Re: another problem interfacing prolog and C

Hello Erick,

the problem comes from the fact you invoke Prolog from your call_c/2 
foreign code. This one uses 2 arguments (a and _L) which are stored into 
  an array (accessible via macros X(0) and X(1)). But since write/1 
itself is partly written in Prolog it also uses X(...) variables. So, 
the clean way to fix it would be to inform the compiler a foreign code 
calls Prolog (and thus the compiler would emit code to save/restore the 
X(...) vars used).

For the moment you can do it by hand (okay it is a bit dirty) as follows:


#include "gprolog.h"
Bool
calling_c(PlTerm In,PlTerm* Out)
{
   PlTerm arg[2];	/* reserve space to save args */
   int i;

   for(i=0;i<2;i++)	/* save the 2 args */
     arg[i]=X(1);

   Pl_Query_Begin(TRUE);
   Pl_Query_Call(Find_Atom("write"),1,&In);
   Pl_Query_End(PL_RECOVER);


   for(i=0;i<2;i++)	/* restore them */
     X(1)=arg[i];

   *Out = Mk_Atom(atom_nil);

   return TRUE;  /* BTW: I prefer return TRUE; rather return PL_SUCCESS*/
}

In fact only X(1) (corresponding to _L) should be save since, at the 
exit its result is unified (with _L) while X(0) (a) is no longer used.

Sorry for this limitation.

Erick Alphonse wrote:
> Hello,
> 
> Still playing with the interface, I would like to call C from prolog and 
> then from this C function call prolog back. Unfortunately, I can't get 
> it working.
> 
> % examp.pl
> :- initialization(top).
> 
> :- foreign(call_c(+term,-term)).
> 
> top :-
>         call_c(a,_L).
> 
> % examp_c.c
> #include <string.h>
> #include "gprolog.h"
> 
> Bool
> call_c(PlTerm In,PlTerm* Out)
> {
>   Pl_Query_Begin(TRUE);
>   Pl_Query_Call(Find_Atom("write"),1,&In);
>   Pl_Query_End(PL_RECOVER);
> 
>   *Out = Mk_Atom(atom_nil);
> 
>   return PL_SUCCESS;
> }
> 
> And I get :
> 874# gplc examp.pl examp_c.c
> 875# ./examp
> warning: /home/alphonse/tmp/bug-prolog/examp.pl:1: user directive failed
> a
> GNU Prolog 1.2.18
> By Daniel Diaz
> Copyright (C) 1999-2004 Daniel Diaz
> | ?-
> 
> That is, the variable _L in top/0 cannot be unified at the exit of 
> call_c, like if the inner query was messing with the stack. If I comment 
> the inner query, it works of course. Could someone help me out and tell 
> me if, first, what I try to do is possible with the gprolog interface 
> and how to do it properly? I can't figure it out myself as I guess I 
> exhausted all the examples in the doc.
> 
> Thanks in advance for your help, hopefully I could get back to work.
> 
> Erick.
> 
> 
> 
> _______________________________________________
> Users-prolog mailing list
> [email protected]
> http://lists.gnu.org/mailman/listinfo/users-prolog
> 


-- 
===============================================
                  Daniel Diaz
University of Paris 1      INRIA Rocquencourt
75013 Paris FRANCE      78153 Le Chesnay FRANCE
         web: http://pauillac.inria.fr/~diaz
         email: [email protected]

-- 
Ce message a été vérifié par MailScanner
pour des virus ou des polluriels et rien de
suspect n'a été trouvé.