Hello,

Alexander Adolf <alexander.adolf@condition-alpha.com> writes:

> [...]
> Browsing through message.el quickly brought me to the sixth of the 21
> FIXMEs in that file, which reads on line 3183:
>
>   ;; FIXME: merge the completion tables from ecomplete/bbdb/...?
>   ;;(add-hook 'completion-at-point-functions #'message-ecomplete-capf nil t)
>   (add-hook 'completion-at-point-functions #'message-completion-function nil t)
> [...]

Thanks to all who cared to share their views, and their suggestions on
my proposed patch. I am attaching an updated version that includes the
improvements that were suggested:

 - remove override, and thus respect user's setting of
   completion-ignore-case (Thomas Fitzsimmons's comment)

 - do not use ":exclusive 'no" in the completion table (Stefan Monnier's
   comment)

Thanks for pointing these out, and for helping to improve the proposed
patch!

Then there was a wider discussion on potential, architectural
improvements of completion handling in message.el.

Before diving into this wider discussion, it would seem worthwhile to me
trying to get to a conclusion on the proposed patch.

Do we have a consensus, that the proposed, updated patch does not solve
all potential issues around email address completion in message mode
once and for all, but that it still provides a useful improvement for
some users without disturbing others, thus making it seem worthwhile to
include the patch?


Many thanks and looking forward to your thoughts,

  --alexander