PutImage vs BackgroundPixmap (was Re: Open Shared Memory and Render Pictures)

Carsten Haitzler (The Rasterman) raster at rasterman.com
Sun Aug 7 23:29:33 UTC 2016


On Sun, 7 Aug 2016 17:43:43 +0200 Thomas L├╝bking <thomas.luebking at gmx.de> said:

> On Sun, Aug 07, 2016 at 06:45:29PM +0900, Carsten Haitzler wrote:
> >things. there is still a race condition here as wm resizes window first, and
> >client some time later responds with a redraw.
> 
> And sometimes never :-)
> But visually, this isn't true. Either the compositor can paint last or
> (and that's probably more relevant) you configure the window w/o
> actually resizing the frame and only resize the frame when the client
> ACKs the change. This is basically what allowed us to remove gaps
> between the decoration and (supporting) clients.
> The pointer will easily go interim off-position, though.

this pointer != size can be super annoying. also from the wm/compositor side
dealing with "bad clients" that sometimes don't ack or take forever to ack can
leave your window "still resizing and waiting for client" and you end up
throwing in timeouts for poorly behaving clients.

> >actually be rather slow on some drivers
> *cough* nvidia *cough*. It's not that bad anymore, though.
> 
> For KWin I had played with lazy configurations (covered by texture
> scales in the compositor rather than in the client) - prone to cause
> some sea-sickness, though ;-)

yeah. tried that too. that was just too horrible to even contemplate shipping
at all. my god was it bad with EVERY client and its content.

> Cheers,
> Thomas
> 


-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler)    raster at rasterman.com



More information about the xorg mailing list