Jean-Christophe ARNU (JX) wrote:
Le Mon, 17 Nov 2003 13:37:10 -0500
Thomas DeWeese <[EMAIL PROTECTED]> me disait que :
Jean-Christophe ARNU (JX) wrote:
Le Mon, 17 Nov 2003 12:46:33 -0500
Thomas DeWeese <[EMAIL PROTECTED]> me disait que :
Ok wild shot in the dark. Do you use invokeAndWait on the update
thread from the event listener thread?
Will look at what I can do with it...
Looking at the thread trace I don't think this is the problem.
Do you print exceptions? The Default User agent may not print all
Exceptions. It looks like someone is 'dropping the ball' when the canvas
gets resized. The only way to debug this tends to be scatter System.err
messages in the various methods and see where the ball is dropped.
Where can I find how the default useragent handles exceptions? (this was the
next point I would discuss on after my resizing/zooming problem). How can I
change UserAgent to get it all exception printed?
Do you think it's necessary to modify JSVGCanvas code to get system.err
traces or juste inherit UserAgent to print all exception?
If you use the SVGUserAgentAdapter or GUIAdapter it will print all
messages. You can provide it in the constructor for the JSVGCanvas.
I would try this first. If it doesn't make it clear what the problem is
then you will need to start modifying the JSVGCanvas code to insert
System.err traces. With Ant it's really trivial to build Batik from
source so it shouldn't be a big deal. You will probably want to
create a 'build.properties' file (in xml-batik) with the one line
'debug=on' so it provides line numbers in stack traces.
If you are currently working from a binary distribution you can
use 'build jars' (same structure a binary dist) or 'build all-jar'
put all of batik in one large jar file (on unix it is 'build.sh' not build).
So the entirety of the bug seems to be in the 'resizing' code. I
think>a thread dump would be really helpful.
Full thread dump Java HotSpot(TM) Client VM (1.4.2_01-b06 mixed mode):
OK so I was wrong, it wasn't helpful :)
Well, every tracks might give its own answer :) Anyway thank you for your
help :)
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
