>
> > > They are generated by make and I actually should remove those others
> > > in the native/* directories.
> >
> > Ok. Are there any plans to provide separate compilation of these
> > native 'modules'? It would certainly be useful to me, but I don't
> > know if there is any other demand for it.
>
> This was asked about nearly one year ago (I can't believe it was that
> long ago) by Godmar Back I think. What is your interpretation of
> "separate compilation"?
Well in my case it would be nice to be able to tell the Makefile to build
a .so library that I can just link in with a normal UNIX executable. But
what I'm planning for my VM is also to deploy it as the kernel of an
experimental OS (where dynamic linking may not be available), and in this
case it would be nice to just a .a file to link to.
When I say 'separate' compilation I mean without compile-time dependencies
on the JVM (except perhaps header files like jni.h) and on the level of
each subdirectory (ie java.io, java.net etc).
> > > Is there anything besides jni.h that you provide with your VM that you
> > > need to build with?
> >
> > As far as I'm aware this is all that is needed.
>
> Okay, I'll probably rework the configure stuff to let you specify
> where to get jni.h from until we actually autogenerate that ourselves.
That would be most appreciated. How functional are the GTK peers at
present?
John
