On Mon, 28 Oct 1996, Francis Leboutte wrote:

> At 09:45 28/10/96 -0500, Bruce Tobin wrote:
> >Suppose you have the following definitions:
> >
> >(ct:defcstruct mystruct
> >   ((data1 win:word)
> >    (data2 win:word)
> >    (data3 :char 4)))
> >
> >(ct:defun-dll my-c-fun ((pmystruct (mystruct *)))
> >   :call-mode :c                                       
> >   :return-type :long
> >   :library-name mylib
> >   :entry-name "test-func")
> >
> >Assuming that test-func takes one argument, a
> >pointer-to-mystruct that will be filled with the
> >address of a newly created mystruct:
> >
> >1.  Is my-c-fun defined correctly?
> 
> Seems OK.
> 
> >2.  What do I pass to my-c-fun when I call it?
> 
> E.g. the following *mystruct*:
> 
> (defparameter *mystruct* (callocate mystruct data1 <value1> data2 <value2>))
> Francis


Would this be correct even if the dll routine is supposed to take an 
unitialized pointer (to be set by the dll routine itself) as argument?

i.e., if the C code for the dll routine looked like

HRESULT testfunc(mystruct *pmystruct)
{
  pmystruct = (mystruct *)malloc(sizeof(mystruct));
  ....
  ....
  return 0;
}