Re: mutt/1116: Fails to thread properly without an @ in msg ID
The following reply was made to PR mutt/1116; it has been noted by GNATS.
From: Christoph Berg <cb@xxxxxxxx>
To: bug-any@xxxxxxxxxxxxx
Cc: Adrian Irving-Beer <wisq-deb@xxxxxxxx>
Subject: Re: mutt/1116: Fails to thread properly without an @ in msg ID
Date: Mon, 5 Mar 2007 11:35:15 +0100
Hi Adrian,
I forwarded your Debian Bug #300327 to the mutt BTS a few days ago
(http://bugs.mutt.org/1116) and there's been some discussion:
Re: Cameron Simpson 2007-03-03 <20070303004550.GA23730@xxxxxxxxxxxxxxx>
> On 02Mar2007 13:05, Thomas Roessler <roessler@xxxxxxxxxxxxxxxxxx> wrote:
> | Part of the problem is that there's little way to extract
> | message-IDs from "In-Reply-To" headers -- except looking for valid
> | syntax.
>
> Well, you could always use a forgiving syntax that allowed "@" to
> be optional for parsing purposes, something like a "<[^>]*>" regexp
> instead of "<[^@>]*@[^@>]*>" (yes, I know the rRFC token stream is more
> complicated than that).
>
> It would let you thread in the face of this particular type of syntax
> bustedness, though of course arbitrary other bustedness may not be handled.
>
> However, mutt should never emit a bad message-id, and so what do you put
> in References: or In-Reply-To: for such a message? It's a slippery slope,
> and I don't like it much.
>
> Maybe we're asking the wrong question.
>
> Christoph, where do these bogus message-ids come from?
Out of curiosity, do you have an example around where these
message-ids are generated?
Christoph
--
cb@xxxxxxxx | http://www.df7cb.de/