gback at cs.utah.edu
Wed Jan 6 21:16:39 PST 1999
> Thanks to Godmar, kaffe now works on my linux box.
> I've decided to start adding some extensions to java.
> First is JVM support for type parametrization.
> I hope, I'll make a preview in a week.
> Any chances this work will be included in main
> source tree (for example, with ./configure --enable-extensions ) ?
> After that I'm going to try make multimethod calls...
> Just to train myself before adding support for prolog engine
> of kiev compiler, wich will be much more complex... ;-)
> Peter, you was intrested in it, wasn't you?
> Will someone help me, when I will be ready, with
> kaffe.def file ?
My personal opinion is that you are certainly encouraged to use kaffe
for your experimentation and participate in its development. We are happy
about everybody who wants to and can contribute.
On the other hand, I personally think it is fair to say that not
everybody would have use for an embedded prolog engine in kaffe.
Not in the short run, at least.
What I would do is write another engine (and then use
--with-engine=kiev or some such to configure)
I realize that some of the stuff in kaffevm has engine specific #ifdefs like
#ifdef TRANSLATOR etc. For modularity, it may actually be a good idea
to factor those out in separate functions and make them explicitly
part of the ``engine'' interface. I think that that should be possible
without incurring too much overhead, and it would be a very welcome
contribution to kaffe anyway.
For those cases where you'll have to change code to do verification,
etc., I'd just duplicate it for your engine, overriding the version in
kaffevm. This would be the cleanest way of doing it, and would keep
the diff distance between your tree and the pure java kaffe tree
More information about the kaffe