work-week ends do not seem to be calculated correctly

[smemsh]

Work week endings may be calculated incorrectly. Observe the following:

$ for ww in {s,e}o{,p,n}ww
  do printf "%s\t%s\n" $ww $(
      faketime '2017-03-05 12:34:56' task calc $ww
  ); done

soww    2017-03-13T00:00:00
sopww   2017-02-27T00:00:00
sonww   2017-03-13T00:00:00
eoww    2017-03-10T23:59:59
eopww   2017-03-03T23:59:59
eonww   2017-03-17T23:59:59

with the reference date of 2017-03-05 12:34:56 (a Sunday), and the calendar:

 $ ncal 03 2017
    March 2017        
Mo     6 13 20 27   
Tu     7 14 21 28   
We  1  8 15 22 29   
Th  2  9 16 23 30   
Fr  3 10 17 24 31   
Sa  4 11 18 25      
Su  5 12 19 26      
    9 10 11 12 13   

these do not look right:

• It seems to consider the next "work week" to have already begun on Saturday, but I should think the next work week doesn't actually begin until Monday. Even though there's no actual work days left, it's still within the same work week.
• soww and sonww are identical
• eoww is earlier than soww (!)

This issue was noticed during testing for #47 (a change to comments documenting behavior of eo*) and is forked into a standalone issue as per the discussion there.

(test version of libshared was 591fa58)

[tbabej]

Sigh. I said #66 partially resolves this, Github. We still need to fix / discuss the "It seems to consider the next "work week" to have already begun on Saturday" aspect of this bug report.

[smemsh]

These look correct now for weekend days such as today, except that eoww and eopww should not be the same:

 $ for ww in {s,e}o{,p,n}ww; do printf "%s\t%s\n" $ww $(
    faketime '2021-11-20 14:21:46' src/task calc $ww); done
soww    2021-11-15T00:00:00
sopww   2021-11-08T00:00:00
sonww   2021-11-22T00:00:00
eoww    2021-11-19T23:59:59
eopww   2021-11-19T23:59:59
eonww   2021-11-26T23:59:59

 $ ncal
    November 2021
Mo  1  8 15 22 29
Tu  2  9 16 23 30
We  3 10 17 24
Th  4 11 18 25
Fr  5 12 19 26
Sa  6 13 20 27
Su  7 14 21 28
   44 45 46 47 48

because, my assertion is, the next work week doesn't actually start until Monday, so the "previous" work week is still the one before this week's Monday. Even though it's after the end of the "work week" (I guess Friday), it's before the beginning of the next one..

All of the following task calc tests should be true, here below run with your patch:

 $ for test in "${tests[@]}"
   do printf "$test -> "; src/task calc "$test"; done
soww == (sonww - 7 days) -> true
eoww == (eonww - 7 days) -> true
(sopww + 7 days) == soww -> true
(eopww + 7 days) == eoww -> false

prior to your patch there were 3 false and a 1 true so it's closer ;-)