I did a GUI install. I have a rivaTNT based card ( Diamond Viper 550AGP ).
It was recognized as "rivaTNT". After I changed the video memory to
16 MB I clicked on "test". Before the new X server was displayed,
there were some dark vertical lines of different width on a black
background. They also appeared when returning to the anaconda screen.
The anaconda screen was all screwed up ! The standard X11 background
pattern was displayed, but in green/black instead of black/white.
The only other objects on the screen were the pointer and the radio
buttons and the arrows of the NEXT/BACK buttons. No text , no anything.
The buttons were still responding , at least the NEXT and BACK.
At this point I switched to VT1 ; black screen. VT2 also black.
VT7 ( still looking green and bugged ) , back to VT1 , now it's OK.
Then I wen't to VT7 and clicked on BACK until I got to the Timezone
Selection. There by moving the pointer over the world map ( which
was not visible ) caused it to be redrawn , mostly around the green
arrow that is drawn when you move the mouse over the map.
This may be several independent bugs , but I'm filing only one bugzilla
Given how late we are in the development cycle for Winston, I really don't think
we're going to be able to fix a driver issue like this before release. The best
thing we can do is work on it for the next release, or if possible, issue an
errata with a driver fix when it's available.
Still there in RC1
This is certainly an X server issue. Some hardware does not respond to the test
sequence well. We have tested cards based on the same chipset without
experiencing this problem. Its a card by card issue, and can only be addressed
by the X server author.