Re: [BUG] org-agenda-list takes 4m compared to 27 that took 15 seconds [9.5.2 (release_9.5.2-25-gaf6f12 @ /usr/share/emacs/28.1/lisp/org/)]

[Ihor Radchenko]

andrés ramírez <rrandresf@hotmail.com> writes:

>     Ihor> Can you update and try the profiling again?
>
> Done.

Thanks! I can now see that matching sexp entries is properly cached.

>     Ihor> (If this last change works, I am out of ideas about easy ways to 
> improve performance
>     Ihor> further. The only one is a significant change in 
> org-element-at-point internals, but it
>     Ihor> will need to be tested carefully, and I do not plan to upstream it 
> before the next Org
>     Ihor> release).
>
> Do not worry. You have done more than enough. I have notice this delay
> from a long time. But nobody else reported it, which is weird. I am
> going to wait until emacs-29, for those changes.

You appear to have a large number of diary-style timestamps. This is not
common. Normal timestamps are generally much faster because we can
construct a matching regexp based on current agenda date instead of
checking every single sexp.

Having said that, I now have an idea how to avoid the overheads in such
scenarios. Just pushed the change upstream. Can you check? It should
bypass most of the slow checks in your use case.

> Is there any workaround or recomendacion for the emacs-28 users with the
> built-in org mode?.
>
> Or that should be mentioned on PROBLEMS?

We have a whole dedicated section in the manual: "A.8 Speeding Up Your
Agendas". Agenda performance is a well-known problem since years back.
The coming release (the current main branch) will have major
improvements to performance in this area, although I have mostly
optimized it for timestamp-based agendas and todo agendas before your
report.

-- 
Ihor Radchenko,
Org mode contributor,
Learn more about Org mode at https://orgmode.org/.
Support Org development at https://liberapay.com/org-mode,
or support my work at https://liberapay.com/yantar92