> But it would be very useful to have the second half of that > available as a separate command in its own right, "ERASE" ... > So Arnt, does it make sense to either propose both commands as > part of this draft and implement one in terms of the other, or > to do an ERASE first and then a separate MOVE referencing ERASE?
As Timo says, you can pipeline the STORE and UID EXPUNGE. Unless you can come up with some reason that an atomic ERASE command would save something significant in the message store -- something I can't imagine to be the case -- there's nowhere near the use for this as there is for MOVE (where it's not just the round trips that are saved, because it allows message-store optimization in some cases). But more to the point: you'll note that the proposed charter is VERY tight on this, and will not be changed... there is exactly ONE thing chartered here, and an ERASE command is not it. This is quite intentional, to head off tangents and feature creep, and to keep us to the thing that's being demanded by, we're told, a significant number of implementors. Please do not try to add anything more to this very tightly focused effort. Barry, Applications AD _______________________________________________ imap5 mailing list [email protected] https://www.ietf.org/mailman/listinfo/imap5
