If an object self-destructs, it will not immediately terminate execution. If the efun this_object() will be called by the destructed object, the result will be 0.
To keep things consistent, most mudlibs frown upon the destruct()ion of other objects, and instead demand call_others to a specific lfun in the object to destruct (traditionally named "remove"). This will then ensure correct update of e.g. weights, volumes etc.
The interpreter does not really destruct the object immediately, but marks it as deleted, removes it from the list of all objects, and puts it onto a list of to-be-destructed objects. The actual freeing occurs only when all references to a destructed object have gone. Thus it is possible, that an object occupies memory long after it has been destructed, although the object is not visible anywhere anymore from outside.
ob->remove(); if(ob) /* still there, probably ob does not provide remove() */ destruct(ob);This is a way of destructing an object but giving it a chance to do it by itself.