From: Angus Dorbie (dorbie++at++sgi.com)
Date: 02/16/2000 11:53:47
If I'm trivialising the input problem you are trivializing the
performance issues.
There's a fundamental design issue with transparently supporting
multiple pipes in this manner which cannot be resolved by a mere
implementation of the concept.
Cheers,ANgus.
Chris Chuter wrote:
>
> Now hold on there. I'll admit I'm a little ingnorant on xinerama (I'll know
> more when I get it running my Linux box - XFree86 4.0). But, I think you're
> trivializing the problem. It's not just an origin reset in most cases - you
> actually lose XEvent information when crossing screens - try this with perfly
> on a dual screen.
>
> load an object
> Do a motion event that crosses the screen (i.e. hold the 1st button down and
> cross the screen)
>
> What you'll see is that the object gets bad x,y data, but what's worse is that
> it doesn't update the x,y as you keep performing the motion event. This is bad.
> Here's some printf output of the Motion event to prove it:
>
> mouseMoved: x:1273, y:471 same screen:1
> mouseMoved: x:1276, y:471 same screen:1
> mouseMoved: x:1277, y:471 same screen:1
> mouseMoved: x:1278, y:471 same screen:1
> mouseMoved: x:1279, y:471 same screen:1
> mouseMoved: x:0, y:1024 same screen:0
> mouseMoved: x:0, y:1024 same screen:0
> mouseMoved: x:0, y:1024 same screen:0
> mouseMoved: x:0, y:1024 same screen:0
> mouseMoved: x:0, y:1024 same screen:0
>
> You don't get correct x,y values unless you release the button and re-pick.
> This is a major pain for Performer programs that use a lot of picking (like
> ours).
>
> I agree with your other point 'Multipipe performer scales well', thats why
> we're here ;)
>
> If you have any solutions to the above problems, I'm all ears. Otherwise, I
> anxiously await SGI's move to X11R6.4 or I'll do some porting to Linux.
>
> -chris
>
> --- Angus Dorbie <dorbie++at++sgi.com> wrote:
> > Well it might add a few problems of it's own.
> >
> > The xinerama stuff AFAIK is a single stream of input.
> >
> > The origin reset is easily fixed in software by tracking input window
> > and the multipipe performance scales well fairly transparently with
> > performer, that's one of the huge advantages of using Performer.
> >
> > Carefull what you wish for.
> >
> > Cheers,ANgus.
> >
> > Chris Chuter wrote:
> > >
> > > Are there any plans for SGI to upgrade to X11R4 or higher in the future?
> > We've
> > > been sitting on X11R3 for a while. X11R4+ promises Xinerama which would
> > solve
> > > many of the multiple head problems (mouse x,y getting reset, dragging
> > windows
> > > across, etc.).
> > >
> > > -chris
> > >
>
>
> =====
> Chris Chuter
> Software Engineer
> Mojave Computer Graphics
> cchuter++at++texas.net
> http://www.texas.net/~cchuter/
> __________________________________________________
> Do You Yahoo!?
> Talk to your friends online with Yahoo! Messenger.
> http://im.yahoo.com
> -----------------------------------------------------------------------
> List Archives, FAQ, FTP: http://www.sgi.com/software/performer/
> Submissions: info-performer++at++sgi.com
> Admin. requests: info-performer-request++at++sgi.com
-- For Performer+OpenGL tutorials http://www.dorbie.com/ My comment on the abuse of Jon Johanson's rights; After giving up raiding their neighboring countries the Norse men have taken to raiding 15 year old kids in their bedrooms. Very sad.
This archive was generated by hypermail 2b29 : Wed Feb 16 2000 - 11:54:09 PST