JDK-4044013 : VM not detecting host timezone correctly (was Clock is not displaying correct ti
  • Type: Bug
  • Component: core-libs
  • Sub-Component: java.util:i18n
  • Affected Version:
    1.1,1.1.2,1.1.3,1.1.4,1.1.6,1.1.7,1.2.0 1.1,1.1.2,1.1.3,1.1.4,1.1.6,1.1.7,1.2.0
  • Priority: P3
  • Status: Closed
  • Resolution: Duplicate
  • OS:
    generic,solaris_2.5.1,solaris_2.6,windows_95,windows_nt generic,solaris_2.5.1,solaris_2.6,windows_95,windows_nt
  • CPU: generic,x86,sparc
  • Submitted: 1997-04-09
  • Updated: 1999-09-21
  • Resolved: 1999-09-21
Related Reports
Duplicate :  
Duplicate :  
Duplicate :  
Duplicate :  
Duplicate :  
Duplicate :  
Duplicate :  
Duplicate :  
Duplicate :  
Description
The clock does not show the correct time while the windows 95 system does.
We are here in Germany using MET and have changed to Daylight Savings
Time some days ago.

It is now 3.50 pm and Hotjava shows 2:50 pm.

Form the personal properties file:
	#HotJava Properties
	#Thu Apr 03 14:52:43 GMT+01:00 1997

===========================================================

Clock gets set to US/Eastern rather than Australian/NSW 

Date:    7 Apr 1997 10:21:54 -0700

===========================================================

The clock seems to display time in GMT-5, but local timezone is GMT-3.

A clocktool running on the same hosts display 11:57PM, while the clock of
hotjava displays 9:57PM. (Company: Institute of Computing, State University of Campinas)

Date:      3 Apr 1997 19:18:48 -0800

===========================================================

I am running HJ 1.0 on Solaris 2.5 on an Ultra 1. The timezone is set to
MET. But when I start HJ, the time is off by 1.5 hours. For example the
realworld time of 2:00PM is displayed at 3:30PM. And the timezone
displayed is GMT+03:30

This is a *VERY* visible problem. Each and every one of our customers
will complain about this!!!

Date: 9 Apr 1997 05:24:35 -0700

===========================================================
Running "date" on the command line gives something like
             Fri Apr 11  18:09:28 MET DST 1997
    which we interpret as Middle EuropeanTime - and which is correct by the way.
    HotJava shows
             Apr 11,=04 7:39 PM
    which is obviously  1.5 hours ahead - and which we cannot
    interpret as another time zone, ....

    Furthermore, we can NOT find anything but "Clock On - or - Off"
    within HotJava.

    When using JDK 1.1.1 -  using the "date class" - produces similar
    errors, we guess that MET is interpreted as something strange in
    MiddleEast or .... 

Date:  11 Apr 1997 09:43:45 -0700

===========================================================
I have this in my .hotjava/properties file:
hotjava.clock=on
hotjava.clock.format=yMMdd 'v'w: kk:mm

the shell environment variable LANG is set to sv (i.e Swedish)
with this configuration the hotjava clock shows:
970414 v15:02:25
when the "date"  command from the shell 
shows the correct time of this writing:

14 april 1997 kl 00:56:28 MET DST

Date: 13 Apr 1997 16:09:58

Comments
EVALUATION This sounds like a dup of the daylight savings time bug that JWS ran into. Reassigning for confirmation bill.foote@Eng 1997-06-02 ------ This is an oft-duped bug in which the VM fails to correctly determine the host OS timezone setting and communicate it to the java.util.TimeZone class (via the user.timezone property). This is partially fixed in 1.1.6 but needs a more robust fix for 1.2. -alanl SEE ALSO 4059578 4064437 4085143 4089103 4092412 4105207 4105412 Platform-to-Java timezone mapping problems have been fixed in Kestrel. Any fixes before Kestrel are not very reliable. masayoshi.okutsu@Eng 1999-09-21
21-09-1999