11

I am having troubles understanding how to manage recurring tasks in taskwarrior

I start with an empty database:

$ task
[task next]
No matches.

I add a recurring daily task:

$ task add recur:daily due:later test

It shows up in the report:

$ task
[task next]

ID Age Recur Due   Description Urg 
 2 -     P1D 19.2y test         2.4

1 task
Creating recurring task instance 'test'

If I mark it done like this:

$ task 2 done
Completed task 2 'test'.
Completed 1 task.
$ task
[task next]
No matches.

it disappears from the report. I believe it makes sense, since "I completed the daily task today".

The problem is it never appears again the next day and further. What am I doing wrong?

AnonymousLurker
  • 1,033
  • 2
  • 12
  • 18

2 Answers2

10

I was messing around with recurring tasks and figured it out:

Some background: Creating a recurring tasks creates a parent task which spawns child tasks.
from: https://taskwarrior.org/docs/recurrence/

with some testing, I've found that with this command:

    (ins)[>]task add "test" recur:5s due:30s
    Created task 27 (recurrence template).

Firstly, due affects the first child task. Every child task spawned after due-time (30 seconds in my example) then is due in recur-time (5 seconds in example)

    (ins)[>]task
    [task next]

ID Age Deps P Project Recur Due  Description
22 -                   PT5S 30s  test

#some time later...

(ins)[>]task
[task next]

ID Age Deps P Project Recur Due  Description
22 60s                 PT5S -30s test
23 30s                 PT5S -25s test
24 25s                 PT5S -20s test
25 20s                 PT5S -15s test
26 15s                 PT5S -10s test
27 10s                 PT5S -5s  test
28  5s                 PT5S -    test
29  0s                 PT5S 5s   test

However due also affects one more thing: when the second child task is spawned (the second child task is due in recur-time so the third child task spawns after recur-time and so on).

This is why you aren't seeing your second child task since you have the due date set to later which is 19.2 years time in the future according to your example; the second child task will therefore spawn only after 19.2 years time later (after which each child task will spawn daily).

Sam Johnson
  • 271
user821596
  • 101
4

I previously answered this same question over in unix & linux. RPosting the answer again, for record.

Before I dive into the working details, note that recurrence is not well-designed. What I bring is empirical observations and liable to change with a rework.

TLDR; How do I use and make sense of recurrence

task add recur:<duration> due:<first task due date> until:<delete first task by date> wait:<date when task will appear> "my task"

Example:

  • task add recur:daily due:9:00 until:12:00 wait:5:00 eat breakfast
  • Eat breakfast at 9:00 AM
  • Hide task until 5:00 AM
  • Remove task by lunch time
  • Create a new occurrence everyday

Details

Recurrence requires a due attribute (enforced with a validation check). Recurrence can also use until and wait attributes, if they exist. Recurrence does not use scheduled. And there is a mask attribute, that tracks which instance of the reoccurence the task is.

When a new recurring instance is created:

  • new task due = parent due + recur * mask
  • identical math for both until and wait
  • scheduled is copied wholesale from the parent task. In practice, this unexpected behaviour won't break your workflow, just affects filters and skews urgency. I call this out because it took me a very long time to discover this was not working as expected.

Every time taskwarrior runs, it checks to see if it's time to create the next occurrence. The psuedo-equation is:

now >= due + recur * (mask + 1)`

Note, there is a config value, rc.recurrence.limit, that can be used to tell taskwarrior to create even more occurrences further out.

You can expect the next occurrence to be created at the due date of the previous.

user10930673
  • 41