Changes between Version 1 and Version 2 of Ticket #421, comment 4


Ignore:
Timestamp:
Oct 23, 2012, 7:30:38 AM (12 years ago)
Author:
olle

Legend:

Unmodified
Added
Removed
Modified

  • Ticket #421, comment 4

    v1 v2  
    55The problems seem to have different causes:
    66
    7  1. The problem for the monthly view was identified by Nicklas Nordborg as being caused by the way the monthly steps are created in the code, where e.g. 2011-01-30 is followed with what would literally be 2011-02-30, which is re-interpreted by the java method to be equal to 2011-03-02, resulting in data for January 2011 being followed by those for March 2011, while data for February 2011 is missing. While the short month of February is the typical case where this problem might appear, all months with less than 31 days might be affected (although only one month for each table).
     7 1. The problem for the monthly view was identified by Nicklas Nordborg as being caused by the way the monthly steps are created in the code, where e.g. 2011-01-30 is followed with what would literally be 2011-02-30, which is re-interpreted by the java method to be equal to 2011-03-02 for a non-leap year, resulting in data for January 2011 being followed by those for March 2011, while data for February 2011 is missing. While the short month of February is the typical case where this problem might appear, all months with less than 31 days might be affected (although only one month for each table).
    88 1. The problem for the weekly view seems to be connected with the fact that many week-number methods in Java and !JavaScript does not comply with the ISO week numbering system used in the EU.
    99 1. The missing data for the first year in the yearly view seems to be the result of a simple bug in the java servlet code (operator `<=` instead of `<`).