[PATCH 2/2] damageext: Xineramify (v4)

[Aaron Plattner]

On 10/09/2013 09:29 AM, Adam Jackson wrote:
> On Tue, 2013-10-08 at 16:43 -0700, Aaron Plattner wrote:
>> On 10/08/2013 03:28 PM, Aaron Plattner wrote:
>
>> Never mind, it looks like this still isn't properly translating the
>> damage by the root window's position.  Was it intentional that v3 of the
>> patch dropped the "Properly translate coordinates for window reports in
>> Xinerama mode" part of v2?
>
> I hadn't thought I'd need it, since I hadn't considered the origin not
> being at +0+0 to be a thing people would do.  Easy enough to revive
> though.

Well, the physical X screen 1 is at +2560+0.  The reported position of 
damage to the root window resets to (0, 0) as soon as the damage crosses 
from physical screen 0 to physical screen 1, but that damage is 
occurring at (2560, 0) w.r.t. the logical root window as far as clients 
are concerned.

>> Also, the initial damage doesn't cover the whole Xineramified root window:
>>
>> $ xdpyinfo | grep dimensions
>>     dimensions:    4480x1440 pixels (1222x385 millimeters)
>> $ ./rootdamage
>> Damage 2560x1440 @ (0, 0)
>
> Which tree are you testing against?  In particular, do you also have:
>
> commit 7d3d4ae55dd6ee338439e2424ac423b1df80501b
> Author: Keith Packard <keithp at keithp.com>
> Date:   Fri Sep 20 14:43:41 2013 -0500
>
>      damage: Must translate initial window damage by window offset

Yes, I'm testing with "dix: Add PostDispatchCallback" and "damageext: 
Xineramify (v4)" applied to commit 6a9bd103cb993e873cb82664b677ceed73c0bd85.

I don't *think* our driver is messing up the root window positions 
somehow, but I'll try testing with a different driver.

-- 
Aaron