"Dan Sugalski" <[EMAIL PROTECTED]> wrote in message news:a05200f00ba1ebb73c6d2@[63.120.19.221]... > There'll definitely be memory address reuse. If .id returns the > current object's memory address, it shouldn't be cached any place, as > otherwise you'll find things going bang with some regularity.
In a multi-threaded environment, even a simple $a.id == $b.id may have significant time between the two lookups. If we are going to rely on addresses for id comparison, then we need to mandate that the address is constant for the life of the object. Brent indicated that he could think of no reason for the "header address" to change: but that isn't strong enough. We need a cast- iron guarantee. Otherwise we can't compare identity using .id. Dave.