From: Felipe Contreras <felipe.contreras@...>
Date: 2013-10-07T13:32:29-05:00
Subject: [ruby-core:57711] Re: [PATCH 1/2] time: fix strptime '%s'

On Mon, Oct 7, 2013 at 10:17 AM, Tanaka Akira <akr@fsij.org> wrote:
> 2013/10/7 Felipe Contreras <felipe.contreras@gmail.com>:
>> '%s' is meant to imply UTC, however:
>>
>>   Time.strptime('0', '%s')
>>   => 1969-12-31 18:00:00 -0600
>>
>> After this patch:
>>
>>   Time.strptime('0', '%s')
>>   => 1970-01-01 00:00:00 +0000

> It seems the patch changes the behavior of Time.strptime("0", "%s").

That's what it says right above in the explanation ^.

> before the patch:
> % ruby -rtime -e 'p Time.strptime("0", "%s")'
> 1970-01-01 09:00:00 +0900
>
> after the patch:
> % ./ruby -rtime -e 'p Time.strptime("0", "%s")'
> 1970-01-01 00:00:00 +0000
>
> Is there a reason for that?

As it says on the explanation above: '%s' is meant to imply UTC.

However, I just read the man pages again:

strptime(3):
%s     The number of seconds since the Epoch, 1970-01-01 00:00:00
+0000 (UTC). Leap seconds are not counted unless leap second support
is available.

strftime(3):
%s     The number of seconds since the Epoch, 1970-01-01 00:00:00
+0000 (UTC). (TZ)

It doesn't really say '%s' is UTC, but that it's the seconds since the
Epoch (which is UTC).

In C strptime('0', '%s') is returning a local timezone, so perhaps my
patch is wrong in changing that, but strptime('0 +0100', '%s %z')
still needs to be fixed.

Cheers.

-- 
Felipe Contreras