Hack for segmentation faults on Apple Silicon(Macos arm64)

There are bugs in pyobject and gobject-introspection.
The API g_callable_info_prepare_closure() only returns the address
of exec not closure.
On linux arm64 and x86_64 platforms, exec address is same as
closure. However, on Macos arm64, it is not true. Then segmentation
fault happens while calling g_callable_info_free_closure() on exec ptr.

Try the C code:

========================================================================

void puts_binding(ffi_cif *cif, void *ret, void* args[],
                  void *stream)
{
  *(ffi_arg *)ret = fputs(*(char **)args[0], (FILE *)stream);
}

typedef int (*puts_t)(char *);

int main()
{
  ffi_cif cif;
  ffi_type *args[1];
  ffi_closure *closure;

  void *bound_puts;
  int rc;

  /* Allocate closure and bound_puts */
  closure = ffi_closure_alloc(sizeof(ffi_closure), &bound_puts);

  if (closure)
    {
      /* Initialize the argument info vectors */
      args[0] = &ffi_type_pointer;

      /* Initialize the cif */
      if (ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 1,
                       &ffi_type_sint, args) == FFI_OK)
        {
          /* Initialize the closure, setting stream to stdout */
          if (ffi_prep_closure_loc(closure, &cif, puts_binding,
                                   stdout, bound_puts) == FFI_OK)
            {
              rc = ((puts_t)bound_puts)("");
              /* rc now holds the result of the call to fputs */
            }
        }

  printf("closure:%p exec: %p\n", closure, bound_puts);
  /* Deallocate both closure, and bound_puts */
  ffi_closure_free(closure);

  return 0;
}
========================================================================

I tried to fix it but there are some bugs in pyobject which also cause
segmentation faults.

I'm not a pundit in these messy code so just do hacks in pyobject to
fix it.

Link: https://gitlab.gnome.org/GNOME/pygobject/-/issues/455
Link: jeffreywildman/homebrew-virt-manager#166 (comment)
Signed-off-by: Su Yue <l@damenly.su>