[13] RFR(S) 8208379: compiler/jvmci/events/JvmciNotifyInstallEventTest.java failed with "Got unexpected event count after 2nd install attempt: expected 9 to equal 2"

Vladimir Kozlov vladimir.kozlov at oracle.com
Thu Jun 6 22:01:16 UTC 2019


On 6/6/19 1:53 PM, Igor Ignatyev wrote:
> Hi Vladimir,
> 
> sounds reasonable. I have one question though. as far as I can see, 8225019 just added "-Djvmci.Compiler=graal" to L#50:
>> - *     -Xbootclasspath/a:. -Xmixed
>> + *     -Djvmci.Compiler=graal -Xbootclasspath/a:. -Xmixed
> 
> why are you removing the whole @run?

By default without explicit -Djvmci.Compiler= JVMCI will select Graal and we will have the same issue as this bug.

Thanks,
Vladimir

> 
> Thanks
> -- Igor
> 
>> On Jun 6, 2019, at 1:14 PM, Vladimir Kozlov <vladimir.kozlov at oracle.com> wrote:
>>
>> http://cr.openjdk.java.net/~kvn/8208379/webrev.00/
>> https://bugs.openjdk.java.net/browse/JDK-8208379
>>
>> Two JVMCI tests use own EmptyCompiler test compiler instead of Graal to control produced JVMCI events.
>> We should not run these tests with Graal.
>>
>> Note,  I removed the run with Graal in JvmciNotifyInstallEventTest.java test which I incorrectly added in my changes for JDK-8225019.
> 
>>
>> Thanks,
>> Vladimir
> 


More information about the hotspot-compiler-dev mailing list