On Fri, Feb 10, 2017 at 03:57:53PM -0600, David Nicol wrote: > At the risk of appearing trolling, I'd like to reiterate that another > possibility is to abandon reference counting altogether and > instead switch to the boehm gc, just to see how well/poorly that performs. > The immediately obvious downside would be that > some thought would have to go into which of the various reliable > destruction order expectations to support, which might require > keeping some reference counting around anyway, at least for things that can > be proven at compile-time to go out of scope, but > that isn't really dynamic reference counting. "patches welcome" -- print+qq&$}$"$/$s$,$a$d$g$s$@$.$q$,$:$.$q$^$,$@$a$~$;$.$q$m&if+map{m,^\d{0\,},,${$::{$'}}=chr($"+=$&||1)}q&10m22,42}6:17a2~2.3@3;^2dg3q/s"&=~m*\d\*.*gThread Previous | Thread Next