RE: fixCRLF

From: Murray S. Kucherawy <msk_at_cloudmark.com>
Date: Tue, 30 Nov 2010 10:13:08 -0800

> -----Original Message-----
> From: opendkim-users-bounce_at_lists.opendkim.org [mailto:opendkim-users-bounce_at_lists.opendkim.org] On Behalf Of R.A. Imhoff
> Sent: Tuesday, November 30, 2010 8:58 AM
> To: opendkim-users_at_lists.opendkim.org
> Subject: fixCRLF
>
> From the description of the conf file, it would appear the FixCRLF will
> add a missing CR if the php were to use only "\n", but in fact at the
> receiver end, what passes the test is LF without CR, yet the LF alone
> passes even when FixCRLF is added, so it seems not to do anything, or
> do I misunderstand its purpose?
>
> I would have imagined that the intention of FixCRLF would be to allow
> the DKIM encoding to succeed irrespective of the EOL codes being used
> in a sending script ?

It's difficult to explain what you're seeing without reproducing it locally. What FixCRLF does is tell the library to "repair" any bare CRs or LFs by converting them to CRLFs. The library is invoked by opendkim which is in turn fed data by postfix. Since your PHP script is invoking postfix, it could be that postfix is doing some or all of the fixing for you.

But more important is the fact that FixCRLF currently only applies to handling of the message body, and your experiment appears to be happening only in the header. Maybe it should be extended to cover the header as well, but the results you're seeing must be coming from something postfix is doing.
Received on Tue Nov 30 2010 - 18:13:17 PST