RFR: Backport of JDK-8044135: Add API to start JMX agent from attach framework

[Staffan Larsen]

On 4 sep 2014, at 14:09, Dmitry Samersoff <dmitry.samersoff at oracle.com> wrote:

> Staffan,
> 
> On 2014-09-04 16:02, Staffan Larsen wrote:
>> 
>> On 4 sep 2014, at 13:52, Dmitry Samersoff <dmitry.samersoff at oracle.com> wrote:
>> 
>>> Staffan,
>>> 
>>> WindowsVirtualMachine.java:109 Should we close PipedInputStream ?
>> 
>> There is a catch-statement that does this for IOExceptions on line 125. Since AttachOperationFailedException is an IOException it the pipe will get closed there. It really should be a finally (or try-with-resources) block instead.
> 
> At ll. 125 we are closing pipe:
> 
>   closePipe(hPipe);
> 
> But It might be necessary to close PipedInputStream as well.

Ah, right - I misunderstood what you meant. Yes, the PipedInputStream should be closed before we throw an exception. Since this is not something I introduced with this change I have filed a bug for it and will deal with it separately: https://bugs.openjdk.java.net/browse/JDK-8057558

Thanks,
/Staffan

> 
> -Dmitry
> 
> 
>> 
>> Thanks,
>> /Staffan
>> 
>>> 
>>> Otherwise looks good.
>>> 
>>> -Dmitry
>>> 
>>> On 2014-09-04 14:11, Staffan Larsen wrote:
>>>> Can I interest any Reviewers in taking a look at this?
>>>> 
>>>> Thanks,
>>>> /Staffan
>>>> 
>>>> On 3 sep 2014, at 14:27, Staffan Larsen <staffan.larsen at oracle.com> wrote:
>>>> 
>>>>> …aaaand the link: http://cr.openjdk.java.net/~sla/8044398-8039173-8044135-jdk8u/webrev.00/
>>>>> 
>>>>> /Staffan
>>>>> 
>>>>> On 3 sep 2014, at 14:25, Staffan Larsen <staffan.larsen at oracle.com> wrote:
>>>>> 
>>>>>> This is a review for a backport of JDK-8044135 to jdk8u. This fix had a few dependencies that also needed backporting, so included in this review are the following fixes from JDK9:
>>>>>> 
>>>>>> • 8044135 Add API to start JMX agent from attach framework (jdk)
>>>>>> • 8039173 Propagate errors from Diagnostic Commands as exceptions in the attach framework (jdk)
>>>>>> • 8044398 Attach code should propagate errors in Diagnostic Commands as errors  (hotspot)
>>>>>> 
>>>>>> The only significant change from the version in JDK 9 is that startManagementAgent() and startLocalManagementAgent() in VirtualMachine.java have "@since 1.8" (instead of 1.9).
>>>>>> 
>>>>>> After speaking with the maintainers and gatekeepers, I plan to push this to the jdk8u-hs-dev repository.
>>>>>> 
>>>>>> Thanks,
>>>>>> /Staffan
>>>>>> 
>>>>> 
>>>> 
>>> 
>>> 
>>> -- 
>>> Dmitry Samersoff
>>> Oracle Java development team, Saint Petersburg, Russia
>>> * I would love to change the world, but they won't give me the sources.
>> 
> 
> 
> -- 
> Dmitry Samersoff
> Oracle Java development team, Saint Petersburg, Russia
> * I would love to change the world, but they won't give me the sources.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/serviceability-dev/attachments/20140904/60c64c4c/attachment.html>