Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!cme!durer!hines
From: hines@cme.nist.gov (Lynnwood Hines - pdes)
Newsgroups: comp.lang.lisp
Subject: passing strings from C into Allegro Franz lisp
Message-ID: <HINES.91May29145007@brew.cme.nist.gov>
Date: 29 May 91 18:50:07 GMT
Sender: news@cme.nist.gov
Distribution: comp
Organization: National Institute of Standards and Technology
Lines: 54

OK, I solved the problem here, and the solution is:

C code that creates the string:

***************************************************


#include <stdio.h>
#include <lisp.h>

char back[] = "Hello from C...";
long lisp_value();

char * fill_string (bindex)
int bindex;
{
	printf("C:setting string to: %s\n", back);
	printf("C:bindex set to: %d\n", bindex);
	fflush(stdout);

	strcpy ((char *)lisp_value(bindex)+2, back);
	
	printf("C:setting lisp_value to: %s\n", (char *)lisp_value(bindex));
	fflush(stdout);

	return ((char *)lisp_value(bindex));
}

************************************************************

and the lisp side looks like this:



(defconstant *msl* 256)
(defvar *lsb* (make-string *msl*))
(defvar *lsbi* (ff:register-value *lsb*))
(load "test.o")
(ff:defforeign 'fstring :entry-point (ff:convert-to-lang "fill_string")
                        :arguments '(integer)
                        :return-type :lisp)

(setq foo (fstring *lsbi*))

Turns out the problem was we had to add 2 to the char * before sending
it over due to the length of the string being stored in the first 2
bytes of the string returned by lisp_value.  Our docs had told us to
use a bunch of elaborate macros in a file called lisp.h (which didn't exist), 
but it looks like one of the macros did a char* and the other added 2
to the pointer! 

Just thought this might be of use to others...

						Lynwood Hines