Hi Michael,
You might just try using 'scheduleGVTRendering' to implement your
refresh.
This should trigger a repaint of the entire visible canvas with the
current settings.
"Scott Ruffner" <[EMAIL PROTECTED]> wrote on 02/09/2006 11:51:23 AM:
> hi, michael --
>
> i'm not sure about the exception you are getting, but
> it may have to do with the multithreaded machinery
> of Batik. do you call setRenderingTransform() in
> your GVTTreeRendererAdapter.gvtRenderingCompleted()
> method? JSVGComponent (superclass of JSVGCanvas)
> also has a GVTTreeRendererListener, and its
> gvtRenderingCompleted() method does some important
> cleanup stuff. If your listener gets called first, maybe
> that's the problem....
>
> underneath the hood of JSVGCanvas, the rendering
> transform ultimately gets reset by a call to
> JSVGComponent.setGraphicsNode(), which is invoked
> after the GVT tree build phase is done, and immediately
> before scheduling the render phase. This happens in
> JSVGComponent.SVGListener.gvtBuildCompleted():
> JSVGComponent.this.setGraphicsNode(...)
> scheduleGVTRendering() ==> calls renderGVTTree().
>
> if it is not too difficult, and if the batik experts out
> there think it is a reasonable solution, you might
> try subclassing JSVGCanvas like so:
>
> class MyCanvas extends JSVGCanvas
> {
> AffineTransform preservedRenderXfm = null;
> boolean usePreservedRenderXfm = false;
> ....
> public void refreshCanvas( SVGDocument theCurrentSVGDoc )
> {
> preservedRenderXfm = new AffineTransform( getRenderingTransform()
);
> usePreservedRenderXfm = true;
> setSVGDocument( theCurrentSVGDoc );
> }
> ....
> protected void renderGVTTree()
> {
> if( usePreservedRenderXfm )
> {
> usePreservedRenderXfm = false;
> setRenderingTransform( preservedRenderXfm, false );
> }
> super.renderGVTTree();
> }
> }
>
> This is simplified because there are multithreading issues to
> worry about (what if the user clicks "Refresh" button in rapid
> succession?). I do something like this in my Batik-based app,
> and it works for me. Doing it this way avoids having to go
> through two render phases, by the way.
>
> good luck!
>
> scott
> --
> Scott A. Ruffner, Scientific Programmer/Analyst
> Lisberger Lab
> W.M. Keck Foundation Center for Integrative Neuroscience
> University of California -- San Francisco
> 513 Parnassus Avenue, Box 0444, S871
> San Francisco, CA 94143-0444
> 415-502-7897
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
