[gclist] Java manchines and when to collect.

P. T. Withington ptw@harlequin.com
Mon, 15 Dec 1997 17:16:22 -0500


At 08:57 -0800 12/15/97, Paul Haahr wrote:
>> Yep.  I suspect you need a considerably more powerful type system than
>> Java's though, but I haven't really looked into it.  Are there any
>> languages which can implement their own storage manager?
>
>I think the Boehm, et al, collector is an existence proof for C.  But
>there are fewer such things for type-safe languages.  (The Lisp machines
>might count, but how do you categorize the hardware support?)

Since to write any storage manager you must convert "address to pile of
bits" into "object of type x" (and back), I would say you cannot write any
storage manager in a type-safe language.  More than needing a subset that
does not allocate, you need a loophole in the type system.

I would categorize the hardware support for GC in LispM's the same as the
hardware support for C++ in a PC:  some operations are faster.