Hi all,

thelarsons3@cox.net wrote:
>>From: Dalibor Topic <robilad@kaffe.org>
>>
>>
>>We're working on improving support for m68k-netbsd in the current CVS head.
>>Current status is here:
>>http://multix.dyndns.org/Sites/kaffe-devel/machines/peregrin/Peregrin.html
>>but we could use some m68k/netbsd hackers for sure :)
> 
> 
> So the current devel branch is not working (yet).  Maybe I can help in some small way with that.  For now, what is the most current working kaffe for mac68k?  Where can I get it, and learn what it supports?

there is apparently a package of 1.0.5 for netbsd-m68k, but that's kaffe 
from a few years ago :(

We're currently busy hunting down all sorts of compiler warnings in 
order to make sure that the kaffe core becomes more stable accross 
platforms. Help on that would be quite appreciated, and doesn't require 
one to be a Java hacker :)

The current problem with m68k-* is, afaik, that the 
kaffe/kaffevm/verify.c assembler file ends up being too large for the 
assembler, so  you get interesting warnings from the assembler [1] and 
crashes later.

So a first step could be to refactor verify.c into smaller bits. I don't 
know if m68k has some object size limit, of if gcc can be persuaded to 
generate code that doesn't cause assembler warnings. Any input from 
m68k-netbsd experts is heartily welcome.

cheers,
dalibor topic

[1]

/var/tmp/ccwOFswR.s: Assembler messages:
/var/tmp/ccwOFswR.s:37503: Warning: .word .L3537-.L3733+0 didn't fit
[snip]
about 20 of those. No idea what they are good for, but they sound like 
problems :)