Re: [PATCH 0/4] Quoting HTML-only emails in replies redux

[Aaron Ecay <aaronecay@gmail.com>]

On Mon, 16 Jan 2012 08:39:30 +0000, David Edmondson <dme@dme.org> wrote:
> On Mon, 16 Jan 2012 02:38:38 -0500, Aaron Ecay <aaronecay@gmail.com> wrote:
> > - Greater flexibility in the construction of address lists.  For example,
> >   there are some email lists where I want replies to list mail to go only
> >   to the list, not also to the original sender.
> 
> Is there a mechanistic way to determine the correct behaviour in this
> respect? I suspect that it's exactly the kind of thing that Carl wanted
> to be included in 'notmuch' itself, so that other UIs can benefit.

I think it requires some amount of configuration, but it can be done
sensibly.  I am much more proficient with Elisp than with C, and
Emacs has prejudiced me towards solutions that allow me to have a
Turing-complete configuration language.  :)

I think a good starting point for thinking about mailing lists is what
Mutt does:
http://www.mutt.org/doc/manual/manual-4.html#using_lists

Notmuch at the CLI/C code level could aim for a comparable level of
expressiveness, and I think it would suffice for most people (including
me).

[...]
> 
> You can do both of these things today using `message-send-hook' (I
> do).

I avoided that, as it seemed to me that just before the message is sent
is too late to be doing these things (I’d like to see confirmation when
writing the message that the address/signature changes were applied
correctly).  But “M-x apropos RET message hook RET” shows that there are
some earlier points to hook into as well.  Thanks.

> I dislike supercite, so no support from me in that direction :-)

One advantage of supercite is that it allows non-English speakers to set
up the “On X, X wrote” line as they prefer.  Notmuch’s current approach
(a hard-coded C string) is the opposite of internationalized.  So it
would be nice to support some customization of that as well, even if we
don’t go the supercite route.

-- 
Aaron Ecay