Re: [Gcl-devel] Re: gcl bad code generation for setf fill-pointer for fixnum value

[Tim Josling]

Thanks Camm

On Tue, 2007-12-18 at 11:04 -0500, Camm Maguire wrote:
> Greetings!
> 
> All of thes ework in cvs head:
> 
> BTW, are you interested in helping with the long-standing goal or
> porting gcl as a new gcc front end for common lisp?
> 

I am currently using gcl to build a front end for gcc, for the language
COBOL. That is my priority now. I have left my job and I'm working full
time on this. There are three objectives in this project

1. Become a highly skilled lisp programmer.
2. Produce the COBOL compiler.
3. Proof of concept of lisp as a highly productive programming
environment.  

Once I have that done, I would probably like to contribute to GCL in
some way. 

Potentially making GCL into a GCC front end would be one option. I don't
know if that would be useful, or not. There may also be potential to
generate better code and better debugging information. I think, again
without much research, that the next priority for gcl might be ANSI
compliance, or perhaps more comprehensive compiler and runtime
diagnostics. But I would be discussing the priorities with others before
embarking on anything.

As for the code generation, gcl generates pretty good code as far as I
can see. For example, for "CASE" with fixnum variable and constant test
values, it generates a fast switch statement. Most of the other free
lisps don't seem to do that.

Regards,
Tim Josling