Guile, C++, and Mac OS X 10.4 (powerpc)

Guile, C++, and Mac OS X 10.4 (powerpc)

[Mike Solomon]

Hey guile users,
    Trying to compile the simple example bessel.c from
Writing-Guile-Extensions.html (renamed bessel.cc because I'm using g++), I
encountered the following error:

bessel.cc: In function 'void init_bessel()':
bessel.cc:13: error: invalid conversion from 'scm_unused_struct*
(*)(scm_unused_struct*)' to 'scm_unused_struct* (*)(...)'
bessel.cc:13: error:   initializing argument 5 of 'scm_unused_struct*
scm_c_define_gsubr(const char*, int, int, int, scm_unused_struct* (*)(...))'

This even comes up when I put everything in extern "C" { ... }.  I have seen
other postings on the net for other software (lilypond, swig) that speaks of
the same issue, and some suggest that it is a problem with g++ and not
guile.  I'm using powerpc-apple-darwin8-g++-4.0.1 (GCC) 4.0.1 (Apple
Computer, Inc. build 5370)

I know that Lilypond uses a workaround where it defines something called
GUILE_ELLIPSIS, but I am not sure how this logic can be integrated into a
small snippet of code.

Anyway, if any of you know how I can make the bessel example compile on OS X
10.4 PPC using g++ 4.0.1, I'd much appreciate it!

Cheers,
Mike

Re: Guile, C++, and Mac OS X 10.4 (powerpc)

[Hans Aberg]

On 28 Jul 2010, at 21:49, Mike Solomon wrote:

> Hey guile users,
>    Trying to compile the simple example bessel.c from
> Writing-Guile-Extensions.html (renamed bessel.cc because I'm using g+ 
> +), I
> encountered the following error:
>
> bessel.cc: In function 'void init_bessel()':
> bessel.cc:13: error: invalid conversion from 'scm_unused_struct*
> (*)(scm_unused_struct*)' to 'scm_unused_struct* (*)(...)'
> bessel.cc:13: error:   initializing argument 5 of 'scm_unused_struct*
> scm_c_define_gsubr(const char*, int, int, int, scm_unused_struct* (*) 
> (...))'

The SCM type is a pointer to an undefined C type - C hack, which  
clashes with C++. Clever in C, but bad for C++ users.

> This even comes up when I put everything in extern "C" { ... }.  I  
> have seen
> other postings on the net for other software (lilypond, swig) that  
> speaks of
> the same issue, and some suggest that it is a problem with g++ and not
> guile.  I'm using powerpc-apple-darwin8-g++-4.0.1 (GCC) 4.0.1 (Apple
> Computer, Inc. build 5370)

When you use g++, all stuff is compiled as C++. So use gcc and add c++  
libraries when linking.

Then in formally correct C++, main() must be C++. Gcc accepts calling C 
++ from C, but attempting to pass an exception through a C function is  
converted to a termination exception.

So I wrote a C++ function scm::init_guile() calling scm_init_guile()  
and other stuff that needs to be initialized. Then
   int main(int argc, char** argv) {
     init_guile();

     try {
       ...
     }
     ...
   }

Then write a header bessel.h

#ifndef BESSEL_H
#define BESSEL_H

/* Copyright ...
    Free Software Foundation GNU General Public License <http://www.gnu.org/licenses/ 
 >
  */

#ifdef __cplusplus
extern "C" {
#endif

void init_bessel();
...

#ifdef __cplusplus
} // extern "C"
#endif

#endif BESSEL_H


The file bessel.c is compiled as C, but your C++ files include bessel.h.

   Hans

Re: Guile, C++, and Mac OS X 10.4 (powerpc)

[Mike Solomon]

Dear Hans,
    Thank you very much for your response.  I'm including here some stuff I
did (a sort of c++ to c and back again) in hopes that it may be useful to
others trying to do the same thing.  Asterisk delimiters mean a new file,
whose name is in the first comment.

***
//bessel-extension.h

#ifndef BESSEL_EXTENSION_H
#define BESSEL_EXTENSION_H

#ifdef __cplusplus 
extern "C" {       
#endif       

int silly (int x);

#ifdef __cplusplus 
} // extern "C"    
#endif       

#endif BESSEL_EXTENSION_H
***
//bessel-extension.cc

#include "bessel-extension.h"
#include "vector"

int
silly (int x)
{
  // include random c++ functionality
  std::vector<int> *foo = new std::vector<int>;
  foo->push_back (1);
  foo->push_back (2);
  int rv = foo->at (0) +foo->at (1)+x;
  delete foo;
  return rv;
}
***
// bessel.h

#ifndef BESSEL_H
#define BESSEL_H

#ifdef __cplusplus
extern "C" {
#endif


void init_bessel();

#ifdef __cplusplus
} // extern "C"
#endif

#endif BESSEL_H
***
// bessel.c
#include <math.h>
#include <libguile.h>
#include <bessel.h>
#include <bessel-extension.h>

SCM
j0_wrapper(SCM x)
{
return scm_make_real(j0 (scm_num2dbl(x,"j0")));
}

SCM
silly_wrapper(SCM x)
{
return scm_from_int(silly (scm_to_int (x)));
}

void
init_bessel()
{
scm_c_define_gsubr("j0",1,0,0,j0_wrapper);
scm_c_define_gsubr("silly",1,0,0,silly_wrapper);
}
***
// bessel.cc

#include "bessel-extension.h"
#include "bessel.h"

***

Then, compile:

g++ -c bessel-extension.cc
gcc -c -I. -fPIC bessel.c
g++ -bundle bessel-extension.o bessel.o -o libguile-bessel.so -lguile
bessel.cc

(this last line will change based on your platform...ie bundle becomes
shared for linux...use libtool if you're really serious about making
something durable!)

Then:

guile> (load-extension "./libguile-bessel" "init_bessel")
guile> (j0 2)
0.223890779141236
guile> (silly 6)
9

Hope this helps the fledgling C++ programmer using guile!

~Mike

On 7/28/10 11:19 PM, "Hans Aberg" <haberg-1@telia.com> wrote:

> On 28 Jul 2010, at 21:49, Mike Solomon wrote:
> 
>> Hey guile users,
>>    Trying to compile the simple example bessel.c from
>> Writing-Guile-Extensions.html (renamed bessel.cc because I'm using g+
>> +), I
>> encountered the following error:
>> 
>> bessel.cc: In function 'void init_bessel()':
>> bessel.cc:13: error: invalid conversion from 'scm_unused_struct*
>> (*)(scm_unused_struct*)' to 'scm_unused_struct* (*)(...)'
>> bessel.cc:13: error:   initializing argument 5 of 'scm_unused_struct*
>> scm_c_define_gsubr(const char*, int, int, int, scm_unused_struct* (*)
>> (...))'
> 
> The SCM type is a pointer to an undefined C type - C hack, which
> clashes with C++. Clever in C, but bad for C++ users.
> 
>> This even comes up when I put everything in extern "C" { ... }.  I
>> have seen
>> other postings on the net for other software (lilypond, swig) that
>> speaks of
>> the same issue, and some suggest that it is a problem with g++ and not
>> guile.  I'm using powerpc-apple-darwin8-g++-4.0.1 (GCC) 4.0.1 (Apple
>> Computer, Inc. build 5370)
> 
> When you use g++, all stuff is compiled as C++. So use gcc and add c++
> libraries when linking.
> 
> Then in formally correct C++, main() must be C++. Gcc accepts calling C
> ++ from C, but attempting to pass an exception through a C function is
> converted to a termination exception.
> 
> So I wrote a C++ function scm::init_guile() calling scm_init_guile()
> and other stuff that needs to be initialized. Then
>    int main(int argc, char** argv) {
>      init_guile();
> 
>      try {
>        ...
>      }
>      ...
>    }
> 
> Then write a header bessel.h
> 
> #ifndef BESSEL_H
> #define BESSEL_H
> 
> /* Copyright ...
>     Free Software Foundation GNU General Public License
> <http://www.gnu.org/licenses/
>> 
>   */
> 
> #ifdef __cplusplus
> extern "C" {
> #endif
> 
> void init_bessel();
> ...
> 
> #ifdef __cplusplus
> } // extern "C"
> #endif
> 
> #endif BESSEL_H
> 
> 
> The file bessel.c is compiled as C, but your C++ files include bessel.h.
> 
>    Hans
> 
> 
> 
> 

Re: Guile, C++, and Mac OS X 10.4 (powerpc)

[Linas Vepstas]

On 29 July 2010 03:04, Mike Solomon <mikesol@ufl.edu> wrote:
>
> Hope this helps the fledgling C++ programmer using guile!

If anyone is interested, I have a C++ template class that is analogous to
scm_c_define_gsubr  for object methods  (and calls scm_c_define_gsubr
under the covers)  One just says:

   MyTestClass *mtc = new MyTestClass(42);
   define_scheme_primitive("bingo", &MyTestClass::my_func, mtc);

and a corresponding scheme function, with appropriate signature
and argument conversion, gets created.

--linas

Re: Guile, C++, and Mac OS X 10.4 (powerpc)

[Ludovic Courtès]

Hi,

Hans Aberg <haberg-1@telia.com> writes:

> On 28 Jul 2010, at 21:49, Mike Solomon wrote:
>
>> Hey guile users,
>>    Trying to compile the simple example bessel.c from
>> Writing-Guile-Extensions.html (renamed bessel.cc because I'm using
>> g+
>> +), I
>> encountered the following error:
>>
>> bessel.cc: In function 'void init_bessel()':
>> bessel.cc:13: error: invalid conversion from 'scm_unused_struct*
>> (*)(scm_unused_struct*)' to 'scm_unused_struct* (*)(...)'
>> bessel.cc:13: error:   initializing argument 5 of 'scm_unused_struct*
>> scm_c_define_gsubr(const char*, int, int, int, scm_unused_struct*
>> (*)
>> (...))'
>
> The SCM type is a pointer to an undefined C type - C hack, which
> clashes with C++. Clever in C, but bad for C++ users.

I don’t think that this is the cause of the problem.

The problem instead stems from use of function declarators with empty
parenthesis, which is also an obsolescent C feature, as discussed here:

  http://savannah.gnu.org/bugs/?23681

I fixed a few of these in Guile 1.9, but not all of them, and not
‘scm_c_define_gsubr’ in particular.

I would appreciate patches in this area.  :-)

Thanks,
Ludo’.

Re: Guile, C++, and Mac OS X 10.4 (powerpc)

[Hans Aberg]

On 18 Aug 2010, at 15:49, Ludovic Courtès wrote:

>>> Hey guile users,
>>>   Trying to compile the simple example bessel.c from
>>> Writing-Guile-Extensions.html (renamed bessel.cc because I'm using
>>> g+
>>> +), I
>>> encountered the following error:
>>>
>>> bessel.cc: In function 'void init_bessel()':
>>> bessel.cc:13: error: invalid conversion from 'scm_unused_struct*
>>> (*)(scm_unused_struct*)' to 'scm_unused_struct* (*)(...)'
>>> bessel.cc:13: error:   initializing argument 5 of  
>>> 'scm_unused_struct*
>>> scm_c_define_gsubr(const char*, int, int, int, scm_unused_struct*
>>> (*)
>>> (...))'
>>
>> The SCM type is a pointer to an undefined C type - C hack, which
>> clashes with C++. Clever in C, but bad for C++ users.
>
> I don’t think that this is the cause of the problem.
>
> The problem instead stems from use of function declarators with empty
> parenthesis, which is also an obsolescent C feature, as discussed  
> here:
>
>  http://savannah.gnu.org/bugs/?23681
>
> I fixed a few of these in Guile 1.9, but not all of them, and not
> ‘scm_c_define_gsubr’ in particular.
>
> I would appreciate patches in this area.  :-)

I got those when doing a templates in a C++ wrap, and just assumed it  
had something to do with that SCM was a pointer.