[Mailman-Developers] "@" in mail **text** gets replaced inarchives

[Harald Meland]

[John A. Martin]

>>>>>> "baw" == Barry Warsaw
>>>>>> "Re: [Mailman-Developers] "@" in mail **text** gets replaced inarchives"
>>>>>>  Sun, 28 Sep 2003 09:45:32 -0400
>
>     baw> On Sun, 2003-09-28 at 05:13, Harald Meland wrote:
>     >> [Barry Warsaw]
>     >>
>     >> > I really really want to use something like message-ids to
>     >> > generate message file names.  I want to be able to generate
>     >> > links to archived messages in the footers,
>
> In the header, please.  All messages have headers, not all have
> footers.  Footers are optional, no?  This will go on the wire, right?

"Be able to" is not (necessarily) the same as "will always".  I read
Barry as saying "it should be _possible_ to configure Mailman to add
the archive URL of the message to the (Mailman-added) footer of the
message".

That being said, it would be nice if a standardized header could be
used for this.  However, to the best of my knowledge, presently there
is no standard specifying an header for such a purpose.

>     baw> we could rewrite all message-id headers when we accept the
>     baw> message.  That would guarantee uniqueness but it would break
>     baw> the correlation of message-ids between list copies and direct
>     baw> copies.  Is that bad?
>
> Yes.  The RFCs are clear that MTAs must not muck with Message-Id
> other that to create one if there is none.

It is not clear to me that Mailman *is* an MTA.  It is not an SMTP
server, and is not (necessarily) an SMTP client.

However, even if Mailman isn't an MTA, it would be nice if it *mostly*
tries to follow the MTA rules.

(As a side note, I am unable to find *clear* references to the effect
of your statement in RFCs 2821 or 2822.)

> What happened to the notion that archives are supposed to represent a
> true record of what was "on the wire"?

Um.  Mailman lists have numerous configuration options for changing
messages (e.g. adding footers) before they are sent to the list
members, and it has had such options since time immemorial.  As such,
Mailman has to choose; should the messages be archived as they
appeared on the wire _coming in_ or _going out_?  To me, it is obvious
that messages in the archives should reflect, as closely as possible,
the messages members receive.

As to the particular issue of changing a message's message-id iff the
incoming message-id is already present in the list's archives:

 * According to RFC 2822, section 3.6.4, it is up to the host that
   "generates" a message to assure that the message's message-id is
   unique:

     The "Message-ID:" field provides a unique message identifier that
     refers to a particular version of a particular message.  The
     uniqueness of the message identifier is guaranteed by the host that
     generates it (see below).  This message identifier is intended to be
     machine readable and not necessarily meaningful to humans.  A message
     identifier pertains to exactly one instantiation of a particular
     message; subsequent revisions to the message each receive new message
     identifiers.

 * To my mind it would not be obviously wrong to view Mailman as the
   *generator* of messages, at the very least in the cases where it is
   obvious that the previous generator didn't do its job of
   guaranteeing message-id uniqueness properly.

> I also urge that the archive not break signatures on signed mail.

I certainly agree that this would be very nice, both for mail obtained
from archives and mail received "live" through a Mailman list.
-- 
Harald