Stackless/microthreads merge news

[Andrew Kuchling]

Christian Tismer <tismer at tismer.com> writes:

> Glyph Lefkowitz wrote:
> > 
> > Christian Tismer <tismer at tismer.com> writes:
> > 
> > > Well, I got the Kaffe sources. This was easy.
> > 
> > Yay open source!  Kaffe is really quite nice.  I hear they've even
> > fixed the lack of BigDecimal, etc...
> > 
> > > Now I'm trying to get the Blackdown Sources, since I'd like to read
> > > the green threads implementation.  These should be sitting in some
> > > patches, I think.  Whether I'll work on either of them is still
> > > open, but I want to read them.
> > 
> > I'd recommend against reading them.  They will make your brain dirty,
> > in two ways: you will have to look at Sun source code (green_threads
> > was not implemented by blackdown) and you will end up being unsuitable
> > for cleanroom efforts such as Kaffe and Classpath.
> 
> Aaahh. Good advice. Is Classpath something where I
> should look at, too?

Probably not; it's an effort to write an LGPL'ed version of Sun's
APIs, so no VM work is required. (http://www.classpath.org)

BTW, another Java VM is Japhar (www.japhar.org).  I can't say which VM
is more complete, or more commonly used.  And can either of them
handle JPython at this point?

(It's surprising how anemic the free Java projects are.  Classpath's
most recent snapshot release is dated 1999-02-06, over a year ago; the
last Kaffe release was in October, and the last Japaher release in
March 1999.  I would think that a Java VM + API, because the APIs are
so large and can be developed incrementally, would be really suited to
frequent smaller releases.)

--amk