Short: Player leaving game vs. call_out()
From: Tmm <tmm@gmx.de>
Date: Tue, 12 Jan 1999 11:04:30 +0100 (MET)
Type: Bug
State: Fixed

> a player enters a room which is loaded thereby.  upon loading, the room
> populates itself with a monster.  this monster starts a call_out() in its
> reset/create/configure/whatever.  before the monster starts this
> call_out, it is made a living by calling the efun enable_commands().  by
> this call it will become this_player for the rest of this call.
>
> so maybe your problem does not come from the new driver at all.  if you
> could provide a minimal example (which should of course not depend on
> your mudlib) which demonstrates the bug you encountered, i would gladly
> try again to reproduce it.

try the following, provided from parsec@uni.mud.de:

void create()
{
    enable_commands() ;   / <- Without this it works!
}


int test()
{
    raum = environment( this_player()) ;

    tell_room( raum, sprintf( "Test: TP n", this_player())) ;
    call_out( "testco", 3) ;
    return 1 ;
}


void testco()
{
    tell_room( raum, sprintf( "Test-CO: TP n", this_player())) ;
}

Clone the object in the room and call test() with a testplayer, then log
him out. The first tell_room will show you, that TP ist correctly set an
the second will show you, that it points to the Testobject, which
shouldn't. As you will see, the 'enable_commands()' causes the problem. We
use enable_commands in all our monsters and players, so we cannot just
leave 'enable_commands' out :)

I hope, you can reproduce this problem.

Greets,
         Tmm@uni.mud.de