Re: [fxtf-drafts] [filter-effects-2] backdrop-filter and visibility (#455)

The CSS Working Group just discussed `[filter-effects-2] backdrop-filter and visibility`.

<details><summary>The full IRC log of that discussion</summary>
&lt;emeyer> Topic: [filter-effects-2] backdrop-filter and visibility<br>
&lt;emeyer> Github: https://github.com/w3c/fxtf-drafts/issues/455<br>
&lt;emeyer> emilio: Browsers are weird and inconsistent about backdrop-filter, which we hit when starting our own implementation.<br>
&lt;emeyer> …We haven’t found content with visibility: visible in the backdrop in the wild, but we should be clear about what should happen.<br>
&lt;emeyer> smfr: My mental model is that you generated a bimap the size of the element with the filter and render the content.  So you end with a texture with transparent around the edge and then use that to composite with the backdrop.<br>
&lt;emeyer> …So to my thinking, the Safari behavior looks correct, minus the toggle bug.<br>
&lt;emeyer> emilio: Doesn’t that mean you should render the whole thing even when the backdrop filter is hidden?<br>
&lt;emeyer> smfr: That is a bit weird, I guess.<br>
&lt;emeyer> smfr: I think we’d be willing to change to what you suggest.<br>
&lt;emeyer> emilio: I think determining the visibility by the element the filter is on is simpler, rather than rendering the whole thing.<br>
&lt;emeyer> …Are there other filters that depend on the background behind the element?  For most stuff you end up filtering transparent.<br>
&lt;emeyer> smfr: mix-blend-mode is the only one I can think of.<br>
&lt;emeyer> …I think the way forward is to test with the filter property and maybe propose that visibility hidden will always hide the whole backdrop.<br>
&lt;emeyer> dbaron: For what it’s worth, I think of visibility: hidden as a paint-time effect.  It should be mostly independent of this compositing things.<br>
&lt;emeyer> …It’s sort of inherited and controls whether things are painted or not.<br>
&lt;emeyer> emilio: That’s what Gecko did, but breaks sites that are set to have backdrops hidden.<br>
&lt;emeyer> …I’ll add some links to examples.<br>
&lt;emeyer> smfr: To dbaron’s point, it’s like saying it’s basically opacity: 0 but affects hit testing.<br>
&lt;emilio> https://bugzil.la/1765862 is the Gecko bug ftr<br>
&lt;emeyer> astearns: Sounds like we need more information and will need to do testing around filter as well as backdrop-filter.  Take back to the issue?<br>
&lt;dbaron> I was saying I think it's like a middle ground between color:transparent and opacity:0<br>
&lt;dbaron> (I might have been muted when I said it)<br>
</details>


-- 
GitHub Notification of comment by css-meeting-bot
Please view or discuss this issue at https://github.com/w3c/fxtf-drafts/issues/455#issuecomment-1124015513 using your GitHub account


-- 
Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config

Received on Wednesday, 11 May 2022 16:54:00 UTC