Bernhard Schmalhofer via RT wrote:
>
> I assume that the Hash PMC should also be used by PIR code in
> 'runtime/parrot/library'. In that case returning PMCNULL is not nice, as
> the returned PMC can't even be queried for it's type.
> So for now a None PMC is returned for non-existent keys.
> However this is not consistent with e.q. FixedPMCArray, which does
> return PMCNULL. ResizablePMCArray is even more inconsistent and seems to
> return an Undef PMC.
Yep, we've to unify aggregates behavior.
Thanks, applied.
leo