[m-dev.] Trailing

[Fergus Henderson]

Peter Schachte, you wrote:
> 
>     2.  It would be nice if there was a mechanism to avoid multiple
> 	untrailings in immediate succession.  Eg, suppose we've been
> 	making one copy of a whole array for each choicepoint.  If a
> 	failure takes us back a few choicepoints, it would be nice if
> 	we could just put back the right copy of the array, rather
> 	than copying back the array once for each choicepoint.  I hope
> 	this is clear.

Yes, it is clear.  But I suspect that any mechanism to do this
will either (a) have "distributed fat", i.e. impose costs on
programs that don't use the feature, or (b) may waste more time
checking whether it can avoid unnecessary untrailings than is saved
by the untrailing it avoids.

-- 
Fergus Henderson <fjh at cs.mu.oz.au>   |  "I have always known that the pursuit
WWW: <http://www.cs.mu.oz.au/~fjh>   |  of excellence is a lethal habit"
PGP: finger fjh at 128.250.37.3         |     -- the last words of T. S. Garp.