This is the mail archive of the
java@gcc.gnu.org
mailing list for the Java project.
Re: Multiplatform object model
- To: Grobarcik Peter <Peter dot Grobarcik at start dot de>
- Subject: Re: Multiplatform object model
- From: Jeff Sturm <jsturm at one-point dot com>
- Date: Mon, 26 Mar 2001 18:37:22 -0500 (EST)
- cc: "'java at gcc dot gnu dot org'" <java at gcc dot gnu dot org>
On Mon, 26 Mar 2001, Grobarcik Peter wrote:
> > As I see it - there is still a problem doing system development with C++,
> > because of the internalls of different compilers, name mangling and other
> > non standartised things.
Not that this is a C++ list, but...
It's true that C++ has a history of binary incompatibilities, however
the new ABI in GCC 3.0 is meant to solve that. It's an industry standard,
not particular to GNU.
> > maybe the Java bytecode or something simmilair (GNU bytecode :-) ) could
> > solve the issue: all compilers would produce some standartised bytecode
> > and then the OS semi-compiler would finalize the task (maybe JIT) - so we
> > can have a fast AND multiplatform code - something like GCJ, but for "all"
> > languages.
> >
> > Would be such a thing possible?
Strictly speaking, it is certainly possible. This is often called an
intermediate language, or IL. It is also a very old idea.
> > Has it sense?
In my opinion, no. We need better source compatibility so that a
recompile becomes nothing more than e.g. "../configure; make; make
install". GNU gives us that, mostly.
Why would you need a portable code format other than source? There are
sound technical and philosophical reasons why a portable IL would
probably not be a good thing for GNU.
Jeff