Beta 3 (Fisher) During installation anaconda correctly detected my Voodoo 5 card (although it detected the memory on board as being 16MB, which is incorrect). The configuration file generated by this works. Post-install, if I run Xconfigurator it always encounters problems during the test phase. I've tried a number of different things and it always says there was a problem. There are two interesting things here: (1) The config file that Xconfigurator generates after post-install does work. startx brings up X just fine. (2) The config file anaconda generates and the config file Xconfigurator generates differ in a number of ways. I will try to attach the files.
Created attachment 8840 [details] anaconda-generated config file
Created attachment 8841 [details] Xconfigurator-generated config file
This defect is considered MUST-FIX for Florence Release-Candidate #1
what would REALLY help is if you can attach the /var/log/XFree86.9.log file generated right after a failed run with Xconfigurator.
The only difference between these two files is that you selected 16bpp with Xconfigurator, and 24bpp with anaconda. It appears the Voodoo5 currently doesn't work in 16bpp, but we don't have hardware to verify this on.
I only have the V5 card for a few more days so I tried to do some more testing (FWIW this was with a reinstall since the problems occurred). Some things I noticed: - Xconfigurator worked fine... most of the time. I tried a few different res/bpp combos and had no problems. Once or twice it would lock up the system, however. More on why this might be later. - Xconfigurator worked twice. The third time it segfaults and I can't run it again until I reboot. - gpm didn't work until I did "service gpm restart" once/twice. It appeared that the lockups occurred if I got gpm started beforehand. If I didn't then Xconfigurator worked fine (until the 3rd time). Some more info about my setup: - MS Intellimouse Explorer connected to USB port - MS Force Feedback 2 connected to USB port Checking /var/log/messages after some successful attempts showed problems with NMI and gpm (see 27393). On a hunch I removed the joystick. After doing this: - the mouse worked immediately upon reboot (no service gpm restart) - no problems with Xconfigurator (at least so far) - no Xconfigurator segfaults after running multiple times I realize all of this is something of a jumble. I don't know exactly what the problem is except that I appear to have a number of problems when the MS joystick is plugged in. I don't expect to be able to use it under Linux (it is a dual boot system), but it would be nice to be able to handle the stick being plugged in.
unfortunately I don't have the V5 card any longer so I can no longer test this issue or confirm a resolution
OK, I'm going to close this as it appears Xconfigurator works OK; it's just running into the joystick problems mentioned in the other bug(s).