[15] RFR: 8234347: "Turkey" meta time zone does not generate composed localized names, 8236548: Localized time zone name inconsistency between English and other locales

Roger Riggs Roger.Riggs at oracle.com
Wed Feb 12 14:44:52 UTC 2020


+1

On 2/11/20 6:12 PM, Joe Wang wrote:
> +1. That's nicer.
>
> -Joe
>
> On 2/11/20 10:17 AM, naoto.sato at oracle.com wrote:
>> Hi,
>>
>> I modified the proposed changeset. Removed the hard coded 3-letter id 
>> compatibility mappings (oldMappings) from CLDRConverter.java. 
>> Instead, using public ZoneId.SHORT_IDS that contain the same set of 
>> mappings (wasn't aware it was defined in the spec!)
>>
>> Here is the updated webrev:
>>
>> http://cr.openjdk.java.net/~naoto/8234347.8236548/webrev.01/
>>
>> Naoto
>>
>> On 2/7/20 1:44 PM, naoto.sato at oracle.com wrote:
>>> Hello,
>>>
>>> Please review the fix to the following issues:
>>>
>>> https://bugs.openjdk.java.net/browse/JDK-8234347
>>> https://bugs.openjdk.java.net/browse/JDK-8236548
>>>
>>> The proposed changeset is located at:
>>>
>>> http://cr.openjdk.java.net/~naoto/8234347.8236548/webrev.00/
>>>
>>> This changeset includes the following changes:
>>>
>>> - English time zone names missing in CLDR source files are no longer 
>>> copied from COMPAT provider at build time. Rather it is synthesized 
>>> at runtime, which has been the way other locales did.
>>>
>>> - Runtime CLDR time zone name provider fallback logic has been 
>>> refined. It now falls back to parent locales per each missing name, 
>>> instead of resource bundle. Also, region fall back is now using 
>>> exemplar city to synthesize the name (e.g., "Turkey" meta zone)
>>>
>>> - Minor fix in DateTimeFormatterBuilder on zone text parsing. It now 
>>> parses zone texts that start with "UTC", yet it is ZoneId.
>>>
>>> Naoto
>



More information about the core-libs-dev mailing list