Bytecode specification

[manu...@gmail.com]

Is there any draft about the bytecode generated by Gazelle.
I would like to interpret it in Java and I would like to know if I
have to hack the C interpreter, or if I can rely on a more abstract
description of how to interpret it.

Or if you're short in time, just the big picture of how it works. The
C code is somewhat clear, and Eclipse CDT is there...


Note : I still need to dig deeper, however I've found the sample
grammars bundled with gazelle both very short and very clear. It is
not cluttered by AST productions.

[Joshua Haberman]

Thanks for your interest! The bit-level format of the byte-code is
called "BitCode." It was created for the LLVM project, and is
documented here:

http://llvm.org/docs/BitCodeFormat.html

In Gazelle, that's implemented in runtime/bc_read_stream.{c,h}. If
you're interested in interpreting Gazelle bytecode in Java (which is
something I'd love to see), I'd start by implementing BitCode.

The Gazelle-specific semantics that are layered on top of BitCode are
described very tersely in docs/FILEFORMAT. The meanings of those
structures are described in the last section of the manual ("The
Gazelle Algorithm"). The manual isn't a full specification, but it
should be a good start.

Josh