Is there something that we're overlooking, or is this a bug? Thanks
for any help.
$ cat DividedDateTimeFieldDaylightSavingRoundingBug.java
import org.joda.time.*;
import org.joda.time.chrono.*;
import org.joda.time.field.*;
public class DividedDateTimeFieldDaylightSavingRoundingBug {
private static final DateTimeZone LOS_ANGELES =
DateTimeZone.forID("America/Los_Angeles");
/**
* Demonstrate that rounding the second 1AM of a DST-change day back to the
* nearest 10 minutes actually rounds it back by an hour.
*/
public static void main(String[] args) {
Instant fallSecond1Am = new Instant(1289120400000L);
// Sanity checks.
/*
* 1. We really do have the second 1:00AM of the fall DST switchover.
* Confirm this (a) by adding one hour to the first 1:00AM and (b) by
* comparing toString() to the expected value:
*/
Instant fallFirst1Am = new DateTime()
.withZone(LOS_ANGELES)
.withDate(2010, 11, 7)
.withTime(1, 0, 0, 0)
.toInstant();
assertEquals(
fallSecond1Am, fallFirst1Am.plus(Hours.ONE.toStandardDuration()));
assertEquals("2010-11-07T01:00:00.000-07:00",
new DateTime(fallFirst1Am.getMillis(), LOS_ANGELES).toString());
assertEquals("2010-11-07T01:00:00.000-08:00",
new DateTime(fallSecond1Am.getMillis(), LOS_ANGELES).toString());
/*
* 2. Rounding the FIRST 1AM back to the nearest 10-minute boundary (i.e.,
* itself) is a no-op, as expected:
*/
assertEquals(fallFirst1Am, floor10Minutes(fallFirst1Am));
/*
* Rounding the SECOND 1AM back to the nearest 10-minute boundary (i.e.,
* itself) should be a no-op, too, but Joda instead rounds a full hour back
* to the FIRST 1:00AM (note that the times in the failure message are UTC):
*
* expected:<2010-11-07T09:00:00.000Z> but was:<2010-11-07T08:00:00.000Z>
*/
assertEquals(fallSecond1Am, floor10Minutes(fallSecond1Am));
}
private static Instant floor10Minutes(Instant in) {
return new Instant(getTenMinuteField().roundFloor(in.getMillis()));
}
private static DividedDateTimeField getTenMinuteField() {
DateTimeFieldType minuteFieldType = DateTimeFieldType.minuteOfHour();
// Note: If we use getInstanceUTC() instead of GST, the test passes.
Chronology chronology = ISOChronology.getInstance(LOS_ANGELES);
DateTimeField oneMinuteField = minuteFieldType.getField(chronology);
return new DividedDateTimeField(oneMinuteField, minuteFieldType, 10);
}
private static void assertEquals(Object expected, Object actual) {
if (!expected.equals(actual)) {
throw new RuntimeException(
"expected:<" + expected + "> but was:<" + actual + ">");
}
}
}
$ javac -classpath .:joda-time-1.6.2.jar
DividedDateTimeFieldDaylightSavingRoundingBug.java && java -classpath
.:joda-time-1.6.2.jar DividedDateTimeFieldDaylightSavingRoundingBug
Exception in thread "main" java.lang.RuntimeException:
expected:<2010-11-07T09:00:00.000Z> but was:<2010-11-07T08:00:00.000Z>
at
DividedDateTimeFieldDaylightSavingRoundingBug.assertEquals(DividedDateTimeFieldDaylightSavingRoundingBug.java:63)
at
DividedDateTimeFieldDaylightSavingRoundingBug.main(DividedDateTimeFieldDaylightSavingRoundingBug.java:46)
------------------------------------------------------------------------------
Benefiting from Server Virtualization: Beyond Initial Workload
Consolidation -- Increasing the use of server virtualization is a top
priority.Virtualization can reduce costs, simplify management, and improve
application availability and disaster protection. Learn more about boosting
the value of server virtualization. http://p.sf.net/sfu/vmware-sfdev2dev
_______________________________________________
Joda-interest mailing list
Joda-interest@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/joda-interest
