Hey Aaron,

Good catch pinpointing the FileMetric data.

So, I did some investigating and found out several interesting things
about these files.  First off, the files that were causing the timezone
problems were actually 10-16-2004.xml and 10-17-2004.xml.  The data on
the 15th was valid and accurate.

The reason these files were causing problems is that in the test data, I
have FileMetric data coming from both LOCC and Sensor Shell (proxy for
BCML).  All the sensor data from LOCC has a timestamp related to the
date of the file.  However, the Sensor Shell FileMetric data is from
unrelated dates.  I did this on purpose to make sure that in the test,
the FileMetric data was coming purely from LOCC and not Sensor Shell.
What I did not take into account was that the somewhat arbitrary
timestamp of the Sensor Shell FileMetric data would cause a problem with
the TimeZoneChanger.  Actually, I believe this test data was created
just before the changer was implemented.

Anyway, I have made fixes to the test data, completed a freshStart and
junitAll and committed the revised test data.  Although, it will not be
used by Ikayzo, it will be safe for future timezone changes.

Thanks,
Mike

Aaron Kagawa wrote:

Hey Guys,

I have figured out that the hackyHPCS contains the FileMetric data
that is
failing the Timezone Changer.

The two files that are causing the problems are located in the
"hackyHPCS\src\org\hackystat\app\hpcs\testdataset\data\FileMetric"
directory (2004-10-15.xml and 2004-10-17.xml). Using Hongbing's
JavaScript
UTC Conversion tool, you will see that this UTC 1099090000041 actually
specifies a time in October 29th.

Mike, can you fix the timestamps in these files? To correctly fix the
problem you have to ensure that all timestamps within the file
2004-10-17.xml are times during October 17, 2004. You need to do the same
thing for 2004-10-15.xml.

thanks, aaron

ps. I've made this issue in Jira for the upcoming 6.5 release. also note
that the Hackystat-standard which I will be using for Ikayzo will not
contain hackyHPCS, therefore you might want to try out the Timezone
changer
on your local machine.

1) The timezone changer detected a problem in a 2 of the testdata.
According to the Timezone changer the timestamps in a couple of the
files
are incorrect. For example, in the FileMetric data on 2004-10-17, at
least
one timestamp references a date in October 30th. This needs to be fixed.
I'm not sure who these testdata belongs to and where to find them at the
moment.

04/20 17:49:07 TimeZoneChangerException: Error processing
/home/hack/hackystat/data/users/testdataset/data/FileMetric/2004-10-17.xml

org.hackystat.kernel.admin.TimeZoneChangerException: The calculated
offset
causes a change in day. [day(from file)=17-Oct-2004|
originalTimeStamp=1099090000041, day=30-Oct-2004, date=Sat Oct 30
07:46:40
JST 2004 | adjustedTimeStamp=1099021600041, day=29-Oct-2004, date=Fri
Oct
29 12:46:40 JST 2004]
04/20 17:49:07 TimeZoneChangerException: Error processing
/home/hack/hackystat/data/users/testdataset/data/FileMetric/2004-10-16.xml

org.hackystat.kernel.admin.TimeZoneChangerException: The calculated
offset
causes a change in day. [day(from file)=16-Oct-2004|
originalTimeStamp=1097830800002, day=15-Oct-2004, date=Fri Oct 15
18:00:00
JST 2004 | adjustedTimeStamp=1097762400002, day=14-Oct-2004, date=Thu
Oct
14 23:00:00 JST 2004]


Reply via email to