On Sat, 2008-06-21 at 07:55 -0700, Mark Crispin wrote: > >>> "If a client issues a CAPABILITY command depite having already received > >>> the capability in the greeting, the server MAY omit the CAPABILITY > >>> response from the LOGIN/AUTHENTICATE commands." > >> This is a very bad idea. It makes assumptions about the client which are not necessarily the case. > > Even if the client didn't ignore them, what's the point of sending the > > same CAPABILITY reply to the client twice? (Assuming of course it > > doesn't change. If it does change, then sure it should be sent.) > > Are you, personally, willing to guarantee that this proposed > additional (and IMHO pointless) complexity will never cause a problem, > now and in the future? > > It IS additional complexity. It is an added behavior rule, with > implications for both server and client. The sole apparent purpose is > to save a few octets of network traffic. The cost is that misbehavior > by either client or server will cause bugs. Well, yes, it is extra complexity.. Maybe it could be said somehow in a bit more simple way that would give the same result for Lemonade clients. But I guess it doesn't really matter, since the environments where those few extra bytes might matter should probably be using Lemonade anyway. But what about the STARTTLS reply having CAPABILITY?
Attachment:
signature.asc
Description: This is a digitally signed message part
_______________________________________________ lemonade mailing list lemonade at ietf.org https://www.ietf.org/mailman/listinfo/lemonade Supplemental Web Site: http://www.standardstrack.com/ietf/lemonade