Code review request: 7086586 Inference producing null type argument
Jonathan Gibbons
jonathan.gibbons at oracle.com
Thu Oct 6 06:36:34 PDT 2011
On 10/06/2011 06:31 AM, Rémi Forax wrote:
> On 10/06/2011 02:34 PM, Maurizio Cimadamore wrote:
>> On 06/10/11 10:47, Maurizio Cimadamore wrote:
>>> On 06/10/11 10:38, Dmeetry Degrave wrote:
>>>>
>>>> hi,
>>>>
>>>> i'm looking for a code review for 7086586 fix for 7u4. It
>>>> implements fix suggested by Maurizio for problem 2 from evaluation.
>>>>
>>>> bug description: http://bugs.sun.com/view_bug.do?bug_id=7086586
>>>> jdk7 evaluation: http://bugs.sun.com/view_bug.do?bug_id=2213800
>>>>
>>>> webrev: http://cr.openjdk.java.net/~dmeetry/7086586/webrev.00/
>>>>
>>>> thanks,
>>>> dmeetry
>>> Approved
>>>
>>> Maurizio
>> Remi has suggested to use 'static final' on botFilter. I'd be ok with
>> final, less ok with 'static' if no other reason than for
>> consistencies with similar declarations in Infer.java. Remi what do
>> you think?
>>
>> Maurizio
>
> I think that all closures that doesn't capture any states should be
> declared static.
> Also a VM may be able to better optimize a static field if Type.filter
> is inlined.
>
> Rémi
>
>
>
>
It would be an interesting annotation processor for someone to write to
analyze code to see which non-static members could plausibly be made static.
-- Jon
More information about the compiler-dev
mailing list