[kaffe] Future directions for Kaffe

[Dalibor Topic]

Guilhem Lavaux wrote:

>* I have a partly working new jit3-v2.0 which is able to cope with 64
>bits arch. AMD64 is still giving some failures in the test suite
>(something like 30 failures for 160 tests). I will update the other jit3
> when I have this one working. I hope to anyway check it in CVS after
>the next release.
>
>  
>
Nice! One more reason for me to get that release out of the door.

>* AWT needs some work. I have not had time to check by myself but
>Riccardo reported there was problems running it. It will also be great
>to finish porting our X native peerless AWT to make it work with
>Classpath's "Swing" toolkit.
>
>  
>
I don't think our native X peers have a bright future in light of Sun 
opening up their own XAWT
peers in the next months.

The AWT problem we have is that the GNU Classpath version in Kaffe needs 
to have its Makefiles
fixed in order to work with automake 1.10, as the linker currently drops 
some libraries due to
changes in automake 1.10, leading to AWT applications breaking all over 
the place when run
with our own build of GNU Classpath. That in turn is best solved by 
removing GNU Classpath from
Kaffe, and just using a prebuilt version, rather than fighting with its 
build system. Or alternatively, by
fixing its Makefiles. :)

I'd prefer to 'spin off' the peerless AWT into its own little project on 
Kaffe.org, if someone wants to
maintain it, and try to spin off non-VM functionality from the tree from 
the next release on into separate
projects from then on. In my opinion, it does not make sense to have our 
own libraries that we don't
have the volunteers to maintain.

cheers,
dalibor topic