[e-lang] Looser memoization and constant-folding: will this work?

[Kevin Reid]

On Jul 25, 2006, at 19:01, Bill Frantz wrote:
> I assumed that preemptive termination consisted of blowing the vat  
> away
> and restarting that vat from its last checkpoint.  That procedure  
> would
> also return th DeepFrozen memoizer to a consistent state.

- The compiler doesn't know anything about vat incarnations.

- Restarting a vat is a heavyweight operation.

- There is probably no general way to tell, in the new incarnation,  
which occurrence of trying-constant-folding is the one that failed  
the last time - the execution path could be arbitrarily different.

-- 
Kevin Reid                            <http://homepage.mac.com/kpreid/>