On Sun, 2004-10-24 at 07:24, Nathaniel Gray wrote:
> On Sat, 23 Oct 2004 12:31:39 +0900 (JST), Jacques Garrigue
> <garrigue@kurims.kyoto-u.ac.jp> wrote:
> 
> > Unlucky!
> > With your example, there is no extra boxing involved.
> > I.e. (1,3) and Rpos(1,3) and Apos(1,3) all share the same physical
> > representation (a block with 2 fields).
> > The problem you describe only occurs when your single constructor has
> > only one argument.
> 
> Really??  You're kidding.  That's kind of cool but kind of strange --
> why does it only optimize for composite types?

For two arguments, it's a pointer to a pair on the heap.
For one, its a pointer to a one element tuple on the heap.
This case might be optimised by replacing the pointer
with the value stored in that single field, but it isn't.

That's what Jacques meant -- there's no issue of optimising
the composite case, so there is 'no problem' then,
the problem 'only occurs when .. has only one argument'
because that's the only time you might get rid of the
extra level of indirection. 

Note this would only be possible for the single constructor 
with one argument, because the tag field (for the variant case) 
stored on the  heap is lost too, but it isn't needed 
for the single constructor case. It isn't needed for
the composite argument case either .. but a heap
block is needed for the tuple, so the tag field is there anyhow.

-- 
John Skaller, mailto:skaller@users.sf.net
voice: 061-2-9660-0850, 
snail: PO BOX 401 Glebe NSW 2037 Australia
Checkout the Felix programming language http://felix.sf.net



-------------------
To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners