New Org (from main) and slow clocking

New Org (from main) and slow clocking

[Marcin Borkowski]

Hi all,

I have just updated my Org (from the main branch, Org mode version
9.6.1, release_9.6.1-253-g74eb0f), and I noticed an issue with clocking.
I have an "ongoing" task which already accumulated 7000+ clock entries,
and clocking in takes close to 90 seconds.  It was instantaneous before
(version 9.6-pre, release_9.5.5-853-g7b9d8e).  What can I do now?

And how can I downgrade?  (I tried the "logical" thing, i.e., I checked
out a "tmp" branch at tag release_9.6.1 and ran "make update", but it
didn't work because it (obviously) couldn't pull the "tmp" branch.

Alternatively, since Org mode is much more mature now than a few years
ago, how can I get rid of my Git-based installation in favor of the Elpa
package (which, I assume, should be much more stable/bug-free)?  (I
understand that by using bleeding edge Org mode I risk a bit, and maybe
it's time to trade new/shiny for safe...)

TIA,

-- 
Marcin Borkowski
http://mbork.pl

Re: New Org (from main) and slow clocking

[Marcin Borkowski]

On 2023-02-17, at 07:51, Marcin Borkowski <mbork@mbork.pl> wrote:

> Hi all,
>
> I have just updated my Org (from the main branch, Org mode version
> 9.6.1, release_9.6.1-253-g74eb0f), and I noticed an issue with clocking.
> I have an "ongoing" task which already accumulated 7000+ clock entries,
> and clocking in takes close to 90 seconds.  It was instantaneous before
> (version 9.6-pre, release_9.5.5-853-g7b9d8e).  What can I do now?
>
> And how can I downgrade?  (I tried the "logical" thing, i.e., I checked
> out a "tmp" branch at tag release_9.6.1 and ran "make update", but it
> didn't work because it (obviously) couldn't pull the "tmp" branch.

Ok, so I think I managed to downgrade (make cleanall, make, sudo make
install - is that correct?), but the problem persists in 9.6.1...

> Alternatively, since Org mode is much more mature now than a few years
> ago, how can I get rid of my Git-based installation in favor of the Elpa
> package (which, I assume, should be much more stable/bug-free)?  (I
> understand that by using bleeding edge Org mode I risk a bit, and maybe
> it's time to trade new/shiny for safe...)

Best,

-- 
Marcin Borkowski
http://mbork.pl

Re: New Org (from main) and slow clocking

[Ihor Radchenko]

Marcin Borkowski <mbork@mbork.pl> writes:

> I have just updated my Org (from the main branch, Org mode version
> 9.6.1, release_9.6.1-253-g74eb0f), and I noticed an issue with clocking.
> I have an "ongoing" task which already accumulated 7000+ clock entries,
> and clocking in takes close to 90 seconds.  It was instantaneous before
> (version 9.6-pre, release_9.5.5-853-g7b9d8e).  What can I do now?

Do you have org-element-use-cache set to nil?

-- 
Ihor Radchenko // yantar92,
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>

Re: New Org (from main) and slow clocking

[Marcin Borkowski]

On 2023-02-17, at 13:02, Ihor Radchenko <yantar92@posteo.net> wrote:

> Marcin Borkowski <mbork@mbork.pl> writes:
>
>> I have just updated my Org (from the main branch, Org mode version
>> 9.6.1, release_9.6.1-253-g74eb0f), and I noticed an issue with clocking.
>> I have an "ongoing" task which already accumulated 7000+ clock entries,
>> and clocking in takes close to 90 seconds.  It was instantaneous before
>> (version 9.6-pre, release_9.5.5-853-g7b9d8e).  What can I do now?
>
> Do you have org-element-use-cache set to nil?

No, should I?

-- 
Marcin Borkowski
http://mbork.pl

Re: New Org (from main) and slow clocking

[Ihor Radchenko]

Marcin Borkowski <mbork@mbork.pl> writes:

>> Do you have org-element-use-cache set to nil?
>
> No, should I?

You should not. But, when it is nil (not default), things will be slow
in your scenario.

Another thing you can try is setting org-element--cache-self-verify to
nil.

If it does not help either, please do M-x profiler-start <RET> cpu <RET>
[clock in] M-x profiler-report <RET> M-x profiler-report-write-profile
<RET> cpu-profile.el <RET>
and then share the resulting profile data.

-- 
Ihor Radchenko // yantar92,
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>

Re: New Org (from main) and slow clocking

[Marcin Borkowski]

On 2023-02-17, at 14:32, Ihor Radchenko <yantar92@posteo.net> wrote:

> Marcin Borkowski <mbork@mbork.pl> writes:
>
>>> Do you have org-element-use-cache set to nil?
>>
>> No, should I?
>
> You should not. But, when it is nil (not default), things will be slow
> in your scenario.
>
> Another thing you can try is setting org-element--cache-self-verify to
> nil.

That did help indeed.  Should I make it nil in my init.el?

Thanks,

-- 
Marcin Borkowski
http://mbork.pl

Re: New Org (from main) and slow clocking

[Ihor Radchenko]

Marcin Borkowski <mbork@mbork.pl> writes:

>> Another thing you can try is setting org-element--cache-self-verify to
>> nil.
>
> That did help indeed.  Should I make it nil in my init.el?

You can, as long as you do not see warnings from Org about
org-element-cache. This variable is enabling extra consistency checks in
org-element-cache.

The plan is setting this to nil in one of the next releases.

Alternatively, you can reduce the value of
org-element--cache-self-verify-frequency. By default, 5% of all the
cache queries are double-checked, which is expensive in your use case.

-- 
Ihor Radchenko // yantar92,
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>

Re: New Org (from main) and slow clocking

[Marcin Borkowski]

On 2023-02-17, at 19:14, Ihor Radchenko <yantar92@posteo.net> wrote:

> Marcin Borkowski <mbork@mbork.pl> writes:
>
>>> Another thing you can try is setting org-element--cache-self-verify to
>>> nil.
>>
>> That did help indeed.  Should I make it nil in my init.el?
>
> You can, as long as you do not see warnings from Org about
> org-element-cache. This variable is enabling extra consistency checks in
> org-element-cache.
>
> The plan is setting this to nil in one of the next releases.
>
> Alternatively, you can reduce the value of
> org-element--cache-self-verify-frequency. By default, 5% of all the
> cache queries are double-checked, which is expensive in your use case.

Thanks!  I added (setq org-element--cache-self-verify nil) to my
init.el.

-- 
Marcin Borkowski
http://mbork.pl