Em quarta, 03/04/2024 às 07:06 (-03), David Bremner <david@tethera.net> escreveu:

> Message properties might work a bit better than tags

Ah, indeed message properties seem to be more appropriate. Are they
persisted, or are they tied to an Emacs session? Because I would like to
have it so that even if I kill my Emacs session and open a new one, if
the message has the `:scheduled' property with the right time, it should
be sent. There's also concerns on what to do when Emacs was not up when
the scheduled time arrived. Do we 1. send it immediately? or 2. inform
the user it was not able to deliver the message? Going with the latter
has the caveat that it might be missed by the user, and one would be
clueless to the fact that their message hadn't arrived.

> but your periodic search would still have to search for all of the
> scheduled=time properties (wildcard search is currently only supported
> using s-exp queries).

How about having a list of '(MESSAGE-ID . SCHEDULED-TIME), with the
periodic check looking in that list for the scheduled times? It would
have the issue of it not being persistent between Emacs sessions, but I
think that could be addressed by letting the users know this lack of
persistance, and documenting how you could achieve it with
`savehist-additional-variables'.

> I would suggest having a look at the existing draft handling, as what
> you describe sounds like it is related.

Taking a quick glance at it, it seems like we could have a similar
thing, where scheduled messages are saved to the database somewhere,
with either the tag, as it is with drafts, or just the message property
as you suggested.

-- 
João Pedro de A. Paula
IT bachelors at Universidade Federal do Rio Grande do Norte (UFRN)
_______________________________________________
notmuch mailing list -- notmuch@notmuchmail.org
To unsubscribe send an email to notmuch-leave@notmuchmail.org