Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!rutgers!seismo!mcvax!enea!toma
From: toma@enea.UUCP
Newsgroups: comp.os.cpm
Subject: Re: C-Compiler for CP/M-80 wanted
Message-ID: <1664@enea.UUCP>
Date: Tue, 27-Jan-87 15:41:54 EST
Article-I.D.: enea.1664
Posted: Tue Jan 27 15:41:54 1987
Date-Received: Tue, 3-Feb-87 19:11:39 EST
References: <3173@brl-adm.ARPA>
Reply-To: toma@enea.UUCP (Tony M}nsson)
Organization: ENEA DATA Svenska AB, Sweden
Lines: 34


As far as I can remember, the c/80 from software toolworks can produce
ROMable code. It is a good compiler, and it definitely generates 8080
code.

When trying 8080 C-code in ROM remember to check the implementation of
the IN and OUT instructions. Since the 8080 has to take the port address
as the byte following the IN or OUT opcode, It really cant be done with
"pure" code. The trick is to reserve some space on the stack, write the
opcode and the parameter plus a ret here, and call your own stack. This
is at least how Whitesmits C does it.

Regarding the bug in SCC8080, it is there allright. I spent some futile
evenings with it before I gave up. I wish somebody else would have a go
at it and mail the result. I got this far:

In the file CRET.ASM (what a name for the startup code!) insert a push
of a parameter just before the call to Xarglist. Xarglist is a C routine
that expects a (char *) parameter. The pointer should point to Hex 80,
since this is where CP/M (CP/M 3.0 in my case) leaves the command line
arguments. So far so good, when I run SID, loads the c-program with SIDs
"I" and "R" commands, the compiler happily compiles anything.

Invoking it directly by "SCC8080 hello.c" from the CP/M prompt however,
produces nothing, the compiler simply ignores the parameters.

For further info contact me as toma@enea or

	                   Tony Mansson
				ENEA DATA Svenska AB
				ENEA SYD
				Box 4150
				S-203 12 MALMO
				SWEDEN