Perhaps a better idea: should we provide a RETRY restart outside the scope of the with-asdf-cache? Then one could establish a handler to add new systems and invoke the RETRY restart?
Actually, if we were to do that in OPERATE, we would have to flip the order of WITH-ASDF-CACHE and LET, but that shouldn't upset anything.
Perhaps a better idea: should we provide a RETRY restart outside the scope of the with-asdf-cache? Then one could establish a handler to add new systems and invoke the RETRY restart?
Actually, if we were to do that in OPERATE, we would have to flip the order of WITH-ASDF-CACHE and LET, but that shouldn't upset anything.