On Wed, Jan 26 2022, Carl Worth wrote:

> On Wed, Jan 26 2022, inwit wrote:
>>> Even that will not be perfect (since the messages matching the query
>>> could change behind emacs back), but close enough for most interactive
>>> use, maybe?
>>
>> I was thinking about saving the IDs of the messages affected by the 
>> change, but I still don't know how would I go about that or if it's 
>> sensible.
>
> One could imagine a history that would enable a conventional undo stack
> for a notmuch interface. The trick with making that usable would be the
> need to refresh views to make what was undone evident, (and the fact
> that some of the operations could be large/slow).
>
> All of those issues kept me from pursuing the idea in early days of
> coding up the emacs notmuch UI. But someone could certainly explore the
> implementation further if desired.

maybe this could be a buffer-local history variable, for notmuch search
and tree search buffers, and the undo feature apply only to the current
search buffer.  the view update is then well-defined, i think, if it's
just changing tags, at least, and per-buffer operation is the normal
behaviour of undo/redo in emacs.  hooking that up with the regular emacs
undo/redo mechanism would be great (in my book at least), but i've never
looked at that, so not sure if it's possible.

cheers,
jao
-- 
Keep away from people who try to belittle your ambitions. Small people
always do that, but the really great make you feel that you, too, can
become great. - Mark Twain
_______________________________________________
notmuch mailing list -- notmuch@notmuchmail.org
To unsubscribe send an email to notmuch-leave@notmuchmail.org