Optimize GC behaviour

JNA is making an insane amount of garbage when I'm trying to find a fake process by name.

Here is a screenshot of the allocations (about 100k.sec)

http://i.stack.imgur.com/u3D2I.png

Here is the test case (used 4.3.0 SNAPSHOT of JNA)

    import com.sun.jna.Native;
    import com.sun.jna.platform.win32.Kernel32;
    import com.sun.jna.platform.win32.Tlhelp32;
    import com.sun.jna.platform.win32.WinDef;
    import com.sun.jna.platform.win32.WinNT;

    /**
     * Created by Jonathan on 8/26/2016.
     */
    public class Main {

        public static void main(String[] args) {
            while (true)
                openProcess("doesntexist.exe");
        }

        private static final WinDef.DWORD DWORD_ZERO = new WinDef.DWORD(0);
        private static final Tlhelp32.PROCESSENTRY32 entry = new Tlhelp32.PROCESSENTRY32.ByReference();

        private static WinNT.HANDLE openProcess(String processName) {
            WinNT.HANDLE snapshot = Kernel32.INSTANCE.CreateToolhelp32Snapshot(Tlhelp32.TH32CS_SNAPALL, DWORD_ZERO);
            try {
                while (Kernel32.INSTANCE.Process32Next(snapshot, entry)) {
                    String fileName = Native.toString(entry.szExeFile);
                    if (processName.equals(fileName))
                        return Kernel32.INSTANCE.OpenProcess(WinNT.PROCESS_ALL_ACCESS, true, entry.th32ProcessID.intValue());
                }
            } finally {
                Kernel32.INSTANCE.CloseHandle(snapshot);
            }
            return null;
        }

    }

And finally here is the memory snapshot
https://dl.dropboxusercontent.com/u/91292881/ShareX/2016/08/JNA%204.3.0.snapshot

Start with the provided Main reproduction and its Kernel32.INSTANCE CreateToolhelp32Snapshot, Process32Next, Native.toString, and CloseHandle calls. Run the loop while profiling allocations to identify the source of the garbage during a failed process lookup. Done means the same lookup behavior with materially lower allocation and GC activity.