Re: Patched Xfree86 4.3.0 in Debian

[Helge H. <hel...@ai...>]

Svetoslav Slavtchev wrote:

>>On mandag 17. mai 2004, 09:45, Aivils wrote:
>>    
>>
>>>>s/PrefBusID/IsolateDevice/g
>>>>right?
>>>>        
>>>>
>>>Yep.
>>>      
>>>
>>Cool! And thanks!
>>    
>>
>
>i'd say that the "SingleCard" would be a better choice,
>no need for sed or perl :-)
>  
>
I believe "device" is a better word than "card" in this context,
it is possible for hw designers to put several devices on a
card (or spread the device over several cards).  Today there
is usually one device per card, but who knows
what the future will bring.

> 
>  
>
>>>>Then, I wondered about the HOWTO: Probably, we should now add some
>>>>Debian-specific notes, now that people most likely can just use
>>>>Sarge or Sid and get the X servers they need, but must use slightly
>>>>different options than the rest of the world.
>>>>        
>>>>
>
>if you are talking about my howto,
>i'd prefer the notes to be generic,
>and i'll add them but first i've to find some time
>and i've to finaly integrate your patch in my Xorg binaries
>
>  
>
>>>If end-user will start muptiple local xf86, then "IsolateDevice" is
>>>not enough. Each xf86 open /dev/ttyXX , which one current is allowed
>>>under Linus-tree kernel. Secondary xf86 become active when user press
>>>Alt-F8 and primary xf86 goes spleep until user activate /dev/tty7.
>>>
>>>Multiple current /dev/ttyXX exist only in linux-ruby.
>>>      
>>>
>>Yes, of course. Debian's approach to that is that you can create 
>>kernel-patch-packages, which the user can trivially apply if s/he 
>>chooses to build a kernel themself. For example, that's how Reiser4 is 
>>supported in Debian: 
>>http://packages.debian.org/unstable/devel/kernel-patch-2.6-reiser4
>>
>>I know Andreas has been thinking about making such a kernel-patch 
>>package for ruby (and suggested I could do it, he'd sponsor it, but I 
>>do not feel quite up to the task... :-) ). It would be really cool to 
>>have the xf86 patches and the kernel-patch package. Meanwhile, there 
>>are a lot more good tools available to compile a kernel, and people do 
>>that more routinely, than to compile xf86, so it is a real advancement. 
>>So, I just use Linus tree with (backstreet) ruby patches now. 
>>
>>
>>    
>>
>>>>Unless you guys find Branden's
>>>>changes reasonable, and adopt and recommend the above patch to
>>>>others that want to get this stuff running, in which case the HOWTO
>>>>could be changed to reflect the new names. So, what do you think
>>>>about that?
>>>>        
>>>>
>>>While single current TTY only allowed Branden have not a raw material
>>>about what write how-to.
>>>      
>>>
>>Uhm, I must admit I didn't fully parse that... :-) 
>>    
>>
My guess: Branden can't write a proper howto on this, as he works
with a standard (non-ruby) kernel that have the single-tty limitation.

>
>mee too, 
>Aivils, what did you had in your mind :-)
>  
>
We live in a world with many different languages, with english being
second or third or something for many of us.  I'm  happy I don't
need to write in German. :-/

Helge Hafting