<i18n dev> testcase failure java/util/Locale/Bug6989440.java

[David Holmes]

This might not be related to the CME problem, but here:

     public static LocaleServiceProviderPool getPool(Class<? extends 
LocaleServiceProvider> providerClass) {
         LocaleServiceProviderPool pool = poolOfPools.get(providerClass);
         if (pool == null) {
             LocaleServiceProviderPool newPool =
                 new LocaleServiceProviderPool(providerClass);
             pool = poolOfPools.put(providerClass, newPool);
             if (pool == null) {
                 pool = newPool;
             }
         }

         return pool;
     }

we should probably be using poolOfPools.putIfAbsent(providerClass, newPool)

David

On 6/10/2011 3:35 AM, Naoto Sato wrote:
> I will look into this. Reopened the original CR.
>
> Naoto
>
> On 10/5/11 9:58 AM, Alan Bateman wrote:
>> Chris Hegarty wrote:
>>> Alan, Naoto, David
>>>
>>> I filed CR 7098100: java/util/Locale/Bug6989440.java fails
>>> intermittently.
>>>
>>> If you're ok with it please review the patch (below) and I can push it
>>> to the tl repo. Job done!
>> I assume there is also some underlying issue in the Locale code and this
>> might get hidden if we fix the test (I"ve no objection to fixing the
>> test of course, just thinking that we should study the Locale code to
>> try to identify the deadlock or hang or whatever it is that is causing
>> the threads in this test not to terminate).
>>
>> -Alan
>