I started out making changes to the interpreter. Mostly, it involved copying/pasting INVOKESTATIC in kaffe.def and adding a couple asm(...) statements. But, it turns out I need to use the JIT, which is not so straightforward. I've been working on it some and think I have an idea of how I need to do things differently, but debugging is going to be tough. As a rudimentary way of debugging, I have been trying to find a way to dump the JITted method so I could look at that, but haven't been able to find a good way to do that.
I expect implementing the new code could be done really cleanly (and quickly) by someone who really understands how the JIT is implemented. I'm just having a tough time figuring that out.
If you can help me out some, I'd really appreciate it. If I get it all figured out, I'd be willing to write up some documentation on how I did what I'm working on so if/when other people try to do something similar they could suffer a little less than me.
About my environment:
My PC is intel/linux
Kaffe is compiled to alpha/digital unix
I'm running the JVM on an alpha simulator on the intel machine
This makes debugging much tougher because if I run gdb I'm debugging the simulator, not kaffe. I might be able to run gdb on the simulator, but I'd like to avoid trying to set that up unless absolutely necessary. It took me long enough to get kaffe running on it.
TIA, Jeremy
_______________________________________________ kaffe mailing list [EMAIL PROTECTED] http://kaffe.org/cgi-bin/mailman/listinfo/kaffe
