For years my workflow has centered in org-agenda and I would go from one clocked item to the next. For instance, I would be clocked into my "Emails" task, which never closes, and then eventually move down in the agenda to "Task B". Then I hit C-x TAB to clock in. It correctly queries for a comment on the task I'm leaving but no longer clocks me in to the new task as I'd asked. Is this a bug or am I missing a new setting? Org mode version 9.4.6 (9.4.6-gf70e36 Thanks! - Tory
torys.anderson@gmail.com (Tory S. Anderson) writes:
> For years my workflow has centered in org-agenda and I would go from
> one clocked item to the next. For instance, I would be clocked into
> my "Emails" task, which never closes, and then eventually move down in
> the agenda to "Task B". Then I hit C-x TAB to clock in. It correctly
> queries for a comment on the task I'm leaving but no longer clocks me
> in to the new task as I'd asked. Is this a bug or am I missing a new
> setting?
This should work AFAICT and it does for me. (With key {I} or {C-c C-x
TAB} in the agenda opposed to your setting with {C-x TAB}.)
Could you provide a complete mini example? Possibly this helps to see
more clearly.
Best regards,
--
Marco
My minimal loadup doesn't exhibit the problem. So, tracing the function "org-agenda-clock-in" that my =C-x TAB= is bound to, I see this.
On the broken one I see the following:
#+begin_src lisp
=====================================================================1 -> (org-agenda-clock-in nil)
1 <- org-agenda-clock-in: !non-local\ exit!
#+end_src
Whereas on the clean working one I see:
#+begin_src lisp
=====================================================================1 -> (org-agenda-clock-in nil)
1 <- org-agenda-clock-in: 0
#+end_src
But find-grep doesn't find anything in my dependencies that matches "non-local". This looks likely to be related to the original problem. Any idea how I can find the source of this !non-local? This is my first time using emacs function tracing.
Marco Wahl <marcowahlsoft@gmail.com> writes:
> torys.anderson@gmail.com (Tory S. Anderson) writes:
>
>> For years my workflow has centered in org-agenda and I would go from
>> one clocked item to the next. For instance, I would be clocked into
>> my "Emails" task, which never closes, and then eventually move down in
>> the agenda to "Task B". Then I hit C-x TAB to clock in. It correctly
>> queries for a comment on the task I'm leaving but no longer clocks me
>> in to the new task as I'd asked. Is this a bug or am I missing a new
>> setting?
>
> This should work AFAICT and it does for me. (With key {I} or {C-c C-x
> TAB} in the agenda opposed to your setting with {C-x TAB}.)
>
> Could you provide a complete mini example? Possibly this helps to see
> more clearly.
>
>
> Best regards,
>>>>> On Thu, 24 Jun 2021 12:42:37 -0600, "Tory S. Anderson" <torys.anderson@gmail.com> said:
Tory> My minimal loadup doesn't exhibit the problem. So, tracing the function "org-agenda-clock-in" that my =C-x TAB= is bound to, I see this.
Tory> On the broken one I see the following:
Tory> #+begin_src lisp
Tory> =====================================================================1 -> (org-agenda-clock-in nil)
Tory> 1 <- org-agenda-clock-in: !non-local\ exit!
Tory> #+end_src
Tory> Whereas on the clean working one I see:
Tory> #+begin_src lisp
Tory> =====================================================================1 -> (org-agenda-clock-in nil)
Tory> 1 <- org-agenda-clock-in: 0
Tory> #+end_src
Tory> But find-grep doesn't find anything in my dependencies that matches
Tory> "non-local". This looks likely to be related to the original
Tory> problem. Any idea how I can find the source of this !non-local? This
Tory> is my first time using emacs function tracing.
'non-local exit' is the tracing telling you that one of the functions
it called returned unexpectedly, rather than returning a result. You
might get a better result by doing
(setq debug-on-error t)
instead of tracing. If org-agenda-clock-in then signals an error
you'll get a backtrace.
Robert
--
After some exploration I found that it appears to be the result of the setting org-log-note-clock-out . It I switch this off, switching tasks works as expected. It used to be that having it on didn't cause the bug of failing to switch to a new task immediately. Can anyone else please verify that having the note pop up with (org-log-note-clock-out t) causes it not to successfully clock in to the new item when previously clocked in to a different item, when switching via the org agenda? It looks that way to me.
- Tory
Marco Wahl <marcowahlsoft@gmail.com> writes:
> torys.anderson@gmail.com (Tory S. Anderson) writes:
>
>> For years my workflow has centered in org-agenda and I would go from
>> one clocked item to the next. For instance, I would be clocked into
>> my "Emails" task, which never closes, and then eventually move down in
>> the agenda to "Task B". Then I hit C-x TAB to clock in. It correctly
>> queries for a comment on the task I'm leaving but no longer clocks me
>> in to the new task as I'd asked. Is this a bug or am I missing a new
>> setting?
>
> This should work AFAICT and it does for me. (With key {I} or {C-c C-x
> TAB} in the agenda opposed to your setting with {C-x TAB}.)
>
> Could you provide a complete mini example? Possibly this helps to see
> more clearly.
>
>
> Best regards,
seems related to a bug I reported a few days ago. link to my bug report: https://lists.gnu.org/archive/html/emacs-orgmode/2021-06/msg00506.html the bug is basically that we you have org-log-note-clock-out set to t, an error "before first headline" would occur when you clock out, thus blocking subsequent actions.
this commit https://code.orgmode.org/bzg/org-mode/commit/8e3e2f667f0b28b85845204b708c3f0aebc9152b probably fixes the issue. Could you perhaps give it a test?
Tory, Dave D <wenlong.dai@gmail.com> write: >...> issues with org-log-note-clock-out... > this commit > https://code.orgmode.org/bzg/org-mode/commit/8e3e2f667f0b28b85845204b708c3f0aebc9152b > probably fixes the issue. Could you perhaps give it a test? Yes, please! I also guess that the commit by Nicolas is a fix for the issue. Thanks Nicolas, Dave and Tory. Ciao, -- Marco