Anaconda and monitor handling

[Mike A. Harris]

Bill Nottingham wrote:
> Peter Jones (pjones at redhat.com) said: 
> 
>>>For the short term, it would be nice if someone were to
>>>reimplement ddc-probe to use x86emu or some other emulator
>>>so that VBE could be used uniformly on all architectures
>>>at least.  That would give reasonable consistency between
>>>architectures even if it wasn't as complete as what the X
>>>server has built in.
>>
>>A better plan might be to make the X server's probe routines into their
>>own DSO, so when other programs need to do this sort of thing, they can
>>use the code which gets the most exposure ;)
> 
> 
> ... along with fixing X so that it can be started with a blank
> monitor section and just figure it out itself. No reason to
> 'run X probing code, write a config file, have X read that' when
> X can just run its own probing code and use the settings directly.

X already does that.  The problem however, is that if DDC probing:

- is not implemented in a driver
- isn't possible due to the attached display not supporting DDC
- is blocked by a KVM switch or other device
- any other reason DDC might fail

... then you get an X server configured to use DDC with horrible
video should DDC not work for any of the above reasons or any
other reason.  You are then forced to manually hand edit your
own config file, wether you are a 1337 hax0r, a developer, a
hobbiest, dentist, lawyer, Grandma, or whatever.

That is a very poor end user experience, and is IMHO worse than
what we have right now, because what we have right now generally
works for the majority of users, wether or not their particular
setup is DDC-enabled and working.

That isn't a reason to not try to improve things however.

Peter, Jeremy and I briefly discussed a possible solution to this
yesterday in IRC, which turns out was already proposed in an
upstream discussion I was partially following.  Kean Johnson
has proposed that we make the X server do the following:

if (UseDDC is set AND we got DDC ranges)
   use DDC probed ranges
else
   if ranges specified in the config file
     use those ranges
   else
     fall back to some pesimistic defaults
   endif
endif


I like Kean's proposal, and he indicated he was working on a patch.
I've attempted to contact him to see if he has finished the
implementation or not.  I'm going to track this closely and
try to help it get merged upstream.  If he has lost interest, I'll
consider finishing it off myself.  This would be something for
X11R7 however, so probably wont make it into FC4 either way, but
it's good that we have had this discussion, in order to pour a bit
of fire into getting DDC to be a bit more sane.

If anyone wants to follow the discussion, join xorg at freedesktop.org,
and search list archives for:

Subject: Proposed change to xorgcfg


Thanks guys.