4 msgForwards vs aliases ?
4 msgno content-filter for authed users
2 msgpostfix does not flush queued mails

Re: [dkim-milter-beta] mlfi_negotiate() problem
\ Tony Earnshaw (15 Nov 2007)

2 msgnewby Q
14 msgPAM 'smtp' file question
5 msgCyrus SASL2 + Postfix + PAM - part 3
9 msgtransport error?
3 msgCan't figure out if this rejection is coming fr...
8 msgAlias with check_recipient_access: /etc/ppostfi...
2 msgNDR reports
7 msgSender-ID and the Hotmail devil...
2 msgPostifx with IMAP
2 msgPostfix and procmail
4 msgdestination limits - what is 'destination'
1 msgRedirect message based on Conditions
1 msgCyrus SASL2 + Postfix + PAM - part 2
6 msgCyrus SASL2 + Postfix + PAM
5 msgheader_checks and content_filter documentation ...
2 msgSome problem configuring postfix/procmail with ...
Subject:Re: [dkim-milter-beta] mlfi_negotiate() problem
Group:Postfix-users
From:Tony Earnshaw
Date:15 Nov 2007


 
Murray S. Kucherawy skrev, on 14-11-2007 23:25:

> I think I've found the problem here with Beta releases and sendmail 8.14.
> It's twofold:
>
> - after Beta3, mlfi_negotiate() began requesting SMFIF_SETSYMLIST from the
> MTA and failing negotiation if it's not offered even though it's by
> definition part of "milter v2", which was the first version to offer
> negotiation in the first place so it should always be there
>
> - prior to sendmail 8.14.1, the MTA would not offer that feature even
> though it was actually available
>
> This contradiction is the issue. So this explains why the upgrade from
> 8.14.0 to 8.14.2 fixed the problem.
>
> However, this doesn't explain the difficulty Tonni was reporting though
> since he's using Postfix. I'll have to wait for some debugging
> information to come back from there before I can make any more progress.

Obviously you didn't have to wait :) Sendmail dkim-milter Beta8, made
available shortly after midnight CET, now runs without fault on Postfix
2.4.6 with custom rpms built on FC6 and RHL5.1, the latter being a
production machine. libmilter is 8.14.1 on the FC6 machine and 8.13.8 on
the RHL5 machine (there's not a hope in hell of Red Hat adopting any
newer version for RHL5, that's not its policy).

I'm sorry I didn't contribute with debug info for Beta7, I was out for a
while until early this morning and didn't see the relevant posts. A diff
-u for Beta7 and Beta9 filter.c shows that you changed a lot of relevant
code.

Thanks for accommodating the Postfix people, now Wietse doesn't have to
change the milter code :)

Best,

--Tonni

--
Tony Earnshaw
Email: tonni at hetnet dot nl


© 2004-2008 readlist.com