Please review java.util.formatter bug fix

[roger riggs]

HI Sherman,

Thanks for the review, I should have restored N=12 after debugging;
the previous  non-error output was hiding what I needed to see.

If we want to hit DST changes then blanketing 2/3 weeks or weekends
in the spring or fall might have a greater chance of a hit then 12/365.

Thanks, Roger

On 8/13/2013 4:16 PM, Xueming Shen wrote:
> Looks fine.
>
> Since you have already turned the verbose off and reduced the locales, it
> might be worth still letting it run N=12. The original idea with 12 is 
> that
> it increases the chance of running into daylight saving on and off cases.
>
> -Sherman
>
> On 08/13/2013 12:06 PM, roger riggs wrote:
>> Hi,
>>
>> The tests for java.util.Formatter were failing in the th_TH locale.
>> The test incorrectly relied on the locale defined calendar 
>> instantiation.
>> Java.util.formatter should be using YEAR_OF_ERA to extract the year
>> to get consistent results with the Japanese Calendar, for the other
>> calendars it does not change the values.
>> The tests are extended to explicitly test the Gregorian, Japanese, and
>> ThaiBuddist calendar formatting using both the Java.util.Calendar and
>> ChronoZonedDateTime calendars.
>>
>> Webrev:
>> http://cr.openjdk.java.net/~rriggs/webrev-fixed-locale-8022193/
>>
>> Thanks, Roger
>>
>