Made all the requested changes.

>
> Tom>  However, I think this is pretty ugly.
> Tom>  It seems like there should be a better way to get this than looking up
> Tom>  a convenience variable.
>
> Sami>  Hmm I looked through the code to find another way but could
> Sami>  not. handle_inferior_event which sets the convenience variable uses
> Sami>  execution_control_state which I don't have access to.
>
> I don't want to hold up this patch for a detail like this, but I wonder
> if we could put the exit code into the struct inferior as well as in the
> convenience variable.
>

Okay, I added exit_code to struct inferior, set it and used it. How does 
that part of the patch look ?

> [create_stop_event_object]
> Tom>  I think it would be better to just have one cast at the end, instead of
> Tom>  lots of casts in the body.
>
> Sami>  Hmm if I change stop_event_object* to event_object it would eliminate
> Sami>  two casts but also add two. One when setting the inferior thread and
> Sami>  one for returning. Same goes for all the create_* functions or should
> Sami>  I change all of those to return more generic objects ?
>
> It seems to me that the callers of the create_* functions then proceed
> to downcast anyway:
>
> Sami>  +  event = (event_object *) create_continue_event_object ();
> Sami>  +  if (event)
> Sami>  +    return evpy_emit_event (event, gdb_py_events.cont);
> Sami>  +  return -1;
>
> Given this, I think just returning the more generic type is ok.
>

Doing this turned my attention to the fact that the individual c types 
were not really needed, and that event_object can be used for all 
events. This resulted in some more clean up. How does it all look now ?