java disassembler/assembler

Tom Tromey tromey@redhat.com
Thu Oct 20 22:15:00 GMT 2005


>>>>> "Daniel" == Daniel Reynaud <reynaud.danyel@wanadoo.fr> writes:

Daniel> And javap doesn't really help. The real mystery is : why do
Daniel> the few existing disassemblers all do the same, they are
Daniel> basically subsets of javap.

I'm not all that familiar with what javap can do.
For gcj we use jcf-dump, which is pretty nice... it has proved
indispensable over the years.  That said, it isn't on the gcjx branch
and I've been hoping a nice replacement would show up.

Daniel> Well, I am defining an updated version of the Jasmin language
Daniel> for that purpose. The new version would offer :

Daniel> - branch targets as offsets or labels (currently there are only labels)

Awesome, since then we could make tests with bad branches.

Daniel> If you have any ideas or features you would like to see
Daniel> included just tell me, because I am actively working on it and
Daniel> plan to release a beta version soon.

Like Jeroen, I think built-in support for the various new 1.5 things
would be helpful.  Maybe it would be nice to have a way to be able to
manage the constant pool explicitly, so we could introduce errors
there as well...?

Ideally I'd like to see this work done as part of the classpath-tools
sub-project.  That would make it a bit simpler to suck it into the
libjava tree :-).  What do you think about that?  FWIW I think there
are already tools in there using gnu.bytecode and also the ASM
bytecode library...

Tom



More information about the Java mailing list