Re: [exim] utf-8 strings in headers

Top Page
Delete this message
Reply to this message
Author: Jeremy Harris
Date:  
To: exim-users
Subject: Re: [exim] utf-8 strings in headers
On 23/02/2020 17:47, Dave Restall - System Administrator, , , via
Exim-users wrote:
> 2020-02-23 16:35:12 1j5uDo-0002yt-Ig H=(NN.NN.NN.NN.example.org) \
>     [NN.NN.NN.NN] F=<User@???> rejected after DATA: \
>     header syntax (unqualified address not permitted: failing \
>     address in "From:" header is: =?utf-8?B?UTFENCODEDSTRING+?=): \
>     unqualified address not permitted: failing address in \
>     "From:" header is: =?utf-8?B?UTFENCODEDSTRING+?=

>
> Excuse the obfuscation.
>
> I've never noticed these before but when I look back in the logs they
> have been happening forever and nobody's complained :-)
>
> When I unpick the actual UTFENCOIDEDSTRING I get what I would call a
> fairly normal address list of the form :-
>
> NAME <user@???>.
>
> These messages are occurring for other headers, e.g. Reply-To.
>
> I had a look at RFC 2822 and NAME <user@???> does seem to be a
> valid value for these headers.


The From: content is supposed to be a human-readable bit and
(separately) a <>-wrapped mail address. Not all munged into
one string and then encoded.

As far as exim cares, it's a string lacking an @ - and therefore
an unqualified mail address (lacking <> too) and no human-readable
bit (legitimate).    Any you haven't specifically permitted
unqualified from that source.
-- 
Cheers,
  Jeremy