On 1101821965 seconds since the Beginning of the UNIX epoch
"Michael Kukat" wrote:
>

>Hm. As the "palette garbling" looks slightly different in XFree86 and
>XdecNetBSD, it could be that something broke there. If i touch this, i'll have
>a look into. The goal is to fix this problem in both variants, XFree86 and
>XdecNetBSD.

Hmmm, you might try XalphaNetBSD with the -noaccel option.  ISTR
that there were issues with the PWS433au and some of the accelerations
on some of these cards.

>candidate for the main box, if it works. Maybe i can move the acceleration
>stuff from XFree86 to XdecNetBSD, the performance difference currently is
>remarkably large. And as XdecNetBSD is part of the official releases (at least
>it was in 2.0RC4), working on this can't be bad. I think of Multias and other
>machines which just come with those old TGAs and setups which simply don't
>need the bloat of a full-featured XFree86.

XdecNetBSD should already have rectangle copy acceleration which
I found to be the most important for my uses.  You probably don't
have to get all of the accelerations moved across to obtain reasonable
performance.

>                                           Did you ever test multi-head with
>XdecNetBSD? The server knows about options line Xinerama IIRC. I had some
>problems with keymaps and so (yes, i use german layout at home :), have to
>dive a bit deeper into all this. But you definitely made me interested in
>having XdecNetBSD work more comfortable and don't consider it obsolete :)

I did get XalphaNetBSD working with -zaphod, which is the predecessor
to Xinerama.  I think that there's an e-mail in the archives where
I describe how I did it (on this mailing list, many years ago.)
There were a couple of odd steps like mknod'ing /dev/ttyF0, etc.

--
    Roland Dowdeswell                      http://www.Imrryr.ORG/~elric/