PATCH: remove all traces of java.io.FileDescriptor from java.net

[Bryce McKinlay]

Jeff Sturm wrote:

>Here's my updated patch.  Hans convinced me that close/finalize ought to
>synchronize access to fnum, I've included that below.
>

Hmm, why is that? Also I noticed that it is actually synchronizing 
twice, both in finalize and close?

Bryce.

>@@ -108,6 +103,22 @@ class PlainSocketImpl extends SocketImpl
>   private native void write(byte[] buffer, int offset, int count)
>     throws IOException;
> 
>+  protected void finalize() throws Throwable
>+  {
>+    synchronized (this)
>+      {
>+	if (fnum != -1)
>+	  try
>+	    {
>+	      close();
>+	    }
>+	  catch (IOException ex)
>+	    {
>+	      // ignore
>+	    }
>+      }
>+    super.finalize();
>+  }
> 
>   /** @return the input stream attached to the socket.
>    */
>
>@@ -413,6 +410,9 @@ java::net::PlainSocketImpl::accept (java
> void
> java::net::PlainSocketImpl::close()
> {
>+  // Avoid races from asynchronous finalization.
>+  JvSynchronize sync (this);
>+
>   // should we use shutdown here? how would that effect so_linger?
>   int res = ::close (fnum);
> 
>