On Saturday 13 September 2008 04:51:04 Vasily Chekalkin wrote: > Another question: there is a lot of PMC_\S+_val accesses to Key pmc. > Should they be replaced with VTABLE analogues as well? Or Key isn't > supposed to be inherited by some other class and it will only sacrifice > performance?
The Key PMC is a confused beast; sometimes it holds an INTVAL, sometimes a STRING, and sometimes it's a PMC. Skip that one for now -- I think we need to do something special there. -- c
