.)

---

/common-microsyntaxes.html ( diff )
/index.html ( diff )
/indices.html ( diff )
/input.html ( diff )

[annevk]

@domenic observed that for <input type=color alpha> we do not have to set the proposed "HTMLCompatible" serialization parameter to true and could instead use the modern serialization. Which would mean instead of rgba(...) (that 2D canvas requires as discussed in #8917) you would get color(srgb ...) presumably. It would still be limited to 8-bits per color component, but it seems reasonable to me.

[domenic]

Editorially looking pretty good, although some questions in #3400 (comment) seem interesting as an uninformed-about-color person.

[josepharhar]

I think it would take us some time to upgrade the color picker popup in chromium (thanks @lukewarlow for mentioning that DevTools has a better one), but if we can just support serialization etc. in IDLs based on these new attributes in the meantime based on WPTs then that sounds good to me.

[lukewarlow]

I think it would be unfortunate to support the alpha attribute but not actually allow the user to select an alpha value? But not fully supporting picking display P3 but supporting the serialisation seems like it could be more acceptable?

[josepharhar]

Yeah I suppose we could easily and quickly add alpha, then hopefully follow up with display p3 later

[annevk]

I believe I have addressed the questions in #3400 and this is now rebased on top of #8917 so it's using all the latest terms. Marking agenda+ to double check on implementer interest, but if that can be provided out-of-band that would be swell as we've already discussed this a number of times.

[svgeesus]

Looks good

[svgeesus]

Looks good overall

[svgeesus]

Suggest "CSS color" link to https://drafts.csswg.org/css-color-5/#typedef-color

[annevk]

I don't think that matches how we talk about "CSS color" in HTML. What we mean here is the parsed representation, the internal object if you will, not the syntax.

[svgeesus]

Is it correct to infer that limited-srgb is also limited to (0 .. 1) ie 0 .. 255 range, as well as being limited to 8 bits?

While display-p3 uses the extended range, so negative values and values greater than 1 (greater than the component value for white) are allowed?

[annevk]

Yes.