(no subject)

[cbj]

Sender: address@hidden
To: Artur Biesiadowski <address@hidden>
Cc: ClassPath List <address@hidden>
Subject: Re: Classpath <-> gcj integration
References: <address@hidden>
From: Brian Jones <address@hidden>
Date: 03 Jan 2001 09:00:42 -0500
In-Reply-To: Artur Biesiadowski's message of "Wed, 03 Jan 2001 02:10:48 +0100"
Message-ID: <address@hidden>
Lines: 31
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii

Artur Biesiadowski <address@hidden> writes:

> What is reference JVM used to test Classpath ? Japhar ? I mean what
> sanity check I should perform after changing some core classes like
> Thread, Class, ClassLoader, String etc.

Since no one else answered this one I'll give it a try.  Everything
we've currently designated as VM dependent (see vm/reference/) would
probably be tested on a VM by VM basis.  I know for a fact that we
have threading issues with Japhar but what I don't know is if it is
because we're broken or Japhar is broken... but it is usually us.  :)

I've mentioned to Tom recently in private email that we should
probably just plan on having both CNI and JNI implementations in
Classpath.  Kaffe and gcj can both use CNI, which makes JNI less
important going forward unless the SableVM, the Intel VM, or some
other VMs really like having that.  

Further, I would rather that someone merge the implementations between
gcj and Classpath without regard to not having this or that
implemented in JNI (just CNI) and someone else can fill in JNI methods
as they come up.  The thing that I need to do though is make a place
we can put CNI related files and then try to add a few things to
support a gcj build of classpath w/CNI; of course anyone is welcome to
beat me to this.  I'll probably release classpath 0.02 shortly with a
Japhar that properly prints exceptions before I work on this CNI build
thing.

Brian
-- 
Brian Jones <address@hidden>