On 2018-02-06 13:07:53 -0300, Lucas Castro wrote: > Em 06-02-2018 10:38, Vincent Lefevre escreveu: > > On 2018-02-06 13:48:19 +0100, Vincent Lefevre wrote: > > > This is completely crazy: > > > > > > zira% date +%Y-%m-%d -d '2003-09-01 1 day ago + 1 month' > > > 2003-09-30 > > > zira% date +%Y-%m-%d -d '2003-09-01 1 day ago' > > > 2003-08-31 > > > zira% date +%Y-%m-%d -d '2003-08-31 + 1 month' > > > 2003-10-01 > > > > > > So, while '2003-09-01 1 day ago' gives 2003-08-31, the following > > > are not equivalent: > > > * '2003-09-01 1 day ago + 1 month' > > > * '2003-08-31 + 1 month' > > > > > > Where is the logic behind that? > > Other ones: > > > > zira% date +%Y-%m-%d -d '2003-02-01 - 1 month' > > 2003-01-01 > > zira% date +%Y-%m-%d -d '2003-02-01 - 31 days' > > 2003-01-01 > > zira% date +%Y-%m-%d -d '2003-02-01 - 31 days + 1 month' > > 2003-01-29 > I guess first step done is the sum and then, subtract 31 days. > So, as the fev is a month just with 28 days, end up at 2003-01-29.
According to --debug, it seems that operations on years are done first, then months, then days. But this is not documented in the Coreutils manual. > > zira% date +%Y-%m-%d -d '2003-02-01 - 1 month + 1 month' > > 2003-02-01 There's a bug on the same subject: https://debbugs.gnu.org/cgi/bugreport.cgi?bug=26101 -- Vincent Lefèvre <vinc...@vinc17.net> - Web: <https://www.vinc17.net/> 100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/> Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)
