On Wed, Jun 6, 2012 at 9:34 AM, Andy Seaborne <[email protected]> wrote: > On 06/06/12 14:59, Andy Seaborne wrote: >> >> On 06/06/12 14:45, Andy Seaborne wrote: >>> >>> On 04/06/12 17:22, Andy Seaborne wrote: >>>> >>>> On 04/06/12 14:29, Spates, Rick wrote: >>>>> >>>>> Andy, >>>>> >>>>> Have you had a chance to look at this any further? Assuming the need >>>>> for a fix is validated I was hoping to get it into a release of our >>>>> clinical application before the Fall. >>>>> >>>>> Thanks, >>>>> Rick >>>> >>>> >>>> It's recorded as JENA-247 [1]. >>>> >>>> It's not an area of the system I'm particularly familiar with and the >>>> forcing to a particular timezone (other than I think GMT should be UTC - >>>> up to a second different!) is presumably because of how Xerces >>>> internally records date/times. My caution is because fixing one thing >>>> must not unfix something else. >>>> >>>> I hope to get it into the release we're currently working, either by >>>> finding a chunk of time to understand that part of the system or by >>>> sync'ing up with someone who does. >>>> >>>> Andy >>>> >>>> [1] https://issues.apache.org/jira/browse/JENA-247 >>> >>> >>> >>> Resolved - the fix is to always rebase times to GMT before further >>> calculation. >>> >>> Please confirm. >> >> >> I spoke too soon - there is a test failure (missed because Eclipse is a >> little keen to choose jars over class folder hierarchies when both have >> a class to match). > > > better now - your test added, learnt something (not exactly sure what) about > java and GregorianCalendars. >
The thing I've learned about Java Calendars is to use Joda Time! Hopefully JSR-310 makes it into Java 8. -Stephen
