> After installing NetBSD/i386 1.6 with XFree86 4.2 as the default X system,
> I've come up a little stuck configuring X.

At least I figured out X for myself.  (^&

In case anyone in the future searches these archives for solutions to
similar problems:


> I used "XFree86 -configure" as root to generate a configuration file.
> This appears now to be more or less correct in many ways.  It is, however,
> totally useless w.r.t. resolutions.  It only comes up in one, very low,
> resolution.  I recall that XFree86 4.0, when *first* released, did a much
 [...]

Not sure about my memory, here, but the fix that worked for me was to
simply use the old, text-based, semi-manual configuration tool
(xf86config).  In that, I told it my monitor frequencies, then found out
how it wrote the config file.  It turns out that I didn't need to muck
with modelines: I just needed to supply the frequency values (horizontal
and vertical).


> better job of determining acceptable output resolutions.  Is there an easy
> way to get XFree86 to fill in whatever 4.x uses as the equivalent of 3.x's

Depending on one's definition of "easy", either "Yes", or "Still looking".
I think it *should* have been able to update my monitor section correctly.
(It did correctly identify the monitor type, but just failed to do
anything useful with that information.)


> Additionally, the last time that I tried to run "startx", I met with
> failure.  (I had to run XFree86 as "XFree86", which didn't use my .xinitrc

Fixed by re-setting the symbolic link of /usr/X11R6/bin/X to point to
/usr/X11R6/bin/XFree86 instead of to the old (3.3.x) XF86_S3V or whatever.


It isn't perfect; XFree86 seems okay in 8bpp, but in 16- and 24bpp, it
generates noise.  I don't know if this is due to overeager server
optimization that needs to be toned down, or if the card is being pushed
harder than it is supposed to be pushed, due to a config error.  (I didn't
think that it would support 1600x1200 at 16- and 24bpp, but it seems to be
trying, with mixed results.)


  ``I probably don't know what I'm talking about.'' --rauch@math.rice.edu