Fix to avoid incompatibility if compile with jdk9

[snnn]

Describe the bug
See apache/felix#114

Urgency
If there are particular important use cases blocked by this or strict project-related timelines, please share more information and dates. If there are no hard deadlines, please specify none.

System information

• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): macOS
• ONNX Runtime installed from (source or binary): source
• ONNX Runtime version: master
• Python version: N/A
• Visual Studio version (if applicable):
• GCC/Compiler version (if compiling from source):
• CUDA/cuDNN version:
• GPU model and memory:

To Reproduce

• Describe steps/code to reproduce the behavior.

Build onnxruntime with java enabled, and run the tests.

    => java.lang.NoSuchMethodError: java.nio.FloatBuffer.rewind()Ljava/nio/FloatBuffer;
       ai.onnxruntime.InferenceTest.testModelInputBuffer(InferenceTest.java:935)
       sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
       sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
       sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
       java.lang.reflect.Method.invoke(Method.java:498)
       org.junit.platform.commons.util.ReflectionUtils.invokeMethod(ReflectionUtils.java:686)
       org.junit.jupiter.engine.execution.MethodInvocation.proceed(MethodInvocation.java:60)
       org.junit.jupiter.engine.execution.InvocationInterceptorChain$ValidatingInvocation.proceed(InvocationInterceptorChain.java:131)
       org.junit.jupiter.engine.extension.TimeoutExtension.intercept(TimeoutExtension.java:149)
       org.junit.jupiter.engine.extension.TimeoutExtension.interceptTestableMethod(TimeoutExtension.java:140)
       [...]
  JUnit Jupiter:InferenceTest:testModelInputFLOAT()
    MethodSource [className = 'ai.onnxruntime.InferenceTest', methodName = 'testModelInputFLOAT', methodParameterTypes = '']
    => java.lang.NoSuchMethodError: java.nio.FloatBuffer.rewind()Ljava/nio/FloatBuffer;
       ai.onnxruntime.OnnxTensor.getFloatBuffer(OnnxTensor.java:167)
       ai.onnxruntime.InferenceTest.testModelInputFLOAT(InferenceTest.java:906)
       sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
       sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
       sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
       java.lang.reflect.Method.invoke(Method.java:498)
       org.junit.platform.commons.util.ReflectionUtils.invokeMethod(ReflectionUtils.java:686)
       org.junit.jupiter.engine.execution.MethodInvocation.proceed(MethodInvocation.java:60)
       org.junit.jupiter.engine.execution.InvocationInterceptorChain$ValidatingInvocation.proceed(InvocationInterceptorChain.java:131)
       org.junit.jupiter.engine.extension.TimeoutExtension.intercept(TimeoutExtension.java:149)

[snnn]

@Craigacp could you please help?

[Craigacp]

Sure, that's easy to fix in the build system. How did you manage to skew JAVA_HOME between the build and test if it was in the same command?

[snnn]

They are on two different machines. We build the jar on Windows then run the tests on Win/Linux/Mac. So very likely the java version mismatches.

[Craigacp]

Ok. Basically we need to conditionally set the -release 8 flag when compiling on 9 or newer. It doesn't exist in 8 so it can't be on permanently. I'm just figuring out the gradle syntax.

[snnn]

Do we need to continue to support JDK8?

[Craigacp]

Personally I would like it to for at least another release, as Tribuo hasn't moved to Java 11 yet, and Tribuo is why I wrote the Java bindings. I'm not sure if our internal users have moved on from Java 8 yet, I'll check.

Changing the minimum Java version is probably worth a separate issue and some discussion from the interested parties in Oracle, MS & AWS. I dunno what version of Java DJL uses. Apache Spark only just released support for Java 11 in version 3.0 which came out in the middle of last year.

The patch in gradle to make it work when compiling on 9+ but running on 8 is small, I'll work up a PR.