Re: [Boost-users] Support for extended Daylight Saving Time in 2007

From: Jeff Garland (jeff_at_[hidden])
Date: 2007-01-10 10:17:38

Stephen Turner wrote:
>> Date: Mon, 8 Jan 2007 21:25:24 -0500
>> From: "Eric Yankowitz" <nnews_at_[hidden]>
>> Subject: [Boost-users] Support for extended Daylight Saving Time in
>> 	2007
>>
>> Been trying to find information describing boost's level of 
>> support for the
>> four week Daylight Saving Time extension in the US that will 
>> occur this
>> March. 
>> (http://geography.about.com/cs/daylightsavings/a/dst.htm), I haven't
>> been able to find very much. A lot of software is having 
>> issues and was
>> wondering if Boost was ready for the event?
>>
>> Was able to find these links, which suggested issues:
>>
>>
>> http://sourceforge.net/tracker/index.php?func=detail&aid=14717
>> 23&group_id=7586&atid=207586
>>
>>
>> http://sourceforge.net/tracker/index.php?func=detail&aid=14786
>> 19&group_id=7586&atid=207586
>>
>> Any idea if these are being resolved? And if any others exist?
>>

Sorry for the extended delay on responses....off the net on vacation for an 
extended period.  Short answer is that no, they haven't been resolved yet.

> Eric,
> 
> I'm a new user of this code, so take everything I say with the appropriate
> amount of caution, but I have looked into this question a little.
> 
> My conclusion is that the date_time library is still shipping with the old
> timezone rules at the moment. It's easy to change the date_time_zonespec.csv
> file to use the new rules instead (you can do that yourself, and I assume
> that the official version will do it soon too). 

Yep, in fact I was looking for someone that has time to look at this carefully 
to do this update and send me a patch -- someone about a month ago said they 
would do this, but haven't seen anything yet.  As for when it will be 
released, I may have to ask the release manager to permit an update.  I wasn't 
expecting 1.34 release to drag into 2007.

> However, at the moment the
> code doesn't seem to support different timezone rules for different years,
> so once you've changed it, you're out of luck if you want to look up times
> in 2006. See my previous post about this subject at
> http://article.gmane.org/gmane.comp.lib.boost.user/23837

This is true -- does your application require historic dates with timezone 
lookups or only current dates?

Jeff