JDK 9 build 109 -> Lucene's Ant build works again; still missing Hotspot patches

Mandy Chung mandy.chung at oracle.com
Thu Mar 17 18:35:44 UTC 2016


Some but not all com.sun.* API that are unsupported.  com.sun.management is an exported API.

Can you try jdk-9+109?  It may be a regression in jdk-9+110.  I suspect it’s a build bug resurrected - the build does some magic to combine the META-INF/services config file from multiple module and looks like jdk-9+110 includes an entry that should have been excluded.  Such build magic is just an interim solution until the module system is integrated.

I tried jake b110.  This problem does not exist as expected.

Mandy

> On Mar 17, 2016, at 11:23 AM, Remi Forax <forax at univ-mlv.fr> wrote:
> 
> It's not a bug, it's a feature :)
> com.sun.management.HotSpotDiagnosticMXBean is a com.sun classes, so maybe not available anymore.
> 
> see https://wiki.openjdk.java.net/display/JDK8/Java+Dependency+Analysis+Tool
> 
> Rémi
> 
> ----- Mail original -----
>> De: "Uwe Schindler" <uschindler at apache.org>
>> À: "Core-Libs-Dev" <core-libs-dev at openjdk.java.net>, "Rory O'Donnell" <rory.odonnell at oracle.com>
>> Envoyé: Jeudi 17 Mars 2016 17:03:25
>> Objet: RE: JDK 9 build 109 -> Lucene's Ant build works again;	still missing Hotspot patches
>> 
>> Hi,
>> 
>> In fact, any static method on ManagementFactory fails with this trace. You
>> cannot get any bean, not even runtime bean!
>> 
>> Uwe
>> 
>> -----
>> Uwe Schindler
>> uschindler at apache.org
>> ASF Member, Apache Lucene PMC / Committer
>> Bremen, Germany
>> http://lucene.apache.org/
>> 
>> 
>>> -----Original Message-----
>>> From: core-libs-dev [mailto:core-libs-dev-bounces at openjdk.java.net] On
>>> Behalf Of Uwe Schindler
>>> Sent: Thursday, March 17, 2016 4:40 PM
>>> To: 'Core-Libs-Dev' <core-libs-dev at openjdk.java.net>; 'Rory O'Donnell'
>>> <rory.odonnell at oracle.com>
>>> Subject: RE: JDK 9 build 109 -> Lucene's Ant build works again; still
>>> missing
>>> Hotspot patches
>>> 
>>> Hi,
>>> 
>>> Unfortunately we found a bug in build 110 - not as bad as the Ant one, but
>>> the Gradle build system does no longer start! :(
>>> 
>>> The following fails:
>>> 
>>> C:\Users\Uwe Schindler\workspace-lucene\tester\src>cat Test.java
>>> import java.lang.management.ManagementFactory;
>>> import com.sun.management.HotSpotDiagnosticMXBean;
>>> 
>>> public final class Test {
>>> 
>>> public static void main(String... args) throws Throwable {
>>> 
>>> ManagementFactory.getPlatformMXBean(HotSpotDiagnosticMXBean.class);
>>> }
>>> 
>>> }
>>> 
>>> C:\Users\Uwe Schindler\workspace-lucene\tester\src>javac Test.java
>>> 
>>> C:\Users\Uwe Schindler\workspace-lucene\tester\src>java Test
>>> Exception in thread "main" java.util.ServiceConfigurationError:
>>> sun.management.spi.PlatformMBeanProvider: Provider
>>> jdk.management.cmm.i
>>> nternal.PlatformMBeanProviderImpl not found
>>>       at java.util.ServiceLoader.fail(ServiceLoader.java:237)
>>>       at java.util.ServiceLoader.access$300(ServiceLoader.java:183)
>>>       at
>>> java.util.ServiceLoader$LazyIterator.nextService(ServiceLoader.java:370)
>>>       at java.util.ServiceLoader$LazyIterator.next(ServiceLoader.java:402)
>>>       at java.util.ServiceLoader$1.next(ServiceLoader.java:478)
>>>       at java.lang.Iterable.forEach(Iterable.java:74)
>>>       at
>>> java.lang.management.ManagementFactory$PlatformMBeanFinder.lambda
>>> $static$0(ManagementFactory.java:890)
>>>       at java.security.AccessController.doPrivileged(Native Method)
>>>       at
>>>       java.security.AccessController.doPrivileged(AccessController.java:428)
>>>       at
>>> java.lang.management.ManagementFactory$PlatformMBeanFinder.<clinit>(
>>> ManagementFactory.java:886)
>>>       at
>>> java.lang.management.ManagementFactory.getPlatformMXBean(Managem
>>> entFactory.java:649)
>>>       at Test.main(Test.java:8)
>>> 
>>> 
>>> Should I open an issue or can we do this through this mail?
>>> 
>>> This breaks several things:
>>> 
>>> - Grade fails to start
>>> - One Lucene test, because it detects Oracle and 64 bits, but then finds
>>> out
>>> that we have no Hotspot bean
>>> - The Lucene Test Runner cannot get the PID of a Process started
>>> - Elasticsearch cannot find its own process ID because of this
>>> 
>>> Uwe
>>> 
>>> -----
>>> Uwe Schindler
>>> uschindler at apache.org
>>> ASF Member, Apache Lucene PMC / Committer
>>> Bremen, Germany
>>> http://lucene.apache.org/
>>> 
>>> 
>>>> -----Original Message-----
>>>> From: Robert Muir [mailto:rcmuir at gmail.com]
>>>> Sent: Thursday, March 17, 2016 4:01 PM
>>>> To: Uwe Schindler <uschindler at apache.org>
>>>> Cc: Rory O'Donnell <rory.odonnell at oracle.com>; Core-Libs-Dev <core-libs-
>>>> dev at openjdk.java.net>; Steve Drach <steve.drach at oracle.com>; Alan
>>>> Bateman <Alan.Bateman at oracle.com>; hotspot compiler <hotspot-
>>>> compiler-dev at openjdk.java.net>
>>>> Subject: Re: JDK 9 build 109 -> Lucene's Ant build works again; still
>>>> missing
>>>> Hotspot patches
>>>> 
>>>> On Thu, Mar 17, 2016 at 10:25 AM, Uwe Schindler
>>> <uschindler at apache.org>
>>>> wrote:
>>>>> 
>>>>> My local tests showed that the MethodHandle-bug is solved, the other
>>> one
>>>> is hopefully fixed, too. Robert may have a way to quickly reproduce.
>>>>> 
>>>> 
>>>> JDK-8150280 is fixed too, I just tested it. Thanks!
>> 
>> 




More information about the core-libs-dev mailing list