RFR 8043138: Attach API should not require jvmstat rmi protocol

[Mandy Chung]

> On Nov 26, 2015, at 2:12 AM, Jaroslav Bachorik <jaroslav.bachorik at oracle.com> wrote:
> 
> Mandy,
> 
> On 25.11.2015 18:22, Mandy Chung wrote:
>> 
>>> On Nov 25, 2015, at 8:53 AM, Jaroslav Bachorik <jaroslav.bachorik at oracle.com> wrote:
>>> 
>>> Updated webrevs:
>>> 
>>> http://cr.openjdk.java.net/~jbachorik/8043138/webrev.01/top
>>> http://cr.openjdk.java.net/~jbachorik/8043138/webrev.01/jdk
>>> 
>> 
>> This looks fine to me.  Is there any test referencing the packages in jdk.jvmstat.rmi?   I only found @modules with
>>    "jdk.jvmstat/sun.jvmstat.monitor”
>>    "jdk.jvmstat/sun.jvmstat.montior.event”
>> 
>> It’d be good to double check and adjust the module name if any.
> 
> The only place jstatd (the only usage of the RMI MonitoredHost provider) is exercised is in jdk/test/sun/tools/jstatd - and it is all done via executing jstatd as an external process so it doesn't require putting jdk.jvmstat.rmi in @modules, IMO.
> 

It’s good.

For test selection, @modules jdk.jvmstat.rmi would be needed in these tests such that when testing an image with a subset of JDK modules, it can select the right set of tests to execute.   This is something to do in the future that Shura and the team will look at.

Mandy