[net.arch] Optimizing assemblers

stuart@BMS-AT.UUCP (Stuart D. Gathman) (09/10/86)

The first thing I'd do for any program development system is to
make a smart assembler so that the code going in is as simple
as possible.  I have often wanted to write a 6502 C compiler
in my spare time :-) and I amuse myself with the preliminary
design: an assembler that generates long and short branches,
keeps track of flags to eliminate set instructions where possible,
defines macro instructions that use zero page as a register bank
with stack pointer, index register, etc. etc.

An assembler should smooth out the indiosyncasies of a machine
while allowing access to specific instructions (but probably
through a macro) for hardware functions.

P.S.  I have an irrational fondness for my first processor.
-- 
Stuart D. Gathman	<..!seismo!{vrdxhq|dgis}!BMS-AT!stuart>

peters@cubsvax.UUCP (Peter S. Shenkin) (09/11/86)

In article <BMS-AT.200> stuart@BMS-AT.UUCP (Stuart D. Gathman) writes:
>P.S.  I have an irrational fondness for my first processor.

H. L. Mencken wrote (as well as I can remember):
	Men have a special fondness for their first loves;  after that
	they tend to bunch them.

Peter S. Shenkin	 Columbia Univ. Biology Dept., NY, NY  10027
{philabs,rna}!cubsvax!peters		cubsvax!peters@columbia.ARPA

jlg@lanl.ARPA (Jim Giles) (09/12/86)

In article <BMS-AT.200> stuart@BMS-AT.UUCP (Stuart D. Gathman) writes:
>P.S.  I have an irrational fondness for my first processor.

I have an irrational distaste for my first processor - it was an IBM
360.  Did I say 'irrational'?

J. Giles
Los Alamos