[Wylug-help] X modes [Was: X fonts]

[John Hodrien]

Dave Fisher wrote:
> On Mon, Jan 16, 2006 at 09:25:45AM +0000, John Hodrien wrote:
>> What have you done to your machine to break it?  
> 
> 
> Premature upgrade, then backed out.
> 
>> Is the font server running?  
> 
> Apparently not, but getting rid of all the font server packages I had (xfs,
> then xfstt) seems to have eliminated the font-related error message. I'm not
> sure whether this is a good thing or not.

I tend to go with running a font server mostly out of habit, but i'd get xfs 
working and happy before I got lost with what I'd broken.

> The new problem seems more severe. From Xorg.log:
> 
>   Fatal server error:
>   no screens found
> 
>>From earlier in the log:
> 
>   (II) I810(0): Using detected DDC timings
>   (II) I810(0):   HorizSync 31-80
>   (II) I810(0):   VertRefresh 56-75
>   (WW) I810(0): config file hsync range 30-65kHz not within DDC hsync range 31-80kHz
>   (WW) I810(0): config file vrefresh range 50-75Hz not within DDC vrefresh range 56-75Hz
> 
> This is mildly perplexing (to say the least), because I specifically set the
> sync and refresh rates to stay within bounds. From my xorg.conf:
> 
>   Section "Monitor"
>         Identifier      "DELL E173FP"
>         Option          "DPMS"
>         HorizSync       31-80
>         VertRefresh     56-75
>   EndSection

I'm tempted to say you're quoting an unrelated bit of the config file (assuming 
you're quoting the correct config file) seeing as those rates don't match what 
it thinks the config file says.  Does the screen section really reference the 
monitor "DELL E173FP"?

> It's as if some part of the Debian system deliberately chose to throw away my
> xorg.conf settings and substitute them with earlier ones from a dodgy
> configuration autoprobe.

Not being a debian user...

> John's suggestion about looking at /tmp permissions was helpful. I didn't find
> any specific problems there, but the idea did bring my attention to an
> unkillable gdm process that might be preserving earlier settings.
> 
> Sending stop signals via /etc/init.d/gdm  produces:
>   
>    * Stopping GNOME Display Manager...  
>    * GNOME Display Manager not running
> 
> Which is a lie, cos ps reveals the same gdm process running bold as brass
> before and after the signal.
> 
> /usr/bin/gdm-signal stop produces:
> 
>   ** (gdm-signal:13334): WARNING **: Failed to establish a connection with GDM: No such file or directory
>   ** (gdm-signal:13334): WARNING **: Failed to establish a connection with GDM: No such file or directory
> 
> And, again the same gdm process remains before and after.
> 
> kill (with -15 or -9) kills the gdm process, but it automatically repawns moments later.

Which will be because you're running in init 5, and init is respawning it for 
you as you've told it to.  If you dropped to init 3 all would go back to normal. 
    I assume the gdm initscript is there to startup gdm before you've finished 
booting, although I could be wrong, I'm guessing without knowing how debian does 
things.

> On the other hand, a reboot seems to clear the old settings. 
> 
> The system runs X when it comes up again, even if the mode is a terrible 8 bit
> 640x480 and the gnome enviroment is thoroughly broken.

So post the full config file and full X log and I'll see if I can suggest 
anything.  Also poke /var/log/messages to see if there are any other interesting 
failure messages.

> I suspect that I can fix the gnome stuff eventually, but I'd like to know what
> went wrong in the first place.  Not least because, I've had similar problems at
> virtually every upgrade on the same machine.

I leave it to a debian user to work out what you've done wrong with your upgrade 
[I know naaaathing].

jh