Re: Rejected messages from the mailing list

From: Rolf E. Sonneveld <>
Date: Mon, 02 Aug 2010 23:46:50 +0200

On 08/02/2010 11:00 PM, Murray S. Kucherawy wrote:
> Hi Rolf,
>> -----Original Message-----
>> From: [mailto:opendkim-users-
>>] On Behalf Of Rolf E. Sonneveld
>> Sent: Monday, August 02, 2010 1:53 PM
>> To: Murray S. Kucherawy
>> Cc: Alessandro Vesely;
>> Subject: Re: Rejected messages from the mailing list
>> So +1 for a new canonicalization scheme.
> Do you (or does anyone) want to take a run at specifying such a thing?
I will try to make a start. My approach would be to make an inventory of
the commonly used MIME header fields and parameters
(RFC2045-2049+updates) and of the headers, most commonly used by DKIM to
generate the signature (the SHOULD header fields of par. 5.5 of
RFC4871). Would that make sense? It's impossible to make a complete list
of all headers that ever can be used to sign, but 4871 uses a 'SHOULD
NOT' for the other header fields, not mentioned in 5.5.

> "relaxed" header canonicalization already lowercase-izes names, so content-type is the same as Content-Type is the same as CoNtEnT-tYpE. It's the value of the field that would need some attention.
> What are some common rewrites? Is the quoting thing Courier does a common one, for example?

Common rewrites are the ones, where the input (slightly) violate the
standards and the MSA/MTA will try to canonicalize the fields, to send
RFC-compliant messages out. The 'be liberal in what you accept - be
strict in what you emit' Internet paradigm. And it is my experience that
common rewrites occur more often than we would like it to be.

Is there any way we can gather statistics on this specific item?

Received on Mon Aug 02 2010 - 21:47:05 PST

This archive was generated by hypermail 2.3.0 : Mon Oct 29 2012 - 23:19:48 PST