[tpop3d-discuss] Reply-To: munging

[Paul Makepeace]

On Wed, Jun 06, 2001 at 08:49:25PM +0100, Chris Lightfoot wrote:
> > But it doesn't. It simply changes the reply-to which any halfway decent
> > email client can override (for example, mutt prompts first the reply-to
> > then the from). I.e. this point is scaremongery & FUD at best.
> 
> I disagree. If I set a Reply-To:, I presumably have some
> good reason to do so; so, if the mailing list software
> goes and overwrites it with something else, my intentions
> will be lost.

Right. What I was just saying was the irony that people configure
majordomo to reply-to: <author> and yet if said author attempts to set
the reply-to in the manner you're describing it gets munged anyway. I've
tried this and yup, I was SOL.

> This sucks if you are forced to use a From:
> header which is not an address at which you can frequently
> read mail.

In practice, I wonder what the %age of people this really applies to
these days. A fun scripting challenge for a rainy day... I suspect in
the sub-0.1% range anecdotally derived from never having seen anyone
ever do it in the ten years I've been using email...

> It's rare indeed to find a message without a message-id
> nowadays. (Checks mailspool... OK, out of 6,573 messages I
> checked, a stunning 8 lacked a message-id, of which I
> suspect most were PINE metadata messages, which don't
> count.)

That's true. Fortunately that wasn't the nub of the argument :-)

> The procmail incantation is... well, it's a procmail
> incantation, so one shouldn't expect to understand it. But
> as such things go, it's not too bad:
> 
> # Discard duplicate mails
> :0 Wh: $HOME/.procmail/msgid.lock
> | formail -D 8192 $HOME/.procmail/msgid.cache

Well, that's interesting. I've often wondered about that... Not that I
use procmail (exim filtering seems good enough). Any scheme whose
argument involves "well, you can use procmail to filter it out" would of
course go to /dev/null by any reasonable standard.

Paul, who'll argue either side depending on phase of moon :)

-- 
Use cliches