On Sat, Aug 7, 2010 at 1:42 AM, Matthieu Herrb <mhe...@gmail.com> wrote: > On Sat, Aug 7, 2010 at 10:29 AM, patrick keshishian <pkesh...@gmail.com> wrote: >> hmm.. I'm leaning towards this being a problem with X now. I'm seeing >> very rare, transient artifacts with my WM's menu (pic added to URL >> below). I also have noticed horizontal "distortion" line through my >> desktop background images (forcing a redraw of same image clears the >> "distortion"). >> >> >> On Fri, Aug 6, 2010 at 2:24 PM, patrick keshishian <pkesh...@gmail.com> >> wrote: >>> Hi, >>> >>> With macppc snapshot install from Aug 3rd I was able to, once again, >>> run firefox (v 3.6.8). But there are some drawing glitches in the UI. >>> The Gimp port finally finished building and it too is demonstrating >>> similar artifacts. They resemble the color shifts one would see when >>> using apps that install their own colormap, but not quite. >>> >>> I'm not certain if this is an X issue or if the problem is at a higher >>> level in the app or the toolkit (gtk+2). I don't run many X apps, the >>> few others (wmapm, xv image viewer, xpdf) have much simpler UI >>> interfaces and don't show issues. >>> >>> I placed three screen shots demonstrating the "glitch" over here: >>> http://sidster.com/scratch/glitch/ >>> >>> With firefox and gmail interface as an example, you see the outline of >>> check-boxes aren't drawn. If I roll-over them with the pointer they >>> redraw but once the pointer is moved away they disappear. But if I >>> check one of them (as shown), the outline for that check-box and all >>> check-boxes below it remain drawn. >>> >>> Also included are dmesg[1], X log and conf. >>> >>> Any ideas as to the cause? >>> >>> I am assuming no one is having issues with -current and associated >>> ports on other platforms. I'm just not sure how many others use macppc >>> -current on a daily basis. >>> >>> Thanks, >>> --patrick >>> >>> [1] Possibly not relevant, but it's my first time seeing this: >>> attempting to restore vector in use vecproc 0 veccpu 6b0ff0 >> >> > > Try without 'Option "AccelMethod" "EXA" in xorg.conf.
You win a cookie. I thought I have always had that on, but a quick search through archives shows that at least in April time-frame snapshot I had it commented out. Thanks for the pointer. --patrick