Today i had the chance to test the g-i on my friend's laptop, who
yesterday reported g-i crash and swapped fb0 fb1 devices (actually he
owns a GeForce 6200 Go)
I repeated the test using the very same ISO as yesterday and, this time,
vesa was correctly mapped fb0 and g-i booted successfully.
We tried many times both cold and warm-booting, after X had started, but
could not reproduce the crash nor the fb0/1 swap.
Note that yesterday's crash happened even if fb1 was sait to be vesa and
"fbdev=/dev/fb1" was set in directfbrc.
Hence, i believe it's safer patching the S60frontend script so that the
GTK frontend, on i386 and AMD64, is started iff fb0 is vesa.
A patch for rootskel which supersedes previous one for rootskel-gtk is
attached, but please verify correctness of line
if [ subarch=="i386/generic" ] || [ subarch=="amd64/generic" ] ; then