On Wed, Jun 12, 2002 at 08:41:22PM +0200, Alain Bench <[EMAIL PROTECTED]> wrote:
> Mutt 1.4 has wonderfull code for threading: powerfull, versatile,
> fast, configurable to everyone's taste, informative (I mean the "&?"
> missing and the "*" broken). But a little bit fragile: Perhaps too much
> confidence in IRT's content.
If you can think of a better heuristic than the current one to weed out
things that aren't message-ids, please do suggest it. We really don't
have much choice but to blindly trust the headers. The code will never
crash or go into an infinite loop, which was not true of the old
threading code. And the new threading code is much more powerful
because of the stuff it does with missing messages: it manages to keep
threads together that otherwise would have been broken up. But in the
very rare cases where several messages appear to have the same ancestor,
and that ancestor is actually not a message-id, and the user part is
longer than 8 characters, so it gets by mutt's non-message-id detection
heuristic, things do get threaded together that shouldn't (or at least
might, depending on what other ancestors there are. IMHO, this is a
very small price to pay for the added power of the new threading
algorithm, especially because from now on, RFC2822 compliant mailers
will never put such things in in-reply-to anyway.
-Daniel
--
Daniel E. Eisenbud
[EMAIL PROTECTED]
"We should go forth on the shortest walk perchance, in the spirit of
undying adventure, never to return,--prepared to send back our embalmed
hearts only as relics to our desolate kingdoms."
--Henry David Thoreau, "Walking"
