On 14/02/15 04:35, Stan Cox wrote:
> The datetime support for the -S/-T options was not supporting 'next
> TODAY' properly. The desired behavior is one week from today. This
> patch fixes that and adds a bit more testing.
>
> git://sourceware.org/git/pcpfans.git
> scox/dev b47af16c219f78d96f3962e61d022ca9006ad3a8
> Stan Cox (1): Match current parse-datetime.y behavior for 'next *TODAY*'
Stan,
I'm working on revamping qa/752 to make it more robust ... and seeing this
issue ...
# grep sunday 752.full
sunday: 2015-02-15 <---- this one is from date(1)
first sunday: 2015-02-22
this sunday: 2015-02-15
next sunday: 2015-02-22
last sunday: 2015-02-08
#1 "sunday" 2015-02-22 00:00:00 <--- from rtimetest
#2 "sunday" 2015-02-22 00:00:00
#3 "sunday" 2014-01-27 11:28:50
#1 "first sunday" 2015-02-22 00:00:00
#2 "first sunday" 2015-02-22 00:00:00
#3 "first sunday" 2014-01-27 11:28:50
#1 "this sunday" 2015-02-15 00:00:00
#2 "this sunday" 2015-02-15 00:00:00
#3 "this sunday" 2014-01-27 11:28:50
#1 "next sunday" 2015-02-22 00:00:00
#2 "next sunday" 2015-02-22 00:00:00
#3 "next sunday" 2014-01-27 11:28:50
#1 "last sunday" 2015-02-08 00:00:00
#2 "last sunday" 2015-02-08 00:00:00
#3 "last sunday" 2014-01-27 11:28:50
The libpcp one as a week ahead of the date(1).
The qa test now exercises _every_ day of the week, so I'll update tomorrow if
monday is now wrong in a similar way.
|