Reconsidering gcjx

[Paolo Bonzini]

> (2) A bytecode versions of ecj.  This is only useful if we also make
> available a bytecode version of libgcj, I think.

For what is worth, this is how OCaml is distributed.  It would be ok for 
me to need a (separate) download of two jars for ecj and libgcj in order 
to bootstrap Java.  If absent, we'd need a pre-existing gcj installation.

We could have the stage1 compiler just compile the bytecode->native Java 
front-end, and use it to compile both to native code.  There is however 
a bootstrapping time problem, if we want to compile libstdc++ and libgcj 
three times.

How big would the mini Java-runtime be?  A bytecode-interpreter, with 
only support for two or three packages, using a simple Baker or 
mark'n'sweep GC, could be done in 10,000 lines of C code or maybe less.

Paolo