imperatorx

As JDK 19 is GA today, can you please re-evaluate this PR?

The benefit is using pgjdbc on virtual threads without performance degradation due to carrier thread pinning :) I think replacing synchronized blocks and await/notify with java.util.concurrent structures is safe, it...

I got the exception once from calling a different, very simple JNI method. The only thing I can think of is that the same `ThreadLocalHandles` object is accessed from multiple...

> I am investigating on this one, can you provide a small reproducer ? Sure: Reproducer: https://github.com/imperatorx/graalvm-threadlocal-reproducer The reproducer - starts 10 virtual threads - and on each virtual thread,...

Sometimes it is even throwing Fatal errors like this: ``` Fatal error: java.lang.ArrayIndexOutOfBoundsException: Index 8 out of bounds for length 4 at com.oracle.svm.core.handles.ThreadLocalHandles.pushFrame(ThreadLocalHandles.java:80) at com.oracle.svm.core.jni.JNIObjectHandles.pushLocalFrame(JNIObjectHandles.java:221) at com.oracle.svm.core.jni.JNIGeneratedMethodSupport.nativeCallPrologue(JNIGeneratedMethodSupport.java:58) at java.io.FileOutputStream.writeBytes(FileOutputStream.java:-2) at...

I think the key vector to this is native code calling back to java code through JNI (java -> native -> java), I could not reproduce with simple java ->...

I'm witnessing the opposite behaviour with ServerSentEvents: - On 0.4.4 the final onComplete event is not passed to the response object to signal the end of the event flow (clien...