On Fri, Aug 2, 2013 at 11:19 AM, Jonathan S. Shapiro <[email protected]>wrote:

> Not in the simple implementation, because the size of the outgoing value
> is statically unknown
>

I see. The first time around I parsed that as a carry-bit. I see now that
you are implementing arbitrary precision, so the size of carry-out is only
bounded by the input arguments.

I think I now better understand why you are trying to solve this with
regions. However, I don't understand what the static-region-scoping
boundary is. The carry may be used in a successive loop-iteration, which
may itself also produce a carry in the process of using the previous carry.
Is there a form of static scope analysis which will allow this to be
possible while also promptly reclaiming the first carry before the second
is handed in to generate a third?

-- 
"*In science if you know what you are doing you should not be doing it. *
*In engineering if you do not know what you are doing you should not be
doing it. *
*Of course, you seldom, if ever, see either pure state.*" - Richard Hamming
_______________________________________________
bitc-dev mailing list
[email protected]
http://www.coyotos.org/mailman/listinfo/bitc-dev

Reply via email to