[Spice-devel] Questions about experimental Spice compositor rebase

Yury Shvedov shved at lvk.cs.msu.su
Thu Mar 24 16:38:57 UTC 2016


Yes, this is the most significant problem of this project. I do not have 
time at all now. All I can to do now is to suggest you to try put the 
line to the repaint function I wrote you some time ago and put an debug 
output about the damage argument.

You will see then how does it affects to the repaint's calls frequency 
(repaint function sends data on each call). And the image on the client.

Compare spice repaint function with rdp one.

All that things I suppose to do further.

On 03/24/2016 07:31 PM, Fabio Fantoni wrote:
> Il 24/03/2016 17:13, Yury Shvedov ha scritto:
>> Fabio, can you please explain me, why do you continue performance 
>> testing the thing what already known to have an awful performance and 
>> the way of fixing that?
>
> You right, I must try to do significant changes instead (probably 
> following rdp compositor and/or other compositor), for example trying 
> to "separate" spice operations for delay repaint without create 
> latency in spice or any other useful improvements based on what I'll 
> learn in the mean time.
> All depends on how much freetime for it I'll have in the weekend 
> (unfortunately only too short time lately).
>

-- 
Kind Regards,
Yury Shvedov



More information about the wayland-devel mailing list