Re: ANNOUNCE: 2a20

On Tue, 4 May 1999, Paul Haldane wrote:

> From: "Paul Haldane" <x_at_y.z>
>
> show the author as "Paul Haldane" (with the quotes). I think it would
> look better without.

I agree. I just put that parser in in a hurry so I haven't made any beauty fixes yet. I think that the comment part can even start with white spaces at times, which should be prevented too.

> Bigger thing - threading. With the example mailbox tacked on to the end
> of this message, hypermail doesn't recognise any of the messages as being
> part of a thread (I guess because there are no in-reply-to/references
> records and hashreplylookup() no longer uses the date/subject fields to
> construct threads).

Right. I haven't added that check to the new thread functions (yet). Anyone with some time would be very welcome to help out (there too).

> I think hypermail should do its best to include all messages in a thread
> even if the user's MUA hasn't bothered to include an in-reply-to. I'd
> have thought that matching on subjects should be good enough for this.

Yes. Many MUAs are silly enough to not do that.

> I personally don't think that hypermail should try to be too clever here
> - if a message does have a valid in-reply-to then we should believe it
> (even if the subject line is completely different).

Of course, many users do change subject in the middle of a thread.

> Second problem - the second and third messages both get 'maybe reply'
> entries pointing at themselves.

The code for the thread links for each mail has not been updated to use the new thread code. It should.

> I think what needs to happen here is for hashreplynumlookup() to be
> passed the number of the current message so that it can reject itself as
> a match (may also apply to hashreplylookup()).

Sounds reasonable, yes.

-- 
             Daniel Stenberg - http://www.fts.frontec.se/~dast
   ech`echo xiun|tr nu oc|sed 'sx\([sx]\)\([xoi]\)xo un\2\1 is xg'`ol