David,
(sorry for the repeat -- I forgot to 'L' when I replied)
* David Kelly <[EMAIL PROTECTED]> [2002-May-14 12:00 AKDT]:
> On Tue, May 14, 2002 at 10:15:12AM -0800, Christopher Swingley wrote:
> > One of the mailing lists I subscribe to uses the Lyris ListManager
> > software, which attaches '[listname]' to the Subject: line of all
> > the messages send out.
>
> I hate Lyris. Lyris is so nonstandard that one would think it was a
> Microsoft in-house project.
I'm right there with you.
> Actually that is not the problem. One threads on the References: or
> In-Reply-To headers, not on subject. Lyris, at least as configured at
> rennlist.org, mangles the Message-ID header beyond belief. Idiots. And
> removes the In-Reply-To and/or References headers entirely. Is
> impossible to thread. Idiots. Oops, said that already. An example
> mangled Message-ID. This is *not* a message I sent to the list but one I
> received.
Actually, I've figured this one out. I archive the messages in
the list in question (which recently ``upgraded'' from listproc to
Lyris), and needed to figure this out to get my the threading scheme
to work with the database system I built. If you apply the following
Perl regex's:
$line =~ m/^Message-ID:\s+<LYRIS-[^-]+-([^-]+-[0-9.]+-[0-9.]+)/i;
$id_field = $1;
or
$line =~ m/^(References:|In-Reply-To:).*
<LYRIS-[^-]+-([^-]+-[0-9.]+-[0-9.]+)/;
$refs = $2;
(the second regex is all one line) you can extract meaningful date from
the Message-ID: and In-Reply-To: / References: fields. So, in your
example:
> Message-Id: <LYRIS-1067-46546-2002.05.14-14.40.42--dkelly#hiwaay.net@
> rennlist.org>
the useful data is the '46546-2002.05.14-14.40.42' value. A reply
that sets the In-Reply-To or References field will contain this unique
value after Lyris gets through munging it.
> > How might one go about fixing this on my end?
>
> Ask the list owner to dump NT and Lyris for a real OS and real list
> server which respects long honored standards and conventions.
Good one!
> Is very easy to pipe thru awk or perl and remove the addition in the
> Subject header. And while you are at it can add Yahoo! Groups
> advertisement filters if you wish. But if the Lyris list is configured
> as above then only a human could read the messages and guess the proper
> thread order. Unless you know how to completely decode the Message-ID
> and there happens to be threading information contained within.
So does this mean that the decoding I mentioned above would need to
be built into mutt? I don't know how much one-off kludging there is
inside mutt to fix non-standard behavior in other parts of the email
conversation, but this one probably wouldn't be too hard to implement.
If this is something that would be accepted, and if someone could
point me in the right direction, I'd certainly try to write a patch.
Thanks!
Chris
--
Christopher S. Swingley phone: 907-474-2689
Computer Systems Manager email: [EMAIL PROTECTED]
IARC -- Frontier Program GPG and PGP keys at my web page:
University of Alaska Fairbanks www.frontier.iarc.uaf.edu/~cswingle
