Lee,
I'm struggling to understand these issues too, but from the standpoint of
trying to control the frame rate. For me, interacting with Swing objects is
somewhere in the future.
I plan to avoid calling StopRenderer() from a Behavior because I don't
understand well enough how the renderer works. I'm working on the
supposition that the execution of the Behavior's processStimulus() is itself
part of the job of the renderer. Yes? No? Maybe?
As part of my simulation, I want to maintain my own simulation clock,
because the way that the j3d (System?) clock and the Interpolators work
causes me big troubles. What I have running now, is a simulation clock
maintained in a Behavior. Its primary functions seem to work ok for simple
models, but there are mysterious secondary effects (related, I think, to the
things you are finding) that persuade me that all of that belongs in a
subclassed Canvas3D.
It's mysterious to me, because I really don't have a grasp of how the
renderer works.
I'm taking the approach that I have to intervene in the process at the
Canvas3D level in mixed-mode. Call StopRenderer() there, do what is
necessary, and then restart it. I haven't used mixed-mode before, so I'm
feeling my way along. It's not at all clear to me at this point how I
distribute the chores between the Canvas3D methods (preRender(),
postRender(), etc.) and the Behaviors I have hanging on the scene graph.
If I get anything running that's comprehensible and interesting, I'll post
it.
Fred
----- Original Message -----
From: "J. Lee Dixon" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Tuesday, September 26, 2000 9:06 AM
Subject: [JAVA3D] Update of swing object during Behavior.processStimulus()
> HELP!
> I have an application that updates several JEditorPanes with new HTML
> pages during a Behavior.processStimulus(). Frequently, right after the
> second or third JEditorPane.setPage(), my whole process hangs up. I
> have called Canvas3D.stopRenderer() before doing the page updates, then
> I do a Canvas3D.startRenderer() after the page update, but it never gets
> to this second call.
>
> I assume it is some type of race condition, but I don't know what to do
> about it. Any ideas? Should I not be updating swing components from
> within a Behavior? If not then where?
>
> This is aggravating because it does not happen all the time...
>
> -Lee
>
> J. Lee Dixon
> Software Engineer
> SAIC - Celebration, FL
> [EMAIL PROTECTED]
===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff JAVA3D-INTEREST". For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".
