Bug 307933 – workspaces.xyz: (Regression: 305212@main) Hovering on menu makes it to be jumpy

• Component: CSS
• Status: RESOLVED FIXED

[Ahmad Saleem]

Hi Team,

While looking into random websites, I noticed that with Safari Technology Preview 237, the below website menu is jumpy on hover on items.

*** Steps to Reproduce ***
1. Go to https://www.workspaces.xyz/
2. From header, hover on `Setups` and then hover on items in the menu list


*** Expected Result ***
No jump on animation while hovering on items.

*** Actual Result ***
Items jump while hovering on the items.

**** Regression ****
Does not reproduce on Safari 26.3.

**** Other Browsers ****
Does not reproduce on Chrome Canary 147.

Just raising, so we can fix it.

Thanks!

[Radar WebKit Bug Importer]

<rdar://problem/170412045>

[Sam Weinig]

Created attachment 478378 [details]
Reduced Test Case

[Sam Weinig]

As far as I can tell, this is technically the correct behavior. Animations and Transitions are defined in terms of computed value, and since the definition of the border-width computed value changed due to https://github.com/w3c/csswg-drafts/issues/11494, the new behavior here is correct.

That said, it is clearly unexpected, and not all that useful. I filed https://github.com/w3c/csswg-drafts/issues/13511 with the CSSWG to discuss updating the specs to specify using the used values for animation and transitions.

For now, I think we should switch to use the used values for checking for a transition property change, as that resolves the regression here and seems to be what at least Firefox is doing.

[Sam Weinig]

Pull request: https://github.com/WebKit/WebKit/pull/58863

[Sam Weinig]

Created attachment 478391 [details]
Slightly different Reduced Test Case

[Sam Weinig]

So, it turns out my install of Firefox was a bit old, so my analysis of it is wrong. As Emilio Cobos Álvarez pointed out in the CSSWG issue, Firefox behaves the same as ToT WebKit here, so Chrome is really the odd one out here.

It might make more sense not to make this change right now, or implement this as a quirk for this website, if we feel it's that important. We should probably also reach out to the website.

[Simon Fraser (smfr)]

I think we should quirk this site.

[Karl Dubost]

I made a first attempt at contacting the website owner Ryan Gilbert through linkedin

[Antti Koivisto]

Pull request: https://github.com/WebKit/WebKit/pull/58927

[EWS]

Committed 307770@main (aeac793ec439): <https://commits.webkit.org/307770@main>

Reviewed commits have been landed. Closing PR #58927 and removing active labels.

[Antti Koivisto]

So this was fixed with a site-specific quirk. The spec/Firefox compliant behavior remains.

[EWS]

Committed 305413.318@safari-7624-branch (d93666fe0672): <https://commits.webkit.org/305413.318@safari-7624-branch>

Reviewed commits have been landed. Closing PR #4505 and removing active labels.