In article <E0xiez6-00004D-00@taurus.cus.cam.ac.uk> prlw1@cam.ac.uk (Patrick Welche) writes:

>It gets as far as reading in /usr/X11R6/lib/X11/rgb.txt, and then
>
>        144 238 144             light green
>        144 238 144             LightGreen
>       "
>  2595 XSuSE_NVidia_Fre RET   read 991/0x3df
>  2595 XSuSE_NVidia_Fre CALL  break(0x23b000)
>  2595 XSuSE_NVidia_Fre RET   break 0
>  2595 XSuSE_NVidia_Fre CALL  read(0x4,0x21d000,0x2000)
>  2595 XSuSE_NVidia_Fre GIO   fd 4 read 0 bytes
>       "" 
>  2595 XSuSE_NVidia_Fre RET   read 0
>  2595 XSuSE_NVidia_Fre CALL  close(0x4)
>  2595 XSuSE_NVidia_Fre RET   close 0
>  2595 XSuSE_NVidia_Fre PSIG  SIGBUS SIG_DFL
>  2595 XSuSE_NVidia_Fre NAMI  "XSuSE_NVidia_Fre.core"

Not so good plan then :-) Try digging in the source to find what happens 
after reading rgb.txt or ktrace the server in a FreeBSD machine and compare
the results. Or finally, find out what instruction caused the core...

christos