RFR JDK 8 javac lint cleanup of java.lang.ref

[Joseph Darcy]

An update, I played around with the declarations a bit more, but wasn't 
about to find something workable so I pushed the already-reviewed version.

If someone else wants to take a crack at improving the generics, I think 
that would be a fine refactoring.

Thanks,

-Joe

On 7/25/2013 2:16 PM, Joe Darcy wrote:
> On 07/25/2013 01:58 PM, Alan Bateman wrote:
>> On 25/07/2013 13:33, Joe Darcy wrote:
>>> Hello,
>>>
>>> Please review these changes to remove the javac lint warnings from 
>>> the java.lang.ref package:
>>>
>>>     8021429 Fix lint warnings in java.lang.ref
>>>     http://cr.openjdk.java.net/~darcy/8021429.0/
>>>
>>> Care was taken to not change any signatures of public API elements. 
>>> After doing a clean build, all the java.lang.ref regression tests 
>>> pass. Full patch below.
>>>
>>> Thanks,
>>>
>>> -Joe
>> This looks okay to me, just wondering Reference.next needs to be a 
>> raw type (you may have tried possible solutions to this already).
>
> After some amount of examination, I didn't find a way to fix this. The 
> description of the protocol around that field is
>
>     /* When active:   NULL
>      *     pending:   this
>      *    Enqueued:   next reference in queue (or this if last)
>      *    Inactive:   this
>      */
>
> None of
>
>     Reference<Object>
>     Reference<?>
>     Reference<T>
>     Reference<? extends T>
>
> work as the declared type of the field given other uses of the field 
> in the package.
>
>>
>> Minor comment on Finalizer is that the comment on the class being 
>> package-private is now misaligned.
>>
>>
>
> I'll fix that before I push.
>
> Thanks all for the quick reviews.
>
> -Joe
>