RE: Bug ID 6342679: NSTimestamp Happy New Year Bug
RE: Bug ID 6342679: NSTimestamp Happy New Year Bug
- Subject: RE: Bug ID 6342679: NSTimestamp Happy New Year Bug
- From: Barði Einarsson <email@hidden>
- Date: Wed, 14 Jan 2009 17:11:12 -0000
- Thread-topic: Bug ID 6342679: NSTimestamp Happy New Year Bug
OK, so when one gives NSTimestamp a null as a time zone
it uses GMT, not the default time zone.
However, GregorianCalendar uses the default time zone.
So, if one wants to use time zones with NSTimestamp, not
just GMT, one has to explicitly specify the default time
zone in the NSTimestamp constructor. Otherwise, GregorianCalendar
will give one unexpected results.
The mismatch in the use of default time zones between NSTimestamp
and GregorianCalendar is a bug.
Comments please.
Best regards,
Bardi
-----Original Message-----
From: Chuck Hill [mailto:email@hidden]
Sent: Wed 14-Jan-09 15:15
To: WebObjects (Group)
Cc: Barði Einarsson
Subject: Re: Bug ID 6342679: NSTimestamp Happy New Year Bug
On Jan 14, 2009, at 3:30 AM, Ken Anderson wrote:
> NO! It is NOT an incorrect result! The reason you get the result
> of 1963 when you use 1964 is because Reykjavik is 1 hour off from
> GMT then.
As confusing as it may be, Ken is correct. It is your understanding
of how date/times are manipulated that is at fault. This is a complex
and confusing area. If you use noon 12:00:00 instead of midnight
00:00:00 in the test dates the result should be as you expect as the
time shift will only affect the time portion. The time shift will
still happen, but it won't affect the day, month, or year.
Chuck
> On Jan 14, 2009, at 5:11 AM, Barði Einarsson wrote:
>
>> Ken
>>
>> If I use Atlantic/Reykjavik the code produces incorrect
>> results. So, it is a bug.
>>
>> If I use GMT instead the code would work. That is a
>> work around for a bug.
>>
>> Do you agree?
>>
>> Best regards,
>> Bardi
>>
>>
>> -----Original Message-----
>> From: Ken Anderson [mailto:email@hidden]
>> Sent: Tue 13-Jan-09 17:59
>> To: Barði Einarsson
>> Cc: email@hidden
>> Subject: Re: Bug ID 6342679: NSTimestamp Happy New Year Bug
>>
>> Bardi,
>>
>> OK - it took me all of 10 minutes to solve this.
>>
>> 1. You choose Atlantic/Reykjavik, not GMT, like we suggested.
>> 2. In 1969, Iceland stopped respecting Daylight Savings Time: http://www.timeanddate.com/worldclock/timezone.html?n=211&syear=1960
>>
>>
>> date in yearFromDate : 1968-01-01 00:00:00 Etc/GMT
>> yearFromDate : 1967
>> TimeZone:Atlantic/Reykjavik
>> date in yearFromDate : 1969-01-01 00:00:00 Etc/GMT
>> yearFromDate : 1969
>>
>> If you had chosen GMT, where there's no shifts for daylight savings
>> time, it's always accurate.
>>
>> Ken
>>
>> On Jan 13, 2009, at 12:02 PM, Barði Einarsson wrote:
>>
>>> It is a bug:
>>>
>>> run happyNewYear(1964)
>>>
>>> output:
>>>
>>> TimeZone:Atlantic/Reykjavik
>>> date in yearFromDate : 1964-01-01 00:00:00 Etc/GMT
>>> yearFromDate : 1963
>>>
>>>
>>> package some.packagename;
>>>
>>> import java.util.GregorianCalendar;
>>> import java.util.TimeZone;
>>>
>>> import com.webobjects.foundation.NSTimeZone;
>>> import com.webobjects.foundation.NSTimestamp;
>>>
>>> public class HappyNewYear {
>>>
>>> static {
>>> TimeZone.setDefault(TimeZone.getTimeZone("Atlantic/
>>> Reykjavik"));
>>> NSTimeZone.setDefault(TimeZone.getTimeZone("Atlantic/
>>> Reykjavik"));
>>> }
>>>
>>>
>>> public static int yearFromDate(NSTimestamp date) {
>>> if (null == date) {
>>> return 0;
>>> }
>>> System.out.println("TimeZone:" +
>>> TimeZone.getDefault().getID());
>>> GregorianCalendar cal = new GregorianCalendar();
>>> cal.setTime(date);
>>> System.out.println("date in yearFromDate : " + date);
>>> int result = cal.get(GregorianCalendar.YEAR);
>>> System.out.println("yearFromDate : " + result);
>>> return result;
>>> }
>>>
>>> public static void happyNewYear(int year) {
>>> NSTimestamp newYear = new NSTimestamp(year, 1, 1, 0, 0, 0, null);
>>> yearFromDate(newYear);
>>> }
>>>
>>> }
>>>
>>>
>>>
>>> -----Original Message-----
>>> From: Ken Anderson [mailto:email@hidden]
>>> Sent: Fri 09-Jan-09 15:23
>>> To: Lachlan Deck
>>> Cc: Barði Einarsson; WebObjects-Dev Mailing List List
>>> Subject: Re: Bug ID 6342679: NSTimestamp Happy New Year Bug
>>>
>>> Bardi,
>>>
>>> I think the most important thing to understand about how dates are
>>> moved around is that the value is always GMT based. Timezones are
>>> there for convenience...
>>>
>>> For instance, when you create a new timestamp with a timezone, the
>>> timezone is used to adjust the date/time you sent in to GMT and the
>>> GMT value is stored. When formatting a date or extracting day/
>>> month/
>>> year from GregorianCalendar, the timezone is used WHEN FORMATTING
>>> THE
>>> RESULT.
>>>
>>> If you work with these tenets, things will hopefully be easier.
>>>
>>> Ken
>>>
>>> On Jan 8, 2009, at 6:54 PM, Lachlan Deck wrote:
>>>
>>>> On 09/01/2009, at 3:10 AM, Ken Anderson wrote:
>>>>
>>>>> I do exactly what you're doing all the time, and I do not have the
>>>>> issue you're experiencing. One thing I do at the beginning of my
>>>>> WOApplication subclass is this:
>>>>>
>>>>> TimeZone.setDefault(TimeZone.getTimeZone("GMT"));
>>>>> NSTimeZone.setDefault(TimeZone.getTimeZone("GMT"));
>>>>>
>>>>> This will make sure that you're really working in GMT across the
>>>>> board. Can you do the above and try your tests again?
>>>>
>>>> Yep.
>>>>
>>>>>> Behaves correctly - no way
>>>>
>>>> Yes way :-)
>>>>
>>>>>> NSTimestamp reports a timezone which is not used by
>>>>>> GregorianCalendar ??
>>>>
>>>> GregorianCalendar does *not* change its timezone based on the date
>>>> passed to it. You are responsible for determining that. That's
>>>> nothing to do with NSTimestamp.
>>>>
>>>>>> This is a difficult bug.
>>>>
>>>> Just more a misunderstanding on your part.
>>>>
>>>> with regards,
>>>> --
>>>>
>>>> Lachlan Deck
>>>>
>>>> _______________________________________________
>>>> Do not post admin requests to the list. They will be ignored.
>>>> Webobjects-dev mailing list (email@hidden)
>>>> Help/Unsubscribe/Update your Subscription:
>>>>
>>>> This email sent to email@hidden
>>>
>>>
>>>
>>> Fyrirvari á tölvupósti / e-mail disclaimer
>>> http://us.is/Apps/WebObjects/US.woa/wa/dp?id=3776
>>>
>>>
>>> _______________________________________________
>>> Do not post admin requests to the list. They will be ignored.
>>> Webobjects-dev mailing list (email@hidden)
>>> Help/Unsubscribe/Update your Subscription:
>>>
>>> This email sent to email@hidden
>>
>>
>>
>> Fyrirvari á tölvupósti / e-mail disclaimer
>> http://us.is/Apps/WebObjects/US.woa/wa/dp?id=3776
>>
>>
>
> _______________________________________________
> Do not post admin requests to the list. They will be ignored.
> Webobjects-dev mailing list (email@hidden)
> Help/Unsubscribe/Update your Subscription:
>
> This email sent to email@hidden
>
--
Chuck Hill Senior Consultant / VP Development
Practical WebObjects - for developers who want to increase their
overall knowledge of WebObjects or who are trying to solve specific
problems.
http://www.global-village.net/products/practical_webobjects
Fyrirvari á tölvupósti / e-mail disclaimer
http://us.is/Apps/WebObjects/US.woa/wa/dp?id=3776
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden