FYI: RMI

[Tom Tromey]

I am checking in the Transvirtual RMI implementation.

I started with the RMI implementation in Kaffe 1.0.6.
Then:

I relabelled classes from kaffe.rmi.* to gnu.java.rmi.*.
I removed all reference to the kaffe.util.NotImplemented exception class.
I changed the headers of all the files to reflect the correct (as I
understand it) copyright status: copyright FSF, part of Classpath,
etc.  I removed the old copyright info.
I wrote all the new Makefile.am's and I've updated configure.in.

I don't have a particularly good way to compile all this code, at
least not in the Classpath tree :-(.  I did build it all in the libgcj
tree, so it does at least compile somewhere.  I'm willing to address
problems I've introduced; send a bug report.

I have no idea whether the code actually works.  I haven't done any
actual testing of it.  In fact, I really don't know much about RMI at
all; I've simply done the import work.  I leave it to others (or
perhaps myself, at some future date) to determine the robustness of
this code.

One known issue is that the rmic implementation relied on kaffe code
outside the copyright assignment (as I interpreted it).  This is the
code that actually invokes the underlying Java bytecode compiler.
This is code that must be reimplemented; until that is done rmic
simply won't work.  Doing this reimplementation is on my to-do list.
I hope to get to it soon.


Many thanks to Transvirtual for assigning this code to the FSF.  This
is a substantial contribution to the Classpath project.

Tom