RFR 8158855: Fix remaining module dependences in java/lang

[Alexandre (Shura) Iline]

Hi.

After some offline conversations it was decided to leave the tests which depend on absence of “-limitmods” in VM options for later. I have created a separate bug ti track that:
https://bugs.openjdk.java.net/browse/JDK-8159523

What left for this review is this, then:
http://cr.openjdk.java.net/~shurailine/8158855/webrev.01/

Please take a look on it.

Next problems are left unfixed in the java/lang tests:
java/lang/invoke/modules/ModuleAccessControlTest.java, java/lang/reflect/Proxy/ProxyClassAccessTest.java ,java/lang/reflect/Proxy/ProxyTest.java JDK-8159523
java/lang/SecurityManager/CheckSecurityProvider.java JDK-8158670
java/lang/String/concat/WithSecurityManager.java JDK-8158851
java/lang/management/MemoryMXBean/Pending.java JDK-8158837
java/lang/reflect/OldenCompilingWithDefaults.java CODETOOLS-7901694
java/lang/StackWalker/TestBCI.java, java/lang/invoke/lambda/LambdaAsm.java CODETOOLS-7901678

Shura


> On Jun 8, 2016, at 5:50 PM, Mandy Chung <mandy.chung at oracle.com> wrote:
> 
> Hi Shura,
> 
> test/java/lang/Class/GetModuleTest.java
>   This can use java.base/jdk.internal.misc.Unsafe instead.
> 
> Other @modules looks okay.
> 
> Regarding the executeTestJava change, if you drop the VM options, this test will only exercise the default launcher setting.  The hotspot nightlies depend on VM options to exercise different configurations.  I think many tests that launch a modular test will run into the same issue.  It’s not ideal to drop the VM options then.  One alternative is to have a new ProcessTool::executeModularTest that will take test.vm.opts and test.java.opts but dropping -limitmods, if any.
> 
> Mandy
>