RFR: JMC-6118 Java Mission Control tests fail on OpenJDK8

Guru guru.hb at oracle.com
Tue Nov 6 21:10:13 UTC 2018


Hi Josh,

Changes looks good to me, But Please Note, We can’t have multiple commit on a single JBS ID. I have raised https://bugs.openjdk.java.net/browse/JMC-6194 <https://bugs.openjdk.java.net/browse/JMC-6194> as a follow on for this one.

Do update the description of JMC-6194 (if its needs correction). 

To make our this uiTest succeeded (along with our compilation) Can we Introduce JDK11_HOME or JDK_HOME which points to OpenJDK_11 and then while running uiTest we use the JDK_HOME in the tycho-surfire (I have not evaluated , its just a thought). 
OpenJFX had this similar idea of defining JAVA_HOME for running tests and JDK_HOME for Compiling (it’s based on gradle) and we can try for our JMC (with maven). 

Thanks,
Guru
> On 06-Nov-2018, at 12:58 AM, Joshua Matsuoka <jmatsuok at redhat.com> wrote:
> 
> Hi,
> 
> The following patch adds four additional tests to the list of excluded
> tests on OpenJDK8:
> 
> ControlRecordingsTest - Attempts to perform various operations on a
> recording from the JVM Browser, fails since we don't have flight recorder
> functionality in OpenJDK8
> 
> JfrWizardTest - Attempts to test various options in the JFR Recording
> wizard. The wizard is started from the JVM Browser, so it fails for the
> same reasons as the previous test.
> 
> TemplateManagerTest - Requires the JVM Browser to have JFR functionality,
> same as above.
> 
> DiagnosticCommandsTabTest - Expects JFR diagnostic commands to be present
> in the list of options. They aren't there for OpenJDK8.
> 
> This allows mvn clean verify -P uitests to be run successfully on OpenJDK8.
> 
> http://cr.openjdk.java.net/~jmatsuoka/JMC-6118/webrev.01/
> 
> Thoughts?
> 
> Cheers,
> 
> - Josh



More information about the jmc-dev mailing list