From: christopher.g.dorosky++at++lmco.com
Date: 12/07/2000 10:17:33
I didn't run your code, but I think you problem may be with the bounds. When you set them static, you prevent automatic recalculation of the bounds, when the item is moved.
The man page for pfNode (I think) goes over the difference between the PFBOUND_STATIC and PFBOUND_DYNAMIC.
The central problem (if this is the case) is that the cull will cut out objects that it doesn't think fit in the viewing frustum. If the bounds are incorrect, you can have things disappearing, based on your view angle.
I would reccommend trying DYNAMIC, or you can stick with static, and set the bounds yourself, each time the object is moved.
Good luck.
Christopher Dorosky
Lead Electronic Systems Engineer - Real Time Simulation
Lockheed Martin Missiles and Fire Control - Dallas
christopher.g.dorosky++at++lmco.com
972-603-2349
-----Original Message-----
From: Roy Ruddle [mailto:Ruddle++at++Cardiff.ac.uk]
Sent: Thursday, December 07, 2000 11:31 AM
To: info-performer++at++sgi.com
Subject: PfChannel bug, I think
Hi there,
After a week of investigation I've narrowed down a problem I've been
experiencing to something quite simple (see attached file). I'm pretty
sure it's a bug in Performer. If it isn't, I'd love to know what I've done
wrong.
In my PF application, I have 2 users collaborating to manipulate a large
object + this means they are looking towards each other. In the example
above the scene consists of one rectangular linestrip. The bottom channel
is looking towards the rectangle with a view heading of 180 deg. The top
channel orbits around the rectangle but always looks towards it (whacky,
but good for illustrating the problem I've found).
If you run the example you have to choose option 0 or 1.
With option 1, the example works as it should.
However, with option 0 you never see the rectangle in the bottom channel,
and it is only visible in the top channel when the view heading is close
to zero.
The difference between options 1 and 0 is that with the latter the
rectangle's vertices are initially all at [0,0,0] but get set to their
proper positions, and the bounds recalculated, before pfFrame() is called
for the first time.
Although the above illustrates the problem I'm having, it isn't exactly
the same. In my full application, I only lose some (not all) of the
geometry in a channel and, for example, I can lose lines or tristrips.
I've also replicated the problem on two completely different SGIs
(MaxImpact/Irix 6.5/PF2.2.7 and a Crimson RE/Irix 6.2).
If anyone can suggest a work around I'd be very grateful
cheers
roy
------------------------------------------------------------------------
Roy Ruddle, Senior Research Associate
Cardiff Human Interfaces and Virtual Environments Laboratory (C-HIVE)
School of Psychology, Cardiff University, PO Box 901, Cardiff CF10 3YG
Tel: +44 (0)29 2087 5030 (direct), Fax: +44 (0)29 2087 4858
Email: Ruddle++at++CARDIFF.AC.UK http://www.cf.ac.uk/psych/ruddle/
This archive was generated by hypermail 2b29 : Thu Dec 07 2000 - 10:16:51 PST