Woody xfree86 configurators broken?

From: G*rd*n (gcf_at_panix.com)
Date: 07/29/03 

Date: 29 Jul 2003 13:03:47 -0400

This is in part the text of a bug I sent in to Debian (203283).
I'm posting it here because the only helpful advice I found
in either straight or news Google was in a Spanish-language
Linux NG which others looking for solution to similar problems
might not see. Also, some of you may be able to point the
way forward.

I am one of the many people (judging by what I read in Google
and Google News) who upgraded to Woody and found that their
X stuff didn't work any more. I went through many hours and
a number of iterations of fooling with the configuration files,
to no avail whatever. However, after a considerable amount
of searching, in which I came across many complaints about
very similar problems, but few suggestions and fewer solutions, I
found this -- in Spanish. I'll translate here. The original is
http://groups.google.com/groups?selm=4g6o.3xd.11%40gated-at.bofh.it

Q:
>>... The problem is that it produces an error. I will
>>append the last lines of the [log] (not all, because it is
>>very long and I don't know how to extract [ just the
>>relevant information ]....
>>
>>
>> (EE) VESA(0): no matching modes
>> (II) Unloading /usr/X11R6/lib/modules/libvgahw.a
>> (EE) Screen(s) found, but none have a usable
>> configuration.

A:
>This error is relative to the file //etc/X11/XF86Config-4,
>which is the configuration of the X server.
>It is necessary to revise this file.
>This problem was discussed a few days ago on this list.
>XF86Config-4 ... has an error and that is what is causing
>the server to break. Do the following as root:
>#XFree86 -configure
>and then
>#XFree86 -xf86config XF86Config-4
>and notice what happens. Probably it will work.

This did work to the extent of bringing up a partially usable,
although somewhat defective X system. It failed to configure
the mouse properly (this is an old PS/2 with a /dev/psaux
mouse) but I was able to fix that by hand. Later I was able
to improve (?) the configuration file by adding mode lines
which the one produced as above didn't have. There are still
some severe problems, but in this message I am trying to focus
on the fact that the automatic configuration programs started
by the installation procedure did not work correctly for my
equipment (IBM PS/2 type, S3 Trio64V+ board, NEC MultiSync
5FGE monitor) and judging by what I saw on the Net, don't
work for a lot of others.

I have not found a reference to this problem on the Debian
site. It is not easy to search for the bug because "no matching
modes" appears to be highly generic, the result of a number
of things going wrong with configuration. In other venues,
the solution presented, if any, is often "I messed around with
the files and suddenly things happened to start working."
This does not lead to much insight.

(I might add that when the X server fails to come up at boot
time, a box comes up asking whether the user wants to run a
program to reconfigure. This turns out to be a graphical
program which, on my system, was seriously broken and impossible
to get out of short of a reboot. I would suggest fixing it
or getting rid of it.)

One problem may be that, according to
http://www.xfree86.org/4.1.0/Status28.html#28
support for my chipset (S3 Trio64V+) "has not yet been ported".
Meanwhile XFree86 has moved on to 4.3.0. However, I don't
know if I can revert to 3.3.6 without screwing up the
dependencies, or go forward to 4.3.0. What I have is what
apt-get got me in the last few days, after I loaded 3.0 from
CDs, so it's either on the CD or at debian.org.

In any case, the X functionality is still very deficient. I
have tried to make XF86Config-4 conform to the S3 Trio64V+
and NEC Multisync 5FGe specs, but at present I'm getting
screens in which pieces of objects are littered around, and
the top 3/4" of the display is the mottled gray background
of the X chaos which precedes the "real" screens.
It is evident that someone has a poor idea of the screen
geometry. I'm running gdm, but I suspect the problems lie at
a lower level. Any suggestions would be greatly appreciated. 

-- 
                (<><>)         /*/
       }"{   G*rd*n   }"{   gcf@panix.com   }"{ 
{ http://www.etaoin.com | latest new material 1/19/03 <-adv't

Relevant Pages

  • Woody xfree86 configurators broken?
    ... This is in part the text of a bug I sent in to Debian. ... a number of iterations of fooling with the configuration files, ...
    (comp.os.linux.x)
  • Re: Getting 10.3 to Boot Again
    ... post link to bugreport here so that people with similar problems can go ... The packager has configurations that he has to test and when bug doesn't ... Packager can't test every configuration, ... looks for reports from users, if they don't come they consider that ...
    (alt.os.linux.suse)
  • Re: Getting 10.3 to Boot Again
    ... post link to bugreport here so that people with similar problems can go ... The packager has configurations that he has to test and when bug doesn't ... Packager can't test every configuration, ... looks for reports from users, if they don't come they consider that ...
    (alt.os.linux.suse)
  • Re: why debian
    ... I don't think what scares people away from Debian is the visual polish ... It's the deeper configuration stuff that is ... my X configuration file and add "1024x768". ... another scsi on my computer. ...
    (Debian-User)
  • Re: XP is great
    ... Adobe; it's a distraction, empty work, noise, chaff ... > configuration" of the machine works well enough for him; ... bug being fixed, you have to be aligned with the crowd. ... > witnessed by the many service packs and security patches Microsoft ...
    (sci.electronics.design)