Calling C function with structure parameter from ADA

Calling C function with structure parameter from ADA

[C.G. Williams]

A friend from China (Dr Kougen Zheng) asked me to post this request...

Help Please!!!
Q: How can an ADA program call a C function with a structure as a parameter?
Any help will be greatly appreciated

Please respond to: cgw1@le.ac.uk
(I will post a summary of any replies in a few days)

Details...

The interpreter used is Adaed-1.11.0a
What I did was... 
$ cc -c sub.c
$ adacomp -nl runlib -i sub.o -a -b struct.ada
$ adaexec runlib
Error: core dumped!!!


******************  struct.ada     *************************************
procedure first is

type Point is record
	x : INTEGER;
	y : INTEGER;
end record;

a : Point;

procedure PRINTA ( pp : in Point);
pragma INTERFACE ( C, PRINTA);
begin
	PRINTA(a);
end;

**************  sub.c           *************************************
#include <stdio.h>

typedef struct {
	int	x;
	int	y;
	} Point;

printa(pp)
Point	pp;
{
	printf("\n coordinate = ( %d , %d )", pp.x, pp.y);
}

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

Ceri Williams                    cgw1@le.ac.uk
PhD Research student
Geology Dept
Leicester University
Leicester
U.K.

Re: Calling C function with structure parameter from ADA

[Theodore Dennison]

C.G. Williams<cgw1@le.ac.uk> 
> A friend from China (Dr Kougen Zheng) asked me to post this request...
> 
> Help Please!!!
> Q: How can an ADA program call a C function with a structure as a parameter?
> Any help will be greatly appreciated
> 
> Please respond to: cgw1@le.ac.uk
> (I will post a summary of any replies in a few days)
> 
..
> typedef struct {
>         int     x;
>         int     y;
>         } Point;
> 
> printa(pp)
> Point   pp;
> 
>         printf("\n coordinate = ( %d , %d )", pp.x, pp.y);
> 

The problem here is that C passes the Point structure to printa by
pushing the Point structure on the stack. C (even ANSI C) does not
define what order the bytes of Point are pushed on the stack! That
means there is no safe way to call printa directly from Ada. In
addition to this, Ada does not define weather it pushes Point on 
the stack, or meerly the Address of Point. However your Ada compiler
docs should either say how it passes parameters, or allow you
to specify.

So...
Make sure that your Ada compiler is passing the ADDRESS of Point (not Point 
itself). Next do either of the following:

2) Change the C printa procedure to read:
      printa (pp)
      Point	*pp;
          printf("\n coordinate = ( %d , %d )", pp->x, pp->y);

3) Create a third C routine:
      printainterface (pp)
      Point	*pp;

      	printa (*pp);
   and change you Ada program to call printainterface instead.


T.E.D. (structured programming bigot)

Re: Calling C function with structure parameter from ADA

[Robert Dewar]

T.E.D. assumes that C passes structures by address. This is wrong, it varies
from one machine environment to another.

The RM essenstially repeats this misconception by requiring Ada to pass
a record by address for Convention C, thus, on some machines ensuring
that Ada is incompatible with C.

Actually this is not a requirement, it is implementation advice.

It is also one of the few cases in which GNAT quite deliberately declines
to follow the implementation advice in the RM.

In GNAT, if you make a procedure convention C, then we pass records
EXACTLY THE SAME WAY THAT C would pass them, guaranteeing compatibility
with C -- This seems a much better approach, and for example, simply
and directly solves the original poster's problems.

I hope other compilers will make this same sensible decision. If not,
watch out, your Ada compiler may go out of its way to ensure that it
is incompatible with C!

Note: in discussion of this point earlier on, the MRT suggested simply
changing the C code to pass the structure by pointer, but that is
irrelevant to the task of creating bindings to existing C interfaces,
where it is essential to be able to model C exactly.

I suppose we could have two conventions

  C
  C_Done_Right

but that seems unnecessary :-)

Re: Calling C function with structure parameter from ADA

[Chris Warack <sys mgr>]

In article <dewar.797190489@gnat>, dewar@cs.nyu.edu (Robert Dewar) writes:
|> T.E.D. assumes that C passes structures by address. This is wrong, it varies
|> from one machine environment to another.
|> 
|> The RM essenstially repeats this misconception by requiring Ada to pass
|> a record by address for Convention C, thus, on some machines ensuring
|> that Ada is incompatible with C.
|> 
|> Actually this is not a requirement, it is implementation advice.
|> 
|> It is also one of the few cases in which GNAT quite deliberately declines
|> to follow the implementation advice in the RM.
|> 
|> In GNAT, if you make a procedure convention C, then we pass records
|> EXACTLY THE SAME WAY THAT C would pass them, guaranteeing compatibility
                            ^^^<-- gcc?
|> with C -- This seems a much better approach, and for example, simply
|> and directly solves the original poster's problems.

Shouldn't the above reference refer to gcc's ability to compile c?  That is,
if I happen to be using a different C compiler to generate the object code
that will be linked, then GNAT makes no such guarantee... or does it?
If so, I'm duly impressed. :-))

|> I hope other compilers will make this same sensible decision. If not,
|> watch out, your Ada compiler may go out of its way to ensure that it
|> is incompatible with C!

Unless they are guaranteed to be using the same conventions in the first place
(by using the same technology perhaps), how is this generally possible?

If it is possible, I am truly interested in how.

-- 
Christopher A. Warack, Capt, USAF
Computer Science Department, US Air Force Academy

cwarack@kirk.usafa.af.mil                (719) 472-2401

Re: Calling C function with structure parameter from ADA

[Robert Dewar]

Christopher, GCC goes out of its way to guarantee data compatibility and
calling sequence compatibility with the native C compilers in a given
configuration. If you ever find a case where there is an incompatibility,
then GCC considers it a bug, and you should report it to the GCC mailing
list as a bug. We know of no such bugs currently in any of the several
hundred configurations supported by GCC.

So When I say that GNAT's calling approach is compatible with C, I really
mean that, I don't just mean it is compatible with GCC. Of course there is no
guarantee that there is not a bug, but bugs can appear anywhere, and it
definitely would be considered to be a bug if GNAT passed a record by
value where C would pass it by address. If you find such a case, report
it as a bug!

Re: Calling C function with structure parameter from ADA

[Fergus Henderson]

kst@thomsoft.com (Keith Thompson) writes:

>In <dewar.797190489@gnat> dewar@cs.nyu.edu (Robert Dewar) writes:
>> T.E.D. assumes that C passes structures by address. This is wrong, it varies
>> from one machine environment to another.
>
>It shouldn't.  The C standard specifies that all parameters, including
>structures, are passed by copy, though of course you can explicitly
>pass an address.

The C standard just specifies that the semantics must be as if they
were passed by copy.  Implementations can and often do pass addresses
but still maintain call-by-value semantics, e.g. by having the caller
make a copy (if necessary) and then passing its address.

-- 
Fergus Henderson            | As practiced by computer science, the study of
fjh@cs.mu.oz.au             | programming is an unholy mixture of mathematics,
http://www.cs.mu.oz.au/~fjh | literary criticism, and folklore. - B. A. Sheil