FW: Problem in DST fall back procedure

[Koski Avi]

Chris,

The result I got refer to Tru64 4.0D (HP) machine. I just modified the
source file "northamerica", before running zic.

/Avi

-----Original Message-----
From: Chris Walton [mailto:Chris.Walton at telus.com] 
Sent: Monday, November 13, 2006 9:54 PM
To: Koski Avi
Cc: tz at elsie.nci.nih.gov
Subject: RE: FW: Problem in DST fall back procedure


Avi,

Paul is correct.
Try setting the time explicitly using UTC: "date -u 110409552007".

As a side issue...
I am puzzled as to the output from "zdump -v Canada/Yukon" you provided
in your e-mail.
It is reporting time zone abbreviations of YST and YDT and showing the
same offsets as AKST and AKDT.
Canada/Yukon is mapped to America/Whitehorse which has been using
PST/PDT for many years.
When I run "zdump -v Canada/Yukon | grep 2007" I get a very different
output from you.
This is what I get:
 Canada/Yukon  Sun Mar 11 09:59:59 2007 UTC = Sun Mar 11 01:59:59 2007
PST isdst=0
 Canada/Yukon  Sun Mar 11 10:00:00 2007 UTC = Sun Mar 11 03:00:00 2007
PDT isdst=1
 Canada/Yukon  Sun Nov  4 08:59:59 2007 UTC = Sun Nov  4 01:59:59 2007
PDT isdst=1
 Canada/Yukon  Sun Nov  4 09:00:00 2007 UTC = Sun Nov  4 01:00:00 2007
PST isdst=0
This is what you reported:
 Canada/Yukon  Sun Mar 11 10:59:59 2007 GMT = Sun Mar 11 01:59:59 2007
YST isdst=0 gmtoff=-32400
 Canada/Yukon  Sun Mar 11 11:00:00 2007 GMT = Sun Mar 11 03:00:00 2007
YDT isdst=1 gmtoff=-28800
 Canada/Yukon  Sun Nov  4 09:59:59 2007 GMT = Sun Nov  4 01:59:59 2007
YDT isdst=1 gmtoff=-28800
 Canada/Yukon  Sun Nov  4 10:00:00 2007 GMT = Sun Nov  4 01:00:00 2007
YST isdst=0 gmtoff=-32400

Did you make any changes to the source files before you compiled and
installed?

-chris

-----Original Message-----
From: Paul Eggert [mailto:eggert at CS.UCLA.EDU] 
Sent: November 13, 2006 2:00 PM
To: avi.koski at comverse.com
Cc: tz at lecserver.nci.nih.gov
Subject: Re: FW: Problem in DST fall back procedure

"Olson, Arthur David \(NIH/NCI\) [E]" <olsona at dc37a.nci.nih.gov> writes:

> We cahnged time to 01:55 , 4 Nov 2007. 
> The time advances normally from 02:00hrs on and "does not fall back to

> 01:00hrs as desired".

My guess is that you set the time to the _second_ 01:55 on that date.
So the clock _should_ advance normally through 02:00.  If you want to
observe a jump, you have to set the time to the _first_ 01:55 on that
date.

> # date
> Sun Nov  4 02:01:23 EST 2007

So far so good; that's a valid time stamp.

> # zdump -v US/Alaska | grep 2007
> US/Alaska  Sat Nov  3 23:05:42 2007 AKDT US/Alaska  Sun Mar 11 
> 10:59:59 2007 GMT = Sun Mar 11 01:59:59 2007 AKST isdst=0 
> gmtoff=-32400 US/Alaska  Sun Mar 11 11:00:00 2007 GMT = Sun Mar 11 
> 03:00:00 2007 AKDT isdst=1 gmtoff=-28800 US/Alaska  Sun Nov  4 
> 09:59:59 2007 GMT = Sun Nov  4 01:59:59 2007 AKDT isdst=1 
> gmtoff=-28800 US/Alaska  Sun Nov  4 10:00:00 2007 GMT = Sun Nov  4 
> 01:00:00 2007 AKST isdst=0 gmtoff=-32400

This is also valid behavior; the clock jumps backward just before 02:00
on 2007-11-04.  Your example for Yukon was similar.

So far, I don't see a bug.