Erwann Chenede wrote:
> Hi Alan,
> 
> Alan Coopersmith wrote:
>> Xorg in 130 and later doesn't deliver drivers under /usr/X11/lib any more
>> (the nvidia driver is still there for now - I don't know if or when that
>>  will change given they build a single package for S10 & Nevada, but the
>>  intel driver is not), but then parsing /var/log/Xorg.0.log has always
>> been
>> a bad idea - it's not even stable enough for ARC to consider it a
>> Volatile
>> interface, it's purely not an interface, and you'd fail ARC review if you
>> told them you were doing it.
>>   
> What interface can this script use ?

Why does the .so modules loaded by the server matter?   They can be renamed
or merged as needed - what's important is the functionality provided, and the
tests for glxinfo & xdpyinfo seem to do that much better.

>> Of course, now that I've looked at the script again, I'm reminded
>> that's far
>> from the worst problem in the script...
>>   
> Could you be more precise ? and/or log bugs against it

I mentioned some before in
http://mail.opensolaris.org/pipermail/jds-review/2009-February/002852.html

I'll bet you've not kept the video card id list up to date with the latest
chipsets supported by newer versions of the Intel & Nvidia drivers in the
kernel.

I was going to point out that the upcoming IPS package renaming would break
any tests you have that use the results of the SVR4 vs. IPS test, but that
seems to be never used.

Of course, /var/log/Xorg.0.log is only for the first X server, and now that
we have VT support and gdm's "switch user" function, there may be more local
X servers running.

-- 
        -Alan Coopersmith-           alan.coopersmith at sun.com
         Sun Microsystems, Inc. - X Window System Engineering

Reply via email to