Hi, is there any intent to add alpha channel support to this property? Firefox allows transparency in RGBA values but Chromium browsers currently do not. Compare this demo in Firefox vs Chrome or Edge to see what I mean:
https://jsfiddle.net/TylerH/jvbznms5/ On Thursday, July 22, 2021 at 12:54:50 PM UTC-4 mk...@chromium.org wrote: > LGTM3. > > -mike > > > On Tue, Jul 20, 2021 at 11:52 PM Daniel Bratell <brat...@gmail.com> wrote: > >> LGTM2 >> >> /Daniel >> On 2021-07-19 23:32, Manuel Rego wrote: >> >> LGTM1 >> >> I see this was re-discussed past week on the CSSWG and the resolution was >> still to keep just 1 color for this property. >> It's nice that you're willing to tweak the color flipping threshold to >> improve interop if needed. >> >> On Thursday, 8 July 2021 at 01:09:04 UTC+2 Joey Arhar wrote: >> >>> I have closed the CSSWG issue >>> <https://github.com/w3c/csswg-drafts/issues/6159#issuecomment-872597989> >>> I referenced before in this thread - I don't think that we are likely to go >>> back and change accent-color to allow multiple colors, and I'd be happy to >>> tweak the color-scheme flipping threshold after shipping this feature in >>> order to improve interop as Emilio mentioned earlier in this thread. >>> >>> On Thu, Jul 1, 2021 at 1:02 PM Joey Arhar <jar...@chromium.org> wrote: >>> >>>> Support for accent-color was added in the DevTools frontend here: >>>> https://chromium-review.googlesource.com/c/devtools/devtools-frontend/+/2759547 >>>> >>>> Here is a picture of the color picker, it appears to work great! >>>> [image: Screen Shot 2021-07-01 at 12.59.20 PM.png] >>>> >>>> On Thu, Jul 1, 2021 at 3:05 AM Yang Guo <yan...@chromium.org> wrote: >>>> >>>>> Is there any special consideration regarding debuggability necessary? >>>>> We should at the least make sure that this new CSS property shows up >>>>> correctly in DevTools and the color picker works in combination with it. >>>>> Please also refer to https://goo.gle/devtools-checklist >>>>> >>>>> On Monday, June 28, 2021 at 10:11:56 PM UTC+2 Joey Arhar wrote: >>>>> >>>>>> Sorry for the delay, I was out last week. >>>>>> Here is a summary of the interop situation, among other things: >>>>>> >>>>>> WebKit hasn't replied to my request for position >>>>>> <https://lists.webkit.org/pipermail/webkit-dev/2021-June/031885.html>. >>>>>> I suspect that if they were to implement accent-color, it would only be >>>>>> for >>>>>> type=radio and type=checkbox since they don't already have any colors >>>>>> for >>>>>> type=range or progress. >>>>>> >>>>>> Mozilla hasn't replied to my request for position >>>>>> <https://github.com/mozilla/standards-positions/issues/540>, but >>>>>> Emilio has already implemented the feature in firefox nightly. >>>>>> >>>>>> The discussions in this thread, as well as the github thread >>>>>> <https://github.com/w3c/csswg-drafts/issues/6159>, are revolving >>>>>> around what exactly to do in order to guarantee contrast. You can see >>>>>> the >>>>>> exact difference between firefox and chrome in these screenshots >>>>>> <https://github.com/w3c/csswg-drafts/issues/6159#issuecomment-858935023> >>>>>> - >>>>>> we don't currently exactly agree on where to flip the color-scheme to >>>>>> ensure contrast, as you can see in the screenshots. Earlier in this >>>>>> thread, >>>>>> Emilio suggested that they should be the same in order for them to be >>>>>> interoperable: >>>>>> > As implemented right now in Chromium and Gecko, some websites would >>>>>> have >>>>>> > pretty different form controls depending on which accent color is >>>>>> > chosen, because we have different thresholds in order to consider >>>>>> > something having little contrast, and Chrome auto-switches to dark >>>>>> mode >>>>>> > controls if that happens (while Firefox just chooses a >>>>>> > better-contrasting foreground color). >>>>>> I wouldn't mind tweaking the chrome behavior slightly in order for >>>>>> them to be the same after shipping, but Emilio also hasn't replied to my >>>>>> comparison images yet. >>>>>> >>>>>> There is also the consideration of having multiple colors provided >>>>>> to accent-color >>>>>> <https://github.com/w3c/csswg-drafts/issues/6159#issuecomment-820881363> >>>>>> versus only using one >>>>>> <https://github.com/w3c/csswg-drafts/issues/6159#issuecomment-864286382>, >>>>>> >>>>>> which Emilio mentioned earlier in this thread. I find it unlikely that >>>>>> we >>>>>> will change to multiple colors since this was already deliberated in >>>>>> CSSWG >>>>>> and we ended up with one color, as well as strong arguments in the >>>>>> github >>>>>> thread by tabatkins and fantasai. >>>>>> >>>>>> On Thu, Jun 24, 2021 at 9:02 AM Yoav Weiss <yoav...@chromium.org> >>>>>> wrote: >>>>>> >>>>>>> Hey folks! Could you provide some clarity on the interop situation >>>>>>> here? There are multiple threads, so a summary would be nice :) >>>>>>> >>>>>>> On Thursday, June 10, 2021 at 7:12:10 PM UTC+2 David Baron wrote: >>>>>>> >>>>>>>> Oops, sorry, it's actually the SVG numbers that slightly disagree >>>>>>>> with the other two, which I suspect are the correct ones (71.52% >>>>>>>> green, >>>>>>>> 21.26% red, 7.22% blue). Though I'm not sure, but either set is >>>>>>>> probably >>>>>>>> close enough. >>>>>>>> >>>>>>>> -David >>>>>>>> >>>>>>>> On Thu, Jun 10, 2021 at 10:06 AM David Baron <dba...@chromium.org> >>>>>>>> wrote: >>>>>>>> >>>>>>>>> On Thu, Jun 10, 2021 at 9:42 AM Mason Freed <mas...@chromium.org> >>>>>>>>> wrote: >>>>>>>>> >>>>>>>>>> Thanks for the explanation and link to the code. That's super >>>>>>>>>> helpful. It looks like the Gecko approach is actually quite similar >>>>>>>>>> to the >>>>>>>>>> Chromium one. The IsDarkColor() >>>>>>>>>> <https://searchfox.org/mozilla-central/source/widget/nsNativeTheme.h#188> >>>>>>>>>> >>>>>>>>>> function just looks at the equal-weighted average of raw RGB >>>>>>>>>> components and >>>>>>>>>> decides "dark" if it is <0.5. And then it uses pure white or >>>>>>>>>> pure black >>>>>>>>>> <https://searchfox.org/mozilla-central/rev/2c991232499e826e46f9d976eb653817340ba389/widget/nsNativeBasicTheme.cpp#192-197> >>>>>>>>>> >>>>>>>>>> as the contrasting color. The Chromium approach differs in two ways. >>>>>>>>>> As >>>>>>>>>> Joey said, we use UX-picked "dark" and "light" colors, which are >>>>>>>>>> close to >>>>>>>>>> black/white, but not identical. And it actually computes the proper >>>>>>>>>> contrast ratio between the accent color and each of these colors, to >>>>>>>>>> ensure >>>>>>>>>> we get the maximum contrast choice, rather than just using averaged >>>>>>>>>> RGB >>>>>>>>>> components. >>>>>>>>>> >>>>>>>>> >>>>>>>>> It's worth emphasizing that this distinction is pretty significant >>>>>>>>> because the contributions of the red, green, and blue color channels >>>>>>>>> in >>>>>>>>> sRGB to perceived luminosity are *very* different. The contributions >>>>>>>>> are >>>>>>>>> 71.54% from the green channel, 21.25% from the red channel, and 7.21% >>>>>>>>> from >>>>>>>>> the blue channel! (Note that those need to be linear-light values, >>>>>>>>> i.e., >>>>>>>>> un-gamma-corrected.) These numbers show up in a number of Web >>>>>>>>> specifications, such as the definition of luminanceToAlpha in SVG >>>>>>>>> <https://www.w3.org/TR/SVG11/filters.html#feColorMatrixElement> or >>>>>>>>> in the middle row of the matrix in lin_sRGB_to_XYZ() >>>>>>>>> <https://drafts.csswg.org/css-color-4/#color-conversion-code>in >>>>>>>>> css-color-4. The WAI rules for computing color contrast ratios >>>>>>>>> <https://www.w3.org/WAI/WCAG21/Understanding/contrast-minimum#dfn-contrast-ratio> >>>>>>>>> >>>>>>>>> (see relative luminance definition >>>>>>>>> <https://www.w3.org/WAI/WCAG21/Understanding/contrast-minimum#dfn-relative-luminance>) >>>>>>>>> >>>>>>>>> use very slightly different numbers but I suspect these may be part >>>>>>>>> of the issue >>>>>>>>> about use of some values from an outdated version of sRGB >>>>>>>>> <https://github.com/w3c/wcag/issues/360> (although I'm not sure). >>>>>>>>> >>>>>>>>> I suspect it's likely a good idea to fix the Gecko code in this >>>>>>>>> case. >>>>>>>>> >>>>>>>>> -David >>>>>>>>> >>>>>>>> -- >> You received this message because you are subscribed to the Google Groups >> "blink-dev" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to blink-dev+...@chromium.org. >> To view this discussion on the web visit >> https://groups.google.com/a/chromium.org/d/msgid/blink-dev/9a27ab88-da84-4f3c-9184-37540d381a1bn%40chromium.org >> >> <https://groups.google.com/a/chromium.org/d/msgid/blink-dev/9a27ab88-da84-4f3c-9184-37540d381a1bn%40chromium.org?utm_medium=email&utm_source=footer> >> . >> >> -- >> You received this message because you are subscribed to the Google Groups >> "blink-dev" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to blink-dev+...@chromium.org. >> > To view this discussion on the web visit >> https://groups.google.com/a/chromium.org/d/msgid/blink-dev/bcca7c09-9632-9164-ac86-f92a54498de2%40gmail.com >> >> <https://groups.google.com/a/chromium.org/d/msgid/blink-dev/bcca7c09-9632-9164-ac86-f92a54498de2%40gmail.com?utm_medium=email&utm_source=footer> >> . >> > -- You received this message because you are subscribed to the Google Groups "blink-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+unsubscr...@chromium.org. To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/46c3f51d-91ba-4b3e-96ce-7a2a0f321382n%40chromium.org.
