[kaffe] Re-used event objects (again)

[Benja Fallenstein]

This is a multi-part message in MIME format.
--------------050209010209070207040405
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit


Dalibor,

Dalibor Topic wrote:
> In the case above, I think patching kaffe would be
> better than working around bugs.

Thanks for your answer. It took me some time to figure out, but in the 
end the patch was really trivial (attached). The reason was that 
KeyEvt.dispatch() reused the current KeyEvt object for KEY_TYPED events 
right after issuing KEY_PRESSED.

I've changed it so that it will create a new KeyEvt object. It will even 
re-use a cached object if caching is activated (it's off by default, see 
java.awt.Defaults.RecycleEvents).

Gzz (compiled against IBM's JDK so far) runs for me under Kaffe, now. 
Success! :-)

Thank you,
- Benja

--------------050209010209070207040405
Content-Type: text/plain;
 name="keyevent.patch"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
 filename="keyevent.patch"

Index: libraries/javalib/java/awt/KeyEvt.java
===================================================================
RCS file: /cvs/kaffe/kaffe/libraries/javalib/java/awt/KeyEvt.java,v
retrieving revision 1.6
diff -u -r1.6 KeyEvt.java
--- libraries/javalib/java/awt/KeyEvt.java	24 Jul 1999 00:56:12 -0000	1.6
+++ libraries/javalib/java/awt/KeyEvt.java	9 Jan 2003 22:52:17 -0000
@@ -39,9 +39,8 @@
 			if ( !consumed && !ShortcutHandler.handle( this) ) {
 				if ( keyChar != 0 ) {         // printable key
 					if ( AWTEvent.keyTgt != null ) {     // maybe a fast finger pulled the keyTgt under our feet
-						id = KEY_TYPED;
-						keyCode = 0;
-						AWTEvent.keyTgt.process( this);
+						KeyEvt typedEvt = getEvent((Component)source, KEY_TYPED, 0, keyChar, modifiers);
+						AWTEvent.keyTgt.process(typedEvt);
 					}
 				}
 			}

--------------050209010209070207040405--