Re: Address munging

Osoba podpisana jako Adam Sjøgren <[email protected]>
w artykule <news:[email protected]> pisze:

> Andrzej writes:
>
>> All - it means, that munging may break:
>> - Content-ID, which is more or less irrelevant here.
>> - Message-ID, which may break all Message-ID based features
>> described in standards, including retrieving messages and threading.
>
> That is a good and valid point.

For many users proper use of Message-ID is one of main reasons to choose
NNTP.
Message-IDs may be broken in message bodies and in standard header fields
defined in RFC 5322:
   Message-ID:
   Resent-Message-ID:
   In-Reply-To:
   References:

I haven't found Message-ID and References munged in headers in my local
archive, but (rarely used) Resent-Message-ID and (frequently used)
In-Reply-To are munged.
There are also other fields which may contain Message-ID, like Archived-At
in examples from gmane.ietf.http-wg below:

news://news.gmane.io/[email protected]
/----
Archived-At: <http://www.w3.org/mid/[email protected]>
Resent-Message-Id: <[email protected]>
\----

news://news.gmane.io/[email protected]
/----
Resent-Message-Id: <[email protected]>
In-Reply-To: <CABkgnnVwD_+zzp8eU8KTiYKLKgAUeG9SQchiOSDSMPbow+4pqA@mail.gmail.com>
Archived-At: <http://www.w3.org/mid/[email protected]>
\----

As you can see, munging rules seem to be applied inconsistently even for
a single group.

There may be also problems with email address of the list itself. In some
cases providing unmunged the list address may be the only way to subscribe
the list and send reply. Currently it seems to be the case of all the
yahoogroups as Yahoo removed web access.

-- 
Andrzej P. Woźniak [email protected] (swap z<->h in address)