···@best.com (.) writes:
>Is anybody actively working on a Lisp-to-JavaVM compiler? I have
>heard several rumors, but nobody has been willing to step forward.
This is a scheme system that is written in Java.
http://www.winternet.com/~sgml/kawa/index.html
I don't know anything more, but it seems it is a pure interpreter, not
a compiler to Java-VM Code.
>The Java community gives us a widely available, pretty efficient
>runtime with extensive native libraries. Compiling Lisp to it
>isn't all that hard--it just needs to be done...
After a first look at the Java VM it seems (for me) that representing
dynamically typed values might become very inefficient. One would have
to write a boxed representation with just Java VM instructions, no
chance to use native machine instructions. Do you have more specific
ideas about this?
Martin
--
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Martin Cracauer <········@wavehh.hanse.de> - BSD User Group Hamburg
BSD, Lisp and other programming info http://www.bik-gmbh.de/~cracauer
From: erik hilsdale
Subject: Re: Lisp-to-JavaVM compiler
Date:
Message-ID: <10402@832179981>
In article <······················@wavehh.hanse.de>,
Martin Cracauer <········@wavehh.hanse.de> wrote:
>···@best.com (.) writes:
>
>>Is anybody actively working on a Lisp-to-JavaVM compiler? I have
>>heard several rumors, but nobody has been willing to step forward.
>
>This is a scheme system that is written in Java.
>
> http://www.winternet.com/~sgml/kawa/index.html
>
>I don't know anything more, but it seems it is a pure interpreter, not
>a compiler to Java-VM Code.
And it is non-tail-recursive (at least it was the last time I
looked at it). One of the most annoying things about doing any
Scheme/Lisp-java conversion (imho) is the gymnastics necessary
for keeping tail-recursive behavior.
-erik
--
erik hilsdale (········@cs.indiana.edu)
http://www.cs.indiana.edu/hyplan/ehilsdal/