Re: [exim] header_syntax

Startseite
Nachricht löschen
Nachricht beantworten
Autor: Fred Viles
Datum:  
To: exim-users
Betreff: Re: [exim] header_syntax
On 3 Mar 2005 at 12:28, Brian wrote about
    "[exim] header_syntax":


| I had recently sent a message to the qmail list and one of the
| moderators/list people tried to send this to me, but it was rejected.
|
| So he told me my mta was 'retarded'.


Seems like a strong opinion from a person sending a malformed message
with no ability to get a DSN if it wasn't delivered.

| I would assume that he being a massive qmail advocate might not be
| wrong, but I don't see what I could have done differently.


Well, you could have not applied the header_syntax test.

| Below is the part from the rejectlog
| 
| 2005-03-03 10:54:58 1D6sfC-000Aln-Aw H=(discworld.dyndns.org) 
| [70.64.60.7] F=<> rejected after DATA: RFC2822: missing or malformed 
| local part: failing address in "From" header is: "Charles Cazabon's 
| correspondence secretary" <>
| Envelope-from: <>
| Envelope-to: <bcook-dated-1110038080.61f2f1@???>
| P Received: from [70.64.60.7] (helo=discworld.dyndns.org) by 
| c.mx.poklib.org with smtp  (Exim 4.44; FreeBSD) id 1D6sfC-000Aln-Aw for 
| bcook-dated-1110038080.61f2f1@??? ; Thu, 03 Mar 2005 10:54:58 -0500
| P Received: (qmail 29630 invoked by uid 500); 3 Mar 2005 15:54:32 -0000
|    Date: 3 Mar 2005 15:54:32 -0000
| I Message-ID: <20050303155432.29629.qmail@???>
| T To: <bcook-dated-1110038080.61f2f1@???>
| F From: "Charles Cazabon's correspondence secretary" <>
| 
| I was looking in rfc822 and 2822 and I couldn't find where it says 
| something relating to " <> names can't have real names in quotes" or 
| something like that.


The quoted string is not the issue. <> is not a valid address
syntax.

| did exim do this erroneously?


No.

| Is the 'qmail guy' right?


No. Your MTA is only doing what you asked it to, which is something
it does not do by default. So the qmail guy is unequivocally wrong
to call the *MTA* retarded.

The message *is* malformed according to RFC2822, and you have made a
policy decision to reject such malformed messages. Apparently the
qmail guy thinks your policy decision is "retarded". He's entitled
to his opinion, but ISTM he doesn't have any data to back it up.

The important data are: how much unwanted mail is the header_syntax
check blocking that wouldn't otherwise be blocked, and what is the
false positive rate. Only you can decide if the cost/benefit ratio
makes sense for you.

- Fred