[ANNOUNCE] xorg-server 1.7.99.902

[Dan Nicholson]

On Thu, Mar 25, 2010 at 11:47 AM, Ruediger Oertel <ro at suse.de> wrote:
> On Thursday 25 March 2010 19:29:37 Dan Nicholson wrote:
>> On Thu, Mar 25, 2010 at 10:56 AM, Ruediger Oertel <ro at suse.de> wrote:
>> > On Wednesday 24 March 2010 17:20:24 Dan Nicholson wrote:
>> >> On Wed, Mar 24, 2010 at 06:05:27AM +0100, Stefan Dirsch wrote:
>> >> > On Mon, Mar 22, 2010 at 07:59:18AM -0700, Dan Nicholson wrote:
>> >> > > 2. Make the handling of missing sections from an existing
>> >> > > configuration behave more like the full autoconfig. In other words,
>> >> > > if there's a missing Screen section, generate multiple
>> >> > > Screen/Device pairs with fallbacks. I think this is the correct
>> >> > > fix, but I'm not that familiar with the autoconfig code so I don't
>> >> > > think I could whip up a patch that fast.
>> >> >
>> >> > I've now attached a patch to
>> >> >
>> >> >   https://bugs.freedesktop.org/show_bug.cgi?id=27229
>> >> >
>> >> > which does that. It works fine for us.
>> >>
>> >> Looks pretty good from my understanding, but I'd like to polish it a bit
>> >> more before applying it. I'm gonna do the review here since I'm not
>> >> that familiar with that part of the code.
>> >>
>> >> The first thing is that it would be great if this was a git formatted
>> >> patch with a commit message. But if it's easier to keep it as a
>> >> separate patch, can you add some information to the patch header? It
>> >> would also be great if there could be some comments in the code. It
>> >> looks like multiple Screen sections are being handled, but not really.
>> >>
>> >> Now to the patch.
>> >
>> > okay, I've reworked the patch again, thanks for all your comments.
>> > I hope the patch looks better now and is easier to read with the
>> > inlined comments.
>>
>> Thanks, it looks a lot nicer.
>>
>> > however there is an issue with the patch: if you really have more than
>> > one possible driver, the Xserver will segfault on exit/cleanup in
>> > randr/rroutput.c:RROutputDestroyResource when cleaning up the resources.
>> > I've tried several approaches to get the copied screen sections
>> > initialized better, but end up in nothing but trouble. Any help would be
>> > appreciated to clean this last glitch.
>>
>> Was it doing this before with your first patch?
> yes, it's reproducable with older versions of the patch as well ...

Is this just the normal case where you have no ServerLayout/Screen in
the conf file, or something more special where two drivers are
actually getting loaded?

>> Do you have any backtrace information?
> sure ...
> Program received signal SIGSEGV, Segmentation fault.
> 0x000000000055fe4f in DeliverPropertyEvent (pWin=0x4545454545454545, value=0x7fffffffdf00) at rrproperty.c:34
> 34      rrproperty.c: No such file or directory.
>        in rrproperty.c
> (gdb) bt
> #0  0x000000000055fe4f in DeliverPropertyEvent (pWin=0x4545454545454545, value=0x7fffffffdf00) at rrproperty.c:34
> #1  0x0000000000443e85 in TraverseTree (pWin=0x87dd00, func=0x55fe30 <DeliverPropertyEvent>, data=0x7fffffffdf00) at window.c:233
> #2  0x000000000055ff4e in RRDeleteAllOutputProperties (output=0x838fe0) at rrproperty.c:77
> #3  0x000000000049d41b in RROutputDestroyResource (value=0x838fe0, pid=<value optimized out>) at rroutput.c:413
> #4  0x0000000000456f13 in FreeClientResources (client=<value optimized out>) at resource.c:810
> #5  0x0000000000456fe9 in FreeAllResources () at resource.c:827
> #6  0x0000000000425bb1 in main (argc=<value optimized out>, argv=<value optimized out>, envp=<value optimized out>) at main.c:300

Well, I'm out of my league here. Hopefully someone else can figure out
what's going on. It doesn't seem like your code is doing anything
really different than xf86AutoConfig(), though.

--
Dan