<i18n dev> RFR: 8051641: Africa/Casablanca transitions is incorrectly calculated starting from 2027

Aleksej Efimov aleksej.efimov at oracle.com
Mon Dec 15 16:22:44 UTC 2014


Sherman, thank you for looking into this problem and helping solve it. I 
will change the formatting before the push.

-Aleksej

On 12/15/2014 07:17 PM, Xueming Shen wrote:
> On 12/15/14 8:06 AM, Aleksej Efimov wrote:
>> Hello Sherman, Masayoshi and other experts,
>>
>> Can I ask for a JDK9 review for the 
>> sun/util/calendar/zi/TestZoneInfo310.java test failure [1] that was 
>> caused by incorrect handling of last rules for Africa/Casablanca time 
>> zone. This timzone has two last rules and three instant rules in a 
>> range of these last rules that causes incorrect instant transitions 
>> table generation. The proposed fix [2] solves this problem. Also 
>> there is a modification to test class that solves the issue with not 
>> optimized instant transitions with the same offset/saving values that 
>> are the last two in instant transitions table. And 
>> TestZoneInfo310.java is removed from problems list.
>>
>> Testing:
>> All tz related test were executed with no failures.
>> JPRT build/testing successfull for all platforms.
>>
>> Best Regards,
>> Aleksej
>>
>> [1] https://bugs.openjdk.java.net/browse/JDK-8051641
>> [2] http://cr.openjdk.java.net/~aefimov/8051641/9/webrev.00/
>>
> Hi Aleksej,
>
> The fix itself is good. The lines in Rules.java need adjust the space 
> and blank
> line at #736 at provider.java need to be removed as well (my fault).
>
> Thanks!
> -Sherman



More information about the i18n-dev mailing list