[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Lightning] Compiling C to Lightning
From: |
Ludovic Courtès |
Subject: |
Re: [Lightning] Compiling C to Lightning |
Date: |
Tue, 23 Aug 2005 11:11:58 +0200 |
User-agent: |
Gnus/5.110004 (No Gnus v0.4) Emacs/21.4 (gnu/linux) |
Hi,
Paolo Bonzini <address@hidden> writes:
> I think Ludovic does not want that, but rather something similar to `c
> (which was based on lcc). Starting from TinyCC to get a working parser,
> could be an idea.
What's `c? Pointers? Googling for it proved to be hard. ;-)
Basically, I'd like to be able to generate a sequence of `jit_'
instructions from C code (or a subset of C). There are many good
reasons for it, one of them being that it's usually easier to write C
code than assembly.
Another good reason is that high-level language run-times may already be
written in C. For instance, Guile's run-time library defines a number
of macros for accessing Scheme objects, like `SCM_CAR ()', etc. Suppose
you want to dynamically compile code that performs an `SCM_CAR ()' on a
run-time object. Either you look at an `SCM_CAR ()' expansion and
re-implement it in Lightning assembly, which is tedious, error-prone,
and unmaintainable. Or you have a tool that directly generates a
sequence of `jit_' insns from the macro expansion.
Of course, the ultimate tool would be a GCC back-end that produces
Lightning assembly...
How did you address similar issues in OCaml and GNU Smalltalk?
Thanks,
Ludovic.