code review for 6949515 (round 0)

[Joe Darcy]

Hi Dan.

Thanks for the fix.  I think this would be a fine change for OpenJDK 6; 
in terms of sequencing, I think it should go back after Andrew's HS17 
merge into OpenJDK 6.

-Joe

Daniel D. Daugherty wrote:
> Joe,
>
> This is a crasher that came in with the HSX-16 update. I think you
> probably want this fix.
>
> Here is the OpenJDK webrev:
>
> http://cr.openjdk.java.net/~dcubed/6949515-webrev/0/
>
> I don't know why the bugs.sun.com link doesn't show anything yet...
>
> I've had three internal reviewers.
>
> Dan
>
>
>
>
> On 5/12/2010 10:29 AM, daniel.daugherty at oracle.com wrote:
>> Greetings,
>>
>> The jmethodID performance fix that we did back in August 2009 has
>> a small problem with GC'ed jmethodIDs. Tomas Hurka reported the
>> crash and was kind enough to test my proposed fix. Strangely
>> enough, we don't have an existing test that runs into this problem.
>>
>> Here is the webrev URL:
>>
>>     http://javaweb.sfbay/~dcubed/6949515-webrev/0/
>>
>> Here is the webrev URL for the original jmethodID performance fix:
>>
>>     http://javaweb.sfbay/~dcubed/6862945-webrev/1/
>>
>> I'm planning to push this fix to OpenJDK6 (HSX-16) and OpenJDK7 
>> (HSX-18).
>>
>> Thanks, in advance, for any comments.
>>
>> Dan
>>