from the quill of Brad Knowles <brad@his.com> on scroll
<v0300781eae94aaf67329@[205.177.25.174]>
> Well, with sendmail 8.7.y, smap won't work -- you have to choose
> one or the other. I don't recall off the top of my head what the bug
> is, but it keeps it from working with sendmail 8.7.y.
Not at all true. This has been discussed in the fwtk mailing list and the
consensus was that many people (including myself at multiple locations) are
running sendmail 8.7.* and smap/smapd.
> The TIS guys
> don't seem to be particularly interested in keeping the fwtk
> up-to-date, either -- they have "real" work to do on that sort of
> stuff for Guantlet.
First, if they were not updating Gauntlet to make a living, they would not
have the time that they do (however little it is) to work on the toolkit.
Making money is not an evil thing. Second, they are now in beta release
with 2.0 of the toolkit. Sure, it's slow going, and their time on the
toolkit is sparing, but they have responded to the pressure to update the
toolkit. I think ideally they let the toolkit go to somebody with more
time and effort to maintain it, but it's theirs to do with what they want.
> How do you do this while holding the other side open? What if
> they have DNS timeouts? What if you have DNS timeouts? Then you're
> in serious violation of RFC 1123 and just about every other major
> Internet Email RFC in existance.
Yeah. Not being a wide open mail relay for anybody to bounce mail off of
violates an RFC or two (so I understand), but Sendmail 8.8 let's you do it.
I think there are parts of the Internet standards documents which are so
out of date with respect to the Internet maturing from a neat expriment to
a full-blown commerical entity that they are going to be violated.
> Trust me, I've discussed this issue with Ned Freed (author of
> said 14 Internet Email RFCs) and Eric Allman (author of sendmail),
> and there's no easy solution to this problem.
Yeah, I realize that. I think my statement was somewhat of a pipe-dream.
If I were going to implement something like that it would be more directed
at mail clients which are on networks which are serviced by the mail hub.
There can be a certain amount of trust in the infrastructure of those
networks where DNS timeouts and mail server uptime is under a certain
amount of control.
This kind of feature is much like ingress routing. Something a provider
can do to be netfriendly by only distributing mail from it's customers
which has legal return addresses.
b.
--
Brian J. Murrell brian@ilinx.com
InterLinx Support Services, Inc. brian@wimsey.com
North Vancouver, B.C. 604 983 UNIX
Platform and Brand Independent UNIX Support - R3.2 - R4 - BSD
Follow-Ups:
References:
|
|
