>
> Hi!
>
> Well, I decided to start a mailing-list dedicated to writing development
> tools for the PIC16C84 (and other PICs too) which will be freely available
> in CopyLeft'd C source code.

Cool! Add me to it.

>
> The first priorities of the list will include:
>
>   o Improving Timo Rossi's assembler to fully handle forward references.
>     Probably, we'll need to add a "pass zero" to build a symbol table.
>  =20
>   o Converting John Favata's compiler to output PIC assembly instructions
>     instead of Parallax's pseudo-8051 code.  It may be best to follow
>     John's strategy, but the assembly code will probably need some
>     optimization once it's converted.

I'd like to add a couple of things to the table. I'm a computer science
professor in Computer Science at Clark Atlanta University (obviously in
Atlanta ;-) Anyway for the last 3 years I've been teaching a class on the
development of Computer Systems Software. We focus on assembler and an
introduction to compilers including projects. Since we professor types
are obligated to at least have a cursory understanding of the projects ;-)
my students and myself have put in quite a lot of work into said projects.
Over the last 3 semesters we have been targeting PICs. The results:

- A fully functional 16C84 assembler. It's one pass and does forward
references. It currently doesn't output the symbol table for linking
purposes but is easy enough to add.

- The foundation to a high level language called NPC (nano-pseudo-C). NPC
strips most of the functionality from C and leaves just the absolute
minimum. It adds some keywords (ENDIF, ENDWHILE, CALL) to simplify the
parsing process. It does most of the C operators and adds a couple
(Like direct bit manipulation and direct assignment of addresses to
variables). It only has pointers and chars right now. I had been targeting
the 17C42 and was about 30% finished with the code generation. It uses
flex for a front end (any GNU problems there?) and is recursive descent
(Didn't want to confuse the students with bison). The heart of it is
the infix-postfix parser that makes code generation straightforward. This
semester I'm splitting the project in two adding a lower half interpreter.

It's all in C and should be no problem to release under GNU. Runs under
Linux right now.

Anyway put me on the list and let me know if there is anything I can do
to help.

BAJ