OpenJDK 6 b19 features?

[Andrew John Hughes]

On 19 February 2010 19:24, Joseph D. Darcy <Joe.Darcy at sun.com> wrote:
> Andrew John Hughes wrote:
>>
>> On 19 February 2010 18:36, Joseph D. Darcy <Joe.Darcy at sun.com> wrote:
>>
>>>
>>> Andrew John Hughes wrote:
>>>
>>>>
>>>> On 18 February 2010 08:27, Gary Benson <gbenson at redhat.com> wrote:
>>>>
>>>>
>>>>>
>>>>> Andrew John Hughes wrote:
>>>>>
>>>>>
>>>>>>
>>>>>> On 17 February 2010 17:55, Joseph D. Darcy <Joe.Darcy at sun.com> wrote:
>>>>>>
>>>>>>
>>>>>>>
>>>>>>> Andrew has already started work on updating HotSpot.
>>>>>>>
>>>>>>>
>>>>>>
>>>>>> This will be ready to push as soon as the issue with Daniel's patch
>>>>>> is resolved.
>>>>>>
>>>>>>
>>>>>
>>>>> Does this updated HotSpot have Zero in it?
>>>>>
>>>>>
>>>>>
>>>>
>>>> No, the patches were applied to hs17.
>>>>
>>>> We could backport it though.  With hs16, we'll be using what is now
>>>> the alternate HotSpot build for IcedTea6 that is closer to the version
>>>> against which Zero was applied upstream.
>>>>
>>>> Joe, what do you think about making Zero available upstream in
>>>> OpenJDK6?  It already ships with IcedTea6 and builds have passed the
>>>> TCK.
>>>>
>>>>
>>>
>>> Hmm.  I'm not sure.  What would making Zero available upstream mean
>>> exactly
>>> for OpenJDK 6?  I'm inferring Zero in is sufficiently new versions of
>>> HotSpot (hs17 vs hs16) given the presence of a src/cpu/zero directory in
>>> the
>>> JDK 7 HotSpot repo.  Would the rough proposal be to port the Zero fixes
>>> applied to hs17 in JDK 7 to the hs16 in OpenJDK 6?
>>>
>>>
>>
>> It would.  In fact, there would be no work to speak of as IcedTea6
>> already supports applying it to both hs14 and hs16 (the latter being
>> much less work).  Having it upstream would simplify the IcedTea build
>> a little and possibly make it available to more people.
>>
>> I'm sure Gary (Zero's main author) can also make some other good
>> points for its inclusion... :-)
>>
>
> What would be the impact of putting Zero in OpenJDK 6 b19 with any future
> update of OpenJDK 6 to HotSpot 17, just doing work now versus later?
>
> -Joe
>

Past behaviour suggests that HotSpot 17 won't be stabilised; the
HotSpot engineers seem to stabilise even numbered versions in a
separate branch.  It's likely to be a while before hs18 appears as a
stable branch and is ready for production use.

The backported changesets would be the same as those applied to 17 so
they should just merge cleanly when we do bump to hs18.  It works with
the IcedTea forest, which often has changesets before they appear in
JDK7 build drops, Zero being one of them.

The vast majority of people using OpenJDK6 do so via IcedTea6 so I
don't think it will make as big a difference as it may appear, but it
will simplify the maintenance of Zero significantly.
-- 
Andrew :-)

Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

Support Free Java!
Contribute to GNU Classpath and the OpenJDK
http://www.gnu.org/software/classpath
http://openjdk.java.net

PGP Key: 94EFD9D8 (http://subkeys.pgp.net)
Fingerprint: F8EF F1EA 401E 2E60 15FA  7927 142C 2591 94EF D9D8