Date: Fri, 12 Jun 1998 22:52:50 +0000 (GMT) From: Terry Lambert <tlambert@primenet.com> To: lyndon@ve7tcp.ampr.org (Lyndon Nerenberg) Cc: scrappy@hub.org, freebsd-isp@FreeBSD.ORG, freebsd-current@FreeBSD.ORG Subject: Re: SASL References Message-ID: <199806122252.PAA23412@usr01.primenet.com> In-Reply-To: <199806121846.MAA15378@ve7tcp.ampr.org> from "Lyndon Nerenberg" at Jun 12, 98 12:46:20 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> I don't have the ACAP references handy. The > proposed SMTP SASL interface (AUTH command) is described in the I-D > draft-myers-smtp-auth-*.txt. I like this (a bit), but I hate the use to which it is supposedly going to be put by Qualcomm: http://www.imc.org/draft-gellens-on-demand Which is basically a workaround for ISP brain damage that can only be implemented by non-brain-damaged ISPs. It's a tautology. This a very annoying. The same functionality for dynamic IP addresses polling a server for queued mail is available through a correctly configured split dynamic DNS and the existing ETRN verb. The real pain here is that the ISP will have to correctly configure things in both instances, and the draft suggests a kludge for an authenticated method for turning the connection around ("ATRN") to no real benefit beyond what a competent ISP would provide in the non-kludged case. Either they have a clue or they don't; it's a binary value in this case. 8-( 8-( 8-( 8-( 8-( 8-( 8-( 8-( 8-( 8-( 8-( 8-( 8-( 8-( I would be much happier if they just documented best known practice, instead of complicating the hell out of SMTP for no good reason. If they have to complicate the hell out of SMTP, the least they could do is not use the semantics they've chosen. Why add a new verb, ATRN, instead of using the existing verb, TURN? To get the additional (and optional and ill-thought-out) domain argument(s)? The semantics of TURN can be redefined in the presence of the AUTH to get this behaviour (should someone want to bogify their SMTP server unnecessarily). Everyone denies TURN as it is, anyway, so it's a non-verb. And the brain damage of domain arguments themselves: why require the 450 response if there is no mail queued for the domain(s)? The 450 response dictates implementation, or it dictates a very large amount of work, depending on how queues are stored. It's stupid, and it's going to take as much time anyway, so the server might as well just turn around immediately, and if it has no mail for the domain(s), as a client it can send a "QUIT" verb to the client-cum-server. At the very least, they should redefine the response: "450 shove off you compute intensive PIG!" So that I can say that I'm not going to traverse my whole queue for another hour (insert configurable interval here) if I didn't find anything the last time you bothered me to run my whole queue for no good reason. I sent mail to Randy a long time ago on this, but haven't heard anything back. I was much more polite in that mail. 8-(. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199806122252.PAA23412>