Hello Volker, ok, thank you. I thought, the converter uses the default timezone, as toString() does, and not GMT by default.
This issue seems to be a reason for many discussions... Regards Helmut >Hi Helmut, > >this is exact the expectet behavior. > >The jsf spec says the default timezone for converter is GMT, a >date.toString() uses the systemDefault timezone, this makes the >difference. > >see this thread for more: >http://www.mail-archive.com/users%40myfaces.apache.org/msg21412.html > >Regards, > Volker > > >2006/12/21, H. Swaczinna <[EMAIL PROTECTED]>: >> Another example: >> >> Initializtion with date = new Date() at 13 minutes after midnight. >> Date.toString() prints "Thu Dec 21 00:13:48 CET 2006", but the >> page shows "20.12.2006" and I got back "Wed Dec 20 01:00:00 CET 2006" >> after submit. So every date with time part less than one our is dated >> back one day. When I execute the same code after one o'clock in the >> morning, the date is calculated correct (but the time part is always >> set to one o'clock, where I expect zero). >> >> Regards >> Helmut >> >> >Hello, >> > >> >I've got a tc:date in my page which represents a java.util.Date in >> >my bean. The Date is initialized with >> >Date date = (new GregorianCalendar(2006, 8, 1)).getTime(), so the >> >time part is 00:00:00. Date.toString() prints "Fri Sep 01 00:00:00 CEST >2006", >> >which is correct. But in the page, I see "31.08.2006" (with >> ><f:convertDateTime pattern="dd.MM.yyyy" />). When I submit the page, the >> >date is written back as "Thu Aug 31 02:00:00 CEST 2006", which is >> >definitely wrong. When I initialize the date with >> >date = (new GregorianCalendar(2006, 11, 1)).getTime(), I got back >> >"Thu Nov 30 01:00:00 CET 2006" from the page. So, the behaviour is >> >different, if the date lies in the daylight saving period or not, and >> >the day of month is always miscalculated by one day minus one or to hours. >> > >> >Maybe, the initialization of the date might be wrong or the converter >> >is not setup correctly. But I think, there's also a problem in the >> >converter. >> > >> >Regards >> >Helmut >>
