RFR 9: MulticastSendReceiveTests.java fails with NumberFormatException

[Chris Hegarty]

Updated diffs based on an offline conversation with Alan.

diff --git 
a/test/java/nio/channels/DatagramChannel/MulticastSendReceiveTests.java 
b/test/java/nio/channels/DatagramChannel/MulticastSendReceiveTests.java
--- a/test/java/nio/channels/DatagramChannel/MulticastSendReceiveTests.java
+++ b/test/java/nio/channels/DatagramChannel/MulticastSendReceiveTests.java
@@ -97,7 +97,7 @@
                  // no datagram received
                  if (sa == null) {
                      if (expectedSender != null) {
-                        throw new RuntimeException("Expected message 
not recieved");
+                        throw new RuntimeException("Expected message 
not received");
                      }
                      System.out.println("No message received (correct)");
                      return;
@@ -109,10 +109,15 @@
                  buf.flip();
                  byte[] bytes = new byte[buf.remaining()];
                  buf.get(bytes);
-                int receivedId = Integer.parseInt(new String(bytes));
-
-                System.out.format("Received message from %s (id=0x%x)\n",
-                    sender, receivedId);
+                String s = new String(bytes, "UTF-8");
+                int receivedId = -1;
+                try {
+                    receivedId = Integer.parseInt(s);
+                    System.out.format("Received message from %s 
(id=0x%x)\n",
+                            sender, receivedId);
+                } catch (NumberFormatException x) {
+                    System.out.format("Received message from %s 
(msg=%s)\n", sender, s);
+                }

                  if (expectedSender == null) {
                      if (receivedId == id)


-Chris.

On 25/03/15 14:26, Chris Hegarty wrote:
> On 25/03/15 12:36, Alan Bateman wrote:
>> On 25/03/2015 12:16, Chris Hegarty wrote:
>>> This is a trivial test only one-liner. The test should attempt to
>>> reconstruct the String message with the same charset as is used to
>>> convert the original message to bytes, UTF-8.
>>>
>> The fix looks okay and matches the sender. That said, it's possible that
>> this is just interference on the network too, in which case the package
>> should be ignored as is does already when it doesn't match the
>> expected id.
>
> Ah right. So, maybe the test should ignore data that cannot be parsed,
> so then it can continue to try to read the expected id. The current code
> cannot do this.
>....
>
> -Chris.
>
>> -Alan