Re: feature request from r5rs

[email protected] wrote:
> On Wed, Jun 11, 2003 at 09:27:06AM -0400, Narang wrote:
> 
>>1st caveat I agree with but the other two caveats are also true for
>>
>>(+ (force (delay (* 3 7)) 13)      ===> 34
> 
> 
> Oh right, as for efficiency, if it was supported by the compiler, and 
> introduced as a new uexp, it shouldn't affect performance anywhere else.

How?

A better example to the above is:

(let ((v (delay * 3 4)))
   (+ v 13))

The compiler doesn't know (in the general case) that v is delayed, so it 
cannot omit special code to force it. This decision needs to be made at 
runtime, e.g. by examining whether any of the arguments need forcing.

Of course then the question is *when* you force a value. When it appears 
as an argument *anywhere*? That wouldn't be very useful. When calling 
primitives? Perhaps, although even that might not be quite good enough.


Matthias.




-------------------------------------------------------
This SF.net email is sponsored by:  Etnus, makers of TotalView, The best
thread debugger on the planet. Designed with thread debugging features
you've never dreamed of, try TotalView 6 free at www.etnus.com.