[kaffe] Small arraycopy patch, may have effects on arraycopy performance

[Dalibor Topic]

Hi all,

In order to try to track down what's causing the trouble on the 
alpha-linux-interpreter tinderbox [1], where the exception is thrown 
from within arraycopy, but doesn't contain useful information to decode 
it, I've moved the error handling code out of the native part of 
arraycopy into a new java part of the method.

While my own measurements using SciMark2 on i386 showed no performance 
decrease or increase by the change, if they do for you, I'd like to hear 
about it.

cheers,
dalibor topic

[1] 
http://tinderbox.anholt.net/tinderbox3/showlog.pl?machine_id=13&logfile=20050102170656.log