[SNMP4J] DateAndTime makeCalender improvement

[Frank Fock]

Hi Pim,

Yes, the issue is already fixed in the latest snapshot for some days now.

Best regards,
Frank


> On 20 Apr 2016, at 10:56, Pim Moerenhout <pim.moerenhout at gmail.com> wrote:
> 
> Hi Frank,
> Could you reproduce the error with this snippet?
> Regards Pim
> 
> 2016-04-09 18:40 GMT+02:00 Pim Moerenhout <pim.moerenhout at gmail.com <mailto:pim.moerenhout at gmail.com>>:
> Hi Frank,
> 
> This test case test the GregorianCalendar with all possible timezones which
> fails with the 2.4.2 DateAndTime.makeCalendar
> 
> public void testDateAndTime() throws Exception {
>   // 1st april 2016 UTC 00:00:00
>   long testMillis = 1459468800000L;
>   for (String timeZoneId : TimeZone.getAvailableIDs()) {
>     GregorianCalendar gc = new
> GregorianCalendar(TimeZone.getTimeZone(timeZoneId));
>     gc.setTimeInMillis(testMillis);
>     OctetString dateAndTime = DateAndTime.makeDateAndTime(gc);
>     int error = DateAndTime.validateDateAndTime(dateAndTime);
>     if (error != 0) {
>       System.out.println("SNMP error: " + error);
>     }
> 
>     // check the makeCalendar from the generated DateAndTime
>     GregorianCalendar parsedGc = DateAndTime.makeCalendar(dateAndTime);
>     long parsedGcMillis = parsedGc.getTimeInMillis();
>     System.out.println(String.format("[%32s] The parsed millis %s: %5.1f
> hours (" + gc
>             .getTime() + " " + parsedGc.getTime() + ")", timeZoneId,
>         (parsedGcMillis != testMillis ? "differs" : "equals "),
>         (parsedGcMillis - testMillis) / 3600000.0));
>   }
> }
> 
> Most timezones fail, except when using the default timezone (or equivalent).
> 
> Regards, Pim
>