15:35
<krosylight (they/them)>

annevk: Regarding https://wpt.fyi/results/html/semantics/forms/the-input-element/color.window.html?label=master&label=experimental&aligned&q=color.window.html 's "Testing '#ffffff08' with color space 'limited-srgb' and with alpha (setAttribute("value"))", the test does:

  1. setAttribute("value", color)
  2. Expects value is set as alpha-less hex
  3. Set .alpha = true
  4. Expect the alpha value revives

Where does alpha revive from? "update a color well color" overwrites element's value and doesn't seem to do getAttribute("value")-equivalent thing. Am I missing something again?

16:05
<annevk>
krosylight (they/them): it's been quite a while since I worked on this, but it's possible that step 2 of https://html.spec.whatwg.org/multipage/input.html#update-a-color-well-control-color should look at the value attribute and not the value concept. Because you do want to be able to set these attributes in a somewhat arbitrary order.
16:07
<annevk>
Although not sure how to match that up with what WebKit is doing.
16:07
<krosylight (they/them)>
but that way it's more complex about which value is used...
16:09
<krosylight (they/them)>
Anyway I'll file a bug, we can talk there
16:09
<krosylight (they/them)>
(happy that I didn't miss details this time 🙂)
16:12
<annevk>
Sounds good. Please explain there how it's way more complex as I'm not immediately seeing it. I do think what that test expects is desirable for web developers.
16:27
<krosylight (they/them)>
(me now thinks it's just a matter of checking dirty flag)
16:27
<krosylight (they/them)>
https://github.com/whatwg/html/issues/12057