I guess I responded instinctively without thinking it through. This is definitely a correct behavior.
Cheers On Jan 30, 2008 9:39 AM, gustav.mauer <[EMAIL PROTECTED]> wrote: > > I am under the impression this is correct behaviour. See for example: > http://java.sun.com/products/jms/tutorial/ page 63 > > > > Sebastjan Trepca wrote: > > > > Hi, > > > > I'm using ActiveMQ(both versions are affected) with STOMP protocol and > > noticed a problem with it. At least I hope it's a problem. > > > > When you ACK a message through stomp, all messages until the right one > > gets ACKed too. > > > > If we check the code: > > > > protected void onStompAck(StompFrame command) throws ProtocolException > > { > > checkConnected(); > > > > // TODO: acking with just a message id is very bogus > > // since the same message id could have been sent to 2 different > > // subscriptions > > // on the same stomp connection. For example, when 2 subs are > > created on > > // the same topic. > > > > Map<String, String> headers = command.getHeaders(); > > String messageId = headers.get(Stomp.Headers.Ack.MESSAGE_ID); > > if (messageId == null) { > > throw new ProtocolException("ACK received without a > > message-id to acknowledge!"); > > } > > > > TransactionId activemqTx = null; > > String stompTx = headers.get(Stomp.Headers.TRANSACTION); > > if (stompTx != null) { > > activemqTx = transactions.get(stompTx); > > if (activemqTx == null) { > > throw new ProtocolException("Invalid transaction id: " > > + stompTx); > > } > > } > > > > boolean acked = false; > > for (Iterator<StompSubscription> iter = > > subscriptionsByConsumerId.values().iterator(); iter.hasNext();) { > > StompSubscription sub = iter.next(); > > MessageAck ack = sub.onStompMessageAck(messageId); > > if (ack != null) { > > ack.setTransactionId(activemqTx); > > sendToActiveMQ(ack, createResponseHandler(command)); > > acked = true; > > break; > > } > > } > > > > if (!acked) { > > throw new ProtocolException("Unexpected ACK received for > > message-id [" + messageId + "]"); > > } > > > > } > > > > > > synchronized MessageAck onStompMessageAck(String messageId) { > > > > if (!dispatchedMessage.containsKey(messageId)) { > > return null; > > } > > > > MessageAck ack = new MessageAck(); > > ack.setDestination(consumerInfo.getDestination()); > > ack.setAckType(MessageAck.STANDARD_ACK_TYPE); > > ack.setConsumerId(consumerInfo.getConsumerId()); > > > > int count = 0; > > for (Iterator iter = dispatchedMessage.entrySet().iterator(); > > iter.hasNext();) { > > > > Map.Entry entry = (Entry)iter.next(); > > String id = (String)entry.getKey(); > > MessageId msgid = (MessageId)entry.getValue(); > > > > if (ack.getFirstMessageId() == null) { > > ack.setFirstMessageId(msgid); > > } > > > > iter.remove(); > > count++; > > > > if (id.equals(messageId)) { > > ack.setLastMessageId(msgid); > > break; > > } > > > > } > > > > ack.setMessageCount(count); > > return ack; > > } > > > > It loops through all messages and ACKs them until it finds the correct > > message id. So when you send few messages and the last one is the > > processed first all previous ones get ACKed? > > > > Is this a feature of STOMP protocol or is it a bug? > > > > Thanks, Sebastjan > > > > > > -- > View this message in context: > http://www.nabble.com/Stomp-protocol-problems-tp15144123s2354p15177216.html > Sent from the ActiveMQ - User mailing list archive at Nabble.com. > > -- Dejan Bosanac www.scriptinginjava.net
