org-clock-rounding-minutes -- non-zero value can cause double clocking

org-clock-rounding-minutes -- non-zero value can cause double clocking

[Raymond Zeitler]

[-- Attachment #1: Type: text/plain, Size: 1400 bytes --]

There is a problem when org-clock-rounding-minutes is non-zero, say N.  The problem is that the clockin-time of a task can be N less than the clockout-time of the previous task at certain times.  Thus, clock reports can show an extra N minutes total time for each occurrence of this effect.
Consider this.
Set org-clock-rounding-minutes to 6. (This allows for tracking time in 0.1 hour increments.)
Create an org file with two tasks:
* Tasks** TODO Foo** TODO Bar

Clock into Task Foo at, say, 12:54 plus or minus 2 minutes.  Its LOGBOOK will show the expected 12:54 timestamp.
Clock into Bar at 12:57, which is halfway between 0.90th hour and 1.00th hour.
Result, the clockout time for Foo is rounded up.  But the clockin time for Bar is rounded *down*.  Thus, two tasks have started at the same time.
* Tasks** TODO Foo   :LOGBOOK:   CLOCK: [2019-03-08 Fri 12:54]--[2019-03-08 Fri 13:00] =>  0:06   :END:** TODO Bar   :LOGBOOK:   CLOCK: [2019-03-08 Fri 12:54]   :END:
So when I use org-mode to track time for my weekly timesheet, it will report that my total clocked time for any given day is several minutes more than the time I've been at work!
I expect that the clockin time of Bar will be the same as the clockout time of Foo.
Is there another variable I need to set in order to enforce clockin-time=clockout-time?
org-version returns 9.1.2-37
- Ray




[-- Attachment #2: Type: text/html, Size: 3562 bytes --]

Re: org-clock-rounding-minutes -- non-zero value can cause double clocking

[Nicolas Goaziou]

Hello,

Raymond Zeitler <zeitra@yahoo.com> writes:

> There is a problem when org-clock-rounding-minutes is non-zero, say
> N.  The problem is that the clockin-time of a task can be N less than
> the clockout-time of the previous task at certain times.  Thus, clock
> reports can show an extra N minutes total time for each occurrence of
> this effect.
> Consider this.
> Set org-clock-rounding-minutes to 6. (This allows for tracking time in 0.1 hour increments.)
> Create an org file with two tasks:
> * Tasks** TODO Foo** TODO Bar
>
> Clock into Task Foo at, say, 12:54 plus or minus 2 minutes.  Its LOGBOOK will show the expected 12:54 timestamp.
> Clock into Bar at 12:57, which is halfway between 0.90th hour and 1.00th hour.
> Result, the clockout time for Foo is rounded up.  But the clockin time for Bar is rounded *down*.  Thus, two tasks have started at the same time.
> * Tasks** TODO Foo   :LOGBOOK:   CLOCK: [2019-03-08 Fri 12:54]--[2019-03-08 Fri 13:00] =>  0:06   :END:** TODO Bar   :LOGBOOK:   CLOCK: [2019-03-08 Fri 12:54]   :END:
> So when I use org-mode to track time for my weekly timesheet, it will report that my total clocked time for any given day is several minutes more than the time I've been at work!
> I expect that the clockin time of Bar will be the same as the clockout time of Foo.
> Is there another variable I need to set in order to enforce
> clockin-time=clockout-time?

What about `org-clock-continuously'?

Regards,

-- 
Nicolas Goaziou

Re: org-clock-rounding-minutes -- non-zero value can cause double clocking

[Raymond Zeitler]

[-- Attachment #1: Type: text/plain, Size: 688 bytes --]

 Nicolas Goaziou <mail@nicolasgoaziou.fr> wrote:
 
>Raymond Zeitler <zeitra@yahoo.com> writes:

>> There is a problem when org-clock-rounding-minutes is non-zero, say
>> N.  The problem is that the clockin-time of a task can be N less than
>> the clockout-time of the previous task at certain times.-snip-
>> I expect that the clockin time of Bar will be the same as the clockout time of Foo.
>> Is there another variable I need to set in order to enforce
>> clockin-time=clockout-time?
> What about `org-clock-continuously'?

It was set to nil.  I toggled it to non-nil, and the clock behavior is as desired.
Thank you for pointing me in the right direction!

- Ray  

[-- Attachment #2: Type: text/html, Size: 2439 bytes --]