RFR: Solaris 64-bit images

[Chris Hegarty]

On 21/05/12 19:26, Kelly O'Hair wrote:
> I think tests that want the 64bit version, should use the regular bin/java -d64   not dive into the bin/sparcv9/ or bin/amd64 directories.
> Just my opinion.

Right, this is exactly what I was thinking. I really don't understand 
why anyone would be executing bin/[sparcv9|amd64]/java directly.

Do these tests predate dual launcher support? Or any other historic 
reason it could be? Anyway, for now I'll just add them to the modules 
problem list and we can revisit when a more permanent solution for the 
images is found.

-Chris.

>
> -kto
>
> On May 21, 2012, at 7:56 AM, Chris Hegarty wrote:
>
>> On 21/05/2012 15:43, Chris Hegarty wrote:
>>> ....
>>>> Chris - I looked at the changes and they look okay to me. I also grabbed
>>>> the patch and it built a solaris-x64 image as I expected. I'm happy to
>>>> push the changes for you. One small thing is that we have a couple of
>>>> tests (launcher and JDI tests come to mind) that launch via amd64/bin or
>>>> sparcv9/bin. We might need to add them to the ModulesProblemList.txt.
>>
>> I wonder why these tests invoke the 64-bit launcher directly? I notice most of them use the Process API to launch the java process. Is there an issue with launching with 'bin/java -d64' (and using the dual mode support of the launcher, if required), and redirecting the output or waiting for the process? Otherwise, I don't see why the test don't simply launch 'bin/java'.
>>
>> I guess what I'm asking is if these tests should be changed (to use bin/java) in the jdk8 mainline, rather than added to the problem list ( could be added temporarily of course)?
>>
>> -Chris.
>>
>>>
>>> Oh, I didn't notice these tests. Let me look into these and for now I'll
>>> add them to the ModulesProblemList.txt.
>>>
>>> Thanks,
>>> -Chris.
>>>
>>>>
>>>> -Alan.
>