Hi Martin,

Martin Constantine <[EMAIL PROTECTED]> wrote on 03/16/2007 06:24:06 
AM:

> I've recently upgraded batik from 1.6 stable to svn  r518077. The 
> application permits users to "scribble" over an svg canvas during a 
> presentation. Scribbles are made with either the mouse or with a tablet 
> pen on an Overlay implemented class which also implements the mouse 
> listeners (very similar to batik's Interactor). This overlay is 
> registered with the JSVGDocument so that scribbles are saved. Updates to 

> the document occur on mousePressed, mouseDragged, and mouseReleased. The 

> GVT tree is updated on mouseReleased where the entire stroke is appended 

> to the current stroke group.

   If I recall correctly I think the Tablet version of Windows is
'mouse move' event happy (presumably for text recog).  This is
probably why it lags on a Tablet but not on other devices.

> Question: Has something changed in the update manager (or somewhere else 

> like the AnimationEngine) that would cause this performance hit? Here's 
> some evidence that suggests that batik is chewing up more CPU cycles in 
> this newer version than the 1.6 stable release:

   It wouldn't surprise me if the update manager had extra work to do
to support animations (like update the state of any animation elements
before redrawing the display).

    That said I'm a little surprised you modify the document at all...
If you are drawing in an Overlay you can (and probably should) modify 
the Overlay in the Swing thread.  What are you changing in the Document
that requires you to do your changes in the UpdateManager thread?


> Our app records the speaker's voice along with scribbles, so running 
> another sound recording application simultaineously produces some 
> contention for sound resources. The older version of the app running 1.6 

> stable does well even in the face of this contention. No pen stroke 
> slowdown issues. However, the new version does not. Pen stroking 
> performance is clearly reduced in the face of this contention. When the 
> other program is removed, pen performance increases to a tolerable 
level.
> 
> Any insight would be greatly appreciated. Thanks.
> 
> Martin.
> 
> 
> 
> ---------------------------------------------------------------------
> 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]

Reply via email to