6800 C compiler survey results

[toma]

	I have been conducting a search for 6800 C compilers, in an
effort to obtain support for the Motorola 6801 and Hitachi 6301 uC's.
By far the best (and really the only) stuff I've found was a re-write
of Ritchie's C compiler by two people formerly at the University of
Delaware, Dave Sincoskie and Rich Hammond.  (from the Delaware USENIX
tape, /delaware/udel-ee/mc/...).  The source was non-functional when it
came off the tape, since it used a defunct stdio library and contained
some pdp-11 dependencies (we're running vax-11/780's); however, I
was able to get both passes running in just a few hours and it appears
to be a very good base to start from.  Unfortunately, the state of C
has changed considerably in the interim; e.g., unions, '=' style
initializations, typedefs, etc. have been added.

	I am about to embark on upgrading the compiler to use the 
additional instructions provided by the 6801/6301.  It is difficult to
say how far I'll get.  First, though, I've started the merge process
of the U of D compiler and a recent version of the pdp-11 compiler (a
large task...) to incorporate some of the new features of C. 

	Regrettably, I don't think we'll be able to provide any source
(presuming I eventually get this thing in reasonable working order)
outside the company, since the compiler is adapted from a licensed
source.   Our licensing agreements allow us only internal usage of
source and binaries, and that on specific machines.  The best advice
I can offer is that if you are interested in a 6800 C compiler, get
ahold of the tape and fix the i/o (the main problems are with getw(),
putw(), the fact that EOF changed from 0 to -1, and a peeksym variable
(containing the next symbol) which contains -1 for NO_SYMBOL_AVAILABLE
and 0 for EOF).

	However, if you are interested in a 6801 or 6301 C compiler
(as I am)  send me a letter expressing your level of interest 
(just curious, would like to have one and are willing to pay for
it, would trade your grandmother for one).  If there is enough 
interest, I'll look into what it would take for fluke to distribute
any results (e.g., licensing hassles).


				Tom Anderson
				John Fluke Mfg.
				Everett, Wa.