Actually, probably not where I would go with this no.

Fair enough.

Sure, executing code always incurs a performance penalty, you of all people should know that ;)

Indeed. Hence not adding my (bad) Moose solution to the benchmark.

I've long since accepted that Moose is not targeted at the vast majority of the stuff I do. That its performance is the Perlish trade-off for its facilities.

I truly wasn't trolling. I'd still like a Moose-compatible solution that: a) avoided as many extra function calls (wrappers) as possible (perhaps even source-filtered to in-line pre & post conditions, type-checks and accessors; b) ditched most of the reflection stuff that I personally see little use for. But I know that goes against the very grain of Moose.

Maybe I'll get around to finishing pudu one day :)