Re: Results of cache usage study

On 7/26/2012 11:29 PM, Jason Duell wrote:
> It's sort of sad, but another fix that could work here would be to 
> record the new cache limit but not enforce it until the next time the 
> cache get blown away for a crash or other unclean close.  That seems 
> to be happening often enough that it's a viable migration strategy...

That's clever! +1