On Wed, 2003-06-04 at 00:31, Michel Dänzer wrote: > On Wed, 2003-05-28 at 15:51, John Leach wrote: > > > > The only thing that makes a difference is changing the color depth. > > At 24 bit everything has a yellow tinge (removing Blue) > > At 16 bit everything is a bit more random, but mostly pink/purple. > > (removing Green) > > At 8 bit everything looks fine until you start opening applications > > (with new colours) that change the palette. This just starts shuffling > > colours around randomly as you change focus. > > Are the colors in the focussed window correct though? If so, that's > normal behaviour of switching between several private colormaps. Yes, but only in 8 bit mode of course. I kind of guessed the switching of colormaps would be to blame, but thought it was interesting none-the-less and best to include it. > > http://www.johnleach.co.uk/documents/powerpc/XFree86.0.log.dri > > This is the "dri-trunk" packages, with 2 monitors, producing the yellowy > > tinge. > > For one thing, behaviour is undefined if you mix Option "UseFBDev" and > dualhead. This config is the closest I can get to a working setup. With FBDev turned off on both screens I get a corrupted screen, zoomed x5 or so with spaced out pixels in randomy colours (and cloned on the second monitor). With FBDev turned on for both screens, I get only flat blocks of correctly rendered colours with no detail rendered, no text, or icons, images etc. (for example, the gdm login window is a block of grey, with a smaller block of white inside it (the username/password box)). The second screen in this configuration has the broken colour rendering plus artifacts from the primary screen (plus lots of other randomness). I can only get sane screen rendering (except the colour problem) with UseFBDev enabled for the primary screen, and disabled for the secondary screen. John. -- GPG KEY: B89C D450 5B2C 74D8 58FB A360 9B06 B5C2 26F0 3047 HTTP: http://www.johnleach.co.uk
Attachment:
signature.asc
Description: This is a digitally signed message part