On Thu, Sep 11, 2003 at 04:24:01PM +0200, Ingolf Steinbach wrote:
> Richard Rauch wrote:
> > So, if I understand the thread correctly, I can register this callback:
> > 
> >   extern "C" static void callback_fun(int button, int state, int x, int
> >   y);
> > 
> > from my C++ class:
> > 
> >   glutMouseFunc (callback_fun);
> > 
> > ...and have callback_fun() call my class?
> 
> Yes, this should work (provided callback_fun matches the
> signature expected by glutMouseFunc()). The only drawbacks
> are

Just tried.  I get the same visible behavior: Most of it works, but
the scroll-wheel crashes.

Also, my compiler complains about multiple storage classes in
the declaration if I have both {extern "C"} and {static}.
(This is g++ 3.3.)  So, I had to remove the {static}.  Now I
have name-space pollution.  (sigh)


>  - one additional level of function calls

I can live with that.  GLUT is a graphic library, and the callbacks are
for handling X events (translated into GLUT events...), so this is a
drop in the bucket I think.  (^&


>  - you'd have to make the static member function public to
>    access it from callback_fun()

That's not as bad as the name-space pollution I mention above.


I consider the uglification of the source code to be the biggest drawback.
Where before I had one glut*() function to register the callback (plus
the "real" callback code), I now have:

  register_*(void) in the root to register the callback.
  extern "C" void ... (...) for the real callback.
  register_*() INVOCATION in the specific classes.
  the original callback code in the class
  the original glut*() call modified to register the C callback

...and it still crashes the same way.


(And, yes, I converted every callback, not just the mouse-button
callback that seemed to directly crash me.)

I guess that it's not a total waste if ISO C++ really requires the
interface to be done this way.

The upshot seems to be that this has had no effect on the bug.



> Good luck

Thanks.  It looks like I need it.  (^&


> -- 
> 
> Ingolf Steinbach                       Jena-Optronik GmbH
> ingolf.steinbach@jena-optronik.de       ++49 3641 200-147
> PGP: 0x7B3B5661  213C 828E 0C92 16B5  05D0 4D5B A324 EC04
> 

-- 
  "I probably don't know what I'm talking about."  http://www.olib.org/~rkr/