calling C (MSC 5.1) from BASIC (MS QuickBasic 4.0)

[John C. Lusth]

I now have my C program compiling successfully under the compact and
large memory models of MSC 5.0.  Thanks to K. Heuer and M. Hunter for their
advice.  However, a more challenging adventure lies ahead in my
journey into PC-land.  I need to interface my mini-expert system
with a database program that is being written in Microsoft's
QuickBasic 4.0.  Unfortunately, the integration is to be done many many
miles from me, and I don't yet have access to a BASIC compiler (interpreter?).
But it's my responsibility to tell the integrators how to call my routines
from their code.  So I need some help.

Assuming the C modules are compiled under the large
memory model,  here are some representative C routine prototypes 
and my best guess at the BASIC declarations that are necessary to call them:

    1)  void esinit ()

	DECLARE SUB Esinit CDECL ()


    2)  int esstart ()

	DECLARE SUB Esstart% CDECL ()

These are easy. Now a slightly more complicated call:

    3)  void estrace (int level)

	DECLARE SUB Estrace CDECL (BYVAL LEVEL AS INTEGER)

A call with a string

    4)  int esreadkb (char *filename)

	DECLARE SUB Esreadkb% CDECL (SEG NAME AS STRING)

and one with a string and two integers whose values need to be modified

    5)  void esvalue (char *var, int *value, int *set);

	DECLARE SUB Esvalue CDECL (SEG VAR$, SEG VALUE%, SEG SET%);


Are these declarations correct? The last two make me uneasy,
especially the string passing, since I think BASIC has real strings
while C does not.  Will the do-hickey which rearranges the order of
the arguments on the stack also patch up strings so they will be
compatible?


Instead of all these function calls, I truly would like to use my
top level varargs routine jess () to simplify the integration.
jess () takes an integer code which tells the routine how many and
what type of arguments to expect.  But I have the feeling that it
is impossible to call a varargs function from BASIC.  Am I right in
my thinking?


John C. Lusth
Becton Dickinson Research Center
Research Triangle Park, NC 27278

...!mcnc!bdrc!jcl