the other fix added is to make sure that initialize() IS run by every
thread, so that the synchronize is encountered, not escaped in the case
that the array being constructed is null. That does the job.
Bob
.
Miguel wrote:
>Nico wrote:
>
>
>>>It really looks like a JVM bug to me ... but that seems
>>>impossible. I will investigate.
>>>
>>>
>>>
>>No need, it was just because initialize() is not static.
>>
>>
>
>Correct! It occurred to me last nite that this was probably the problem.
>
>The method signature needs to be:
>
> synchronized static initialize() {
> }
>
>
>If initialize() is not 'static' then the synchronization occurs on the
>object, not on the class.
>
>
>Miguel
>
>
>
>_______________________________________________
>Jmol-developers mailing list
>[email protected]
>https://lists.sourceforge.net/lists/listinfo/jmol-developers
>
>
_______________________________________________
Jmol-developers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/jmol-developers
