[Webkit-unassigned] [Bug 249583] [GTK][WPE][Compositing] Dynamically drawn HTML elements are not correctly composed when css-transformed with the EGL renderer

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Mon Dec 19 10:32:50 PST 2022


https://bugs.webkit.org/show_bug.cgi?id=249583

--- Comment #2 from Mario Sanchez Prada <mario at webkit.org> ---
(In reply to Loïc Le Page from comment #0)
> [...]
> This issue happens with the following components and systems:
> - WebKitGTK on Wayland desktop with NVidia or Intel cards
> - WPE MiniBrowser on desktop with NVidia card (Intel not tested by it would
> be likely the same as for WebKitGTK)
> - wpe-android with the Android emulator and on real phone devices
> 
> The issue is systematic, it happens 100% of the time.
> 
> At the contrary everything works well and as expected on WebKitGTK on X11
> desktop with NVidia card (Intel not tested).

Thanks Loïc for the report! For what it's worth I run a quick test on my laptop with the following configuration and I **also reproduced the issue**:
  - Intel® Iris® Xe Graphics GPU (integrated in Intel Core i7120P)
  - GNOME Shell session on X11 (on Fedora 37 workstation)
  - Epiphany browser 43.0 (based on 2.38.2), using GLX

In such configuration, and even though Epiphany 43.0 uses GLX on a X11 session (something I double checked with strace -f anyway), I can reproduce the problem both with the videos and the webgl demos 100% of the time, so this might not be 

> [...]
> Thus, everything seems pointing towards an issue with the EGL renderer as it
> seems to be the common thing to all failling platforms.

As my quick test shows, this might be not related to EGL after all, since I can reproduce the problem on WebKitGTK 2.38.2 on a X11 session when using GLX instead of EGL, which is something that worked well for you on your NVIDIA card.

I've recorded a short video on https://bugs.webkit.org/attachment.cgi?id=464104 to showcase the issue, hopefully this will help getting this annoying bug fixed alongside the great test cases bug report provided by Loïc already.

PS: I've tested this on Firefox 107.0.1 and Chrome 108.0.5359.98 and works well there

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20221219/9a90bc11/attachment-0001.htm>


More information about the webkit-unassigned mailing list