Review request: JDK-7162111 TEST_BUG: change tests run in headless mode [macosx] (open part) (original) (raw)

Alexey Utkin alexey.utkin at oracle.com
Mon Nov 19 09:57:48 UTC 2012


Bug description: https://jbs.oracle.com/bugs/browse/JDK-7162111

Here is the suggested fix: http://cr.openjdk.java.net/~uta/openjdk-webrevs/JDK-7162111/

Summary:

Tests that are not changed and pass successfully (all tests run for jdk8):

java/io/Serializable/serialver/classpath/run.sh java/io/Serializable/serialver/nested/run.sh java/util/ResourceBundle/Control/Bug6530694.java javax/script/CauseExceptionTest.java javax/script/GetInterfaceTest.java javax/script/JavaScriptScopeTest.java javax/script/NullUndefinedVarTest.java javax/script/PluggableContextTest.java javax/script/ProviderTest.sh javax/script/RhinoExceptionTest.java javax/script/StringWriterPrintTest.java javax/script/Test1.java javax/script/Test2.java javax/script/Test3.java javax/script/Test4.java javax/script/Test5.java javax/script/Test6.java javax/script/Test7.java javax/script/Test8.java javax/script/UnescapedBracketRegExTest.java javax/script/VersionTest.java sun/tools/jrunscript/jrunscript-argsTest.sh sun/tools/jrunscript/jrunscript-cpTest.sh sun/tools/jrunscript/jrunscript-DTest.sh sun/tools/jrunscript/jrunscript-eTest.sh sun/tools/jrunscript/jrunscript-fTest.sh sun/tools/jrunscript/jrunscriptTest.sh

demo/jvmti/mtrace/TraceJFrame.java:

Inapplicable in headless mode: JFrame creation time is under investigation. Test was changed to pass successfully if headless mode was detected.

java/io/Serializable/resolveClass/deserializeButton/run.sh:

The file [java/io/Serializable/resolveClass/deserializeButton/Foo.java] was changed. The pure Vector collection substitutes the AWT Button class in mixed ClassLoader objects serialization. The problem verification procedure was conserved.

The switch "-Djava.awt.headless=true" is useless in all CoreLib tests. AWT uses the property to force running in headless mode. There are two cases:

If the test fails with headless exception that is a good signal: test in not a pure CoreLib algorithm and code reach GUI initialization stage. Borderline case is the testing scripts. Script engine can preliminary instantiate GUI sub-system, but for that case inability to execute GUI-less scenario could be interpreted as bug.

Currently all mentioned tests with suggested changes pass on Mac OS in ssh session from Windows machine (no chance for X11).

Regards, -uta



More information about the core-libs-dev mailing list