The key name in `java.test.config` should align with the launch configuraiton

[astrelsky]

I have the following defined in my workspace /.vscode/settings.json

"java.test.config": {
        "name": "GhidraTest",
        "vmArgs": ["-Djava.system.class.loader=ghidra.GhidraClassLoader", "-Declipse.project.dir=${workspaceFolder}", "-Dpython.console.encoding=UTF-8", "-Dfile.encoding=UTF8", "-Dlog4j.skipJansi=true", "-Duser.country=US", "-Duser.language=en", "-Dsun.stdout.encoding=UTF-8", "-Dsun.err.encoding=UTF-8", "-Dvisualvm.display.name=Ghidra"]
    }

Upon running a test and explicitly selecting the GhidraTest configuration option the following output is produced. Note the incorrect vmargs. (If args is set they also get ignored.)

Launching with the following launch configuration: '{
  "name": "Launch Java Tests",
  "type": "java",
  "request": "launch",
  "mainClass": "org.eclipse.jdt.internal.junit.runner.RemoteTestRunner",
  "projectName": "Ghidra-Cpp-Class-Analyzer",
  "classPaths": [
    "C:\\Users\\astre\\Documents\\Ghidra-Cpp-Class-Analyzer\\bin\\main",
    "C:\\Users\\astre\\Documents\\Ghidra-Cpp-Class-Analyzer\\bin\\test",
    -snipped- # ghidra jars. .classpath works
    "C:\\Users\\astre\\.vscode-insiders\\extensions\\redhat.java-0.52.0\\server\\config_win\\org.eclipse.osgi\\75\\0\\.cp",
    "C:\\Users\\astre\\.vscode-insiders\\extensions\\redhat.java-0.52.0\\server\\config_win\\org.eclipse.osgi\\46\\0\\.cp"
  ],
  "modulePaths": [],
  "args": [
    "-version",
    "3",
    "-port",
    "58930",
    "-testLoaderClass",
    "org.eclipse.jdt.internal.junit4.runner.JUnit4TestLoader",
    "-loaderpluginname",
    "org.eclipse.jdt.junit4.runtime",
    "-test",
    "ghidra.app.cmd.data.rtti.gcc.TypeInfoTest:test",
    "-port",
    "58929",
    "-port",
    "58929"
  ],
  "vmArgs": [
    "-ea"
  ],
  "encoding": "utf8",
  "env": {
    -snipped-
  },
  "console": "internalConsole",
  "noDebug": true
}'
%TESTC  1 v2
%TSTTREE1,test(ghidra.app.cmd.data.rtti.gcc.TypeInfoTest),false,1,false,-1,test(ghidra.app.cmd.data.rtti.gcc.TypeInfoTest),,
%TESTS  1,test(ghidra.app.cmd.data.rtti.gcc.TypeInfoTest)
%ERROR  1,test(ghidra.app.cmd.data.rtti.gcc.TypeInfoTest)
%TRACES 
java.lang.NullPointerException
	at ghidra.app.cmd.data.rtti.gcc.TypeInfoTest.test(TypeInfoTest.java:19)
	at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
	at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
	at java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
	at java.base/java.lang.reflect.Method.invoke(Method.java:567)
	at org.junit.runners.model.FrameworkMethod$1.runReflectiveCall(FrameworkMethod.java:59)
	at org.junit.internal.runners.model.ReflectiveCallable.run(ReflectiveCallable.java:12)
	at org.junit.runners.model.FrameworkMethod.invokeExplosively(FrameworkMethod.java:56)
	at org.junit.internal.runners.statements.InvokeMethod.evaluate(InvokeMethod.java:17)
	at generic.test.TestThread.run(TestThread.java:106)
%TRACEE 
%TESTE  1,test(ghidra.app.cmd.data.rtti.gcc.TypeInfoTest)
%RUNTIME6062

The NullPointerException that is causing my test to fail is because the extension points aren't being loaded which I think is because the classloader isn't being set.

[jdneo]

Hi @astrelsky,

I'm not quite sure, but I guess you have also set the setting: java.test.defaultConfig. If that is the case, you can just remove that setting or set it to "GhidraTest".

Please feel free to let me know if this can resolve your problem

[astrelsky]

Hi @astrelsky,

I'm not quite sure, but I guess you have also set the setting: java.test.defaultConfig. If that is the case, you can just remove that setting or set it to "GhidraTest".

Please feel free to let me know if this can resolve your problem

I initially had it set to that but removed it and explicitly selected "GhidraTest" as a sanity check. The same results are produced.

[jdneo]

Hmm, I was trying to repro this issue but not able to.

The implementation is here: https://github.com/microsoft/vscode-java-test/blob/master/src/utils/launchUtils.ts#L58-L60

The vmargs should be inserted. (Suppose you are using JUnit 4)

[astrelsky]

Hmm, I was trying to repro this issue but not able to.

The implementation is here: https://github.com/microsoft/vscode-java-test/blob/master/src/utils/launchUtils.ts#L58-L60

The vmargs should be inserted. (Suppose you are using JUnit 4)
I'm attaching the full unedited (except for a few privacy snips) output of the java language server and java test runner to this comment.

If I must I'll attempt to debug the test runner tomorrow. No matter what configuration I make or set I get the same results.

language_server.log
test_runner.log

These are both my global and workspace settings.
workspace:

{
    "java.autobuild.enabled": true,
    "java.test.config": {
        "name": "GhidraTest",
        "workingDirectory": "${workspaceFolder}",
        "args": ["-c", "ghidra.GhidraClassLoader"],
        "vmArgs": ["-Djava.system.class.loader=ghidra.GhidraClassLoader", "-Declipse.project.dir=${workspaceFolder}", "-Dpython.console.encoding=UTF-8", "-Dfile.encoding=UTF8", "-Dlog4j.skipJansi=true", "-Duser.country=US", "-Duser.language=en", "-Dsun.stdout.encoding=UTF-8", "-Dsun.err.encoding=UTF-8", "-Dvisualvm.display.name=Ghidra"]
    },
    "java.dependency.packagePresentation": "flat"
}

global:

{
    "telemetry.enableTelemetry": false,
    "java.format.settings.url": "${env:GHIDRA_SRC}\\eclipse\\GhidraEclipseFormatter.xml",
    "java.format.settings.profile": "GhidraEclipseFormatter",
    "editor.suggestSelection": "first",
    "vsintellicode.modify.editor.suggestSelection": "automaticallyOverrodeDefaultValue",
    "java.configuration.checkProjectSettingsExclusions": false,
    "java.trace.server": "verbose",
    "terminal.integrated.rendererType": "dom",
    "workbench.startupEditor": "newUntitledFile",
    "java.codeGeneration.useBlocks": true,
    "java.completion.guessMethodArguments": true,
    "java.completion.importOrder": [
        "java",
        "ghidra",
        "javax",
        "com",
        "org"
    ],
    "java.implementationsCodeLens.enabled": true,
    "java.referencesCodeLens.enabled": true,
    "java.errors.incompleteClasspath.severity": "ignore",
    "java.test.report.position": "currentView",
    "python.jediEnabled": false,
    "python.insidersChannel": "off",
    "java.test.report.showAfterExecution": "always",
    "java.test.message.hintForSetingDefaultConfig": false,
    "explorer.confirmDelete": false,
    "java.test.log.level": "verbose",
    "files.associations": {
        "*.jj": "java"
    },
    "editor.tokenColorCustomizations": {
        "[sleigh]": {
            "comments": "#229977"  
        }
    },
    "java.codeGeneration.generateComments": true,
    "java.completion.overwrite": false,
    "java.maven.downloadSources": true,
    "java.debug.logLevel": "verbose"
}

[jdneo]

Hi @astrelsky,

I guess I found the root cause. It should be vmargs, not vmArgs.

I need to find out a way to validate the key names in the implementation. Or make it align with the launch configuration. Sorry for the confusion.

[astrelsky]

Hi @astrelsky,

I guess I found the root cause. It should be vmargs, not vmArgs.

I need to find out a way to validate the key names in the implementation. Or make it align with the launch configuration. Sorry for the confusion.

Yes this was the cause. The good old copypasta strikes again. Thank you for your help. I'll leave this open until the naming situation is resolved.