RFR: 8054823: Add size_t as a valid VM flag type

[Staffan Larsen]

About the new AttachSetGetFlag test:

- To get the pid of the newly launched process use the new Process.getPid() API instead. This is also available cross platform and you can remove the only non-windows restriction.

- To avoid the annoying and unstable Thread.sleep(5000) let the started program print that it is ready and wait for that string in the output. This also avoids the need for -XX:+PauseAtStartup. 

- To quit the started program use Process.destroyForcibly() in addition to having the process end by itself after a long timeout.

Thanks,
/Staffan

On 11 aug 2014, at 19:46, Stefan Karlsson <stefan.karlsson at oracle.com> wrote:

> On 2014-08-11 18:35, Vladimir Kozlov wrote:
>> Stefan,
>> 
>> Seems fine to me,
> 
> Thanks for reviewing!
> 
>> except new tests don't have correct @bug.
>> 
>> 27  * The test helsp verifying that size_t flags can be set/read.
>>               ^ typo
>> 
>> SizeTTest.java - remove @author
> 
> Too much copy-n-paste. I've updated the patch and did some additional cleanups to the jtreg annotations:
> http://cr.openjdk.java.net/~stefank/8054823/webrev.01.delta/
> http://cr.openjdk.java.net/~stefank/8054823/webrev.01/
> 
> thanks,
> StefanK
> 
>> 
>> Thanks,
>> Vladimir
>> 
>> On 8/11/14 6:48 AM, Stefan Karlsson wrote:
>>> Hi all,
>>> 
>>> I propose that we add 'size_t' as a valid VM flag type. This will help with type safety, since we won't have to cast
>>> back and forth between the two types.
>>> 
>>> This patch adds the framework, but only changes one of the flags. I choose to change one of our experimental flags, so
>>> that I could write tests for this change.
>>> 
>>> Please, review this patch:
>>>  http://cr.openjdk.java.net/~stefank/8054823/webrev.00/
>>>  https://bugs.openjdk.java.net/browse/JDK-8054823
>>> 
>>> thanks,
>>> StefanK
>