RE: dkim-rep works (was 2.4.0.Beta3 available )

From: Murray S. Kucherawy <msk_at_cloudmark.com>
Date: Fri, 20 May 2011 10:40:15 -0700

> -----Original Message-----
> From: opendkim-dev-bounce_at_lists.opendkim.org [mailto:opendkim-dev-bounce_at_lists.opendkim.org] On Behalf Of Andreas Schulze
> Sent: Friday, May 20, 2011 6:33 AM
> To: Andreas Schulze
> Cc: Murray S. Kucherawy; opendkim-dev_at_lists.opendkim.org
> Subject: Re: dkim-rep works (was 2.4.0.Beta3 available )
>
> some other open points to me:
> - we diskussed a stronger default for ReputationReject some times ago.
> The current value of 1001 is safe but very defensive: It will *never* reject
> any mail because 1000 is the larges possible value in the wild ...
> This is dokumented in the manpage.
>
> I personaly configure my systems with "ReputationReject 30" and dont miss any mail.
> Maybe I could do a little bit statistics ...

My preference is to use a default that doesn't interfere with delivery at all.

That said, if dkim-reputation.org wants to post a recommended default value, I'd be fine with using that.

> - The lasts months I run all dkim-signer with "BodyLengths yes". This option does not
> exist anymore. Without this setting the whole body is signed an any ML appending a footer
> will destroy my signatures. Right ?
> What is the suggested way / recomendation ??

The general chatter about use of "l=" reached the point where the working group considered removing it completely. Based on the technical arguments presented, I decided to force people to be more deliberate in how it's used.

If you want the previous functionality of BodyLengths, use:

BodyLengthDB csl:*

This means you're explicitly saying you always want it used, rather than the previous fire-and-forget Boolean setting that can result in you not realizing you're using it when you might not really want to.

> - I build opendkim always with a little patch to tell me if a default configfile will be used.
> I find it usefull ...
> I think, I should open a feature request !?

Possibly; can you post an example of the output you're looking for?

-MSK
Received on Fri May 20 2011 - 17:40:24 PST