https://bugs.freedesktop.org/show_bug.cgi?id=52454
--- Comment #5 from Kristian Høgsberg <k...@bitplanet.net> --- (In reply to comment #4) > Created attachment 69268 [details] [review] > patch > > (In reply to comment #3) > > I'm still able to reproduce this issue using cairo-image on today's master > > bits (I have not tried with cairo-egl). > > I was able to reproduce this with --with-cairo-glesv2 and decided to look > into it. The problem is that window->cairo_surface is NULL when > get_text_extents() is called because this surface is destroyed on attach, > for the shm path. I've attached a patch to fix the problem. I'm not sure if > it's correct since I'm not sure why cairo_surface is destroyed on the shm > path only. Yeah, I think this is a good fix. I thought we relies on the surface being destroyed after flush, but that's obviously not how the EGL surface works. Your patch brings the two backends closer in how they work and fix the problem... let's go with that. Mind sending it to the list? -- You are receiving this mail because: You are the assignee for the bug.
_______________________________________________ Wayland-bugs mailing list Wayland-bugs@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/wayland-bugs
