Re: minor fix for glib-genmarshal

[Johannes Stezenbach <js convergence de>]

Owen Taylor wrote:
> 
> While this works, it strikes me as ugly -- to do a boolean test on an
> object of unknown type. (You could imagine v_return being a structure
> passed by value, though that won't happen currently.)

Hm, you're right.

On a related note, I quite frequently (e.g. in many finalize
handlers) use code like this:

  if (foo)
    {
      g_object_unref (foo);
      foo = NULL;
    }

IMHO it yould be more elegant to code it like this:

   g_object_unref_x (foo);
   foo = NULL;

where g_object_unref_x (which lacks a good name ;-) would be just a
NULL pointer tolerant version of g_object_unref.

Using g_object_unref_x() in sig->rarg->release would also
solve this genmarshal problem.


Regards,
Johannes