| Applying Colin's patch https://issues.apache.org/activemq/browse/AMQ-2129 Thanks! |
|
More...
|
almost 16 years ago
|
| Switch to using AsciiBuffers instead of strings wherever possible to avoid UTF8 marshalling.. also someimes you can avoid demarshalling alltogether. |
|
More...
|
almost 16 years ago
|
| Added a proto2 wire which uses the new externalizable style encoding/decoding. Just so we can benchmark protobuf encoding vs. more standard java style externalizable encoding. |
|
More...
|
almost 16 years ago
|
| Made payload generation more efficent. |
|
More...
|
about 16 years ago
|
| - You can now force marshalling in the pipe transport - Better toString in the SingleFlowRelay - Producer can now send big messages - Support doing a traditional sleep in the consumer to simulate think time. |
|
More...
|
about 16 years ago
|
| deploy the test jar too |
|
More...
|
about 16 years ago
|
| Making the MockBrokerTest easier to extend |
|
More...
|
about 16 years ago
|
| lets prototype a better nio module in a differnt sandbox area.. we are tackling enough as it is here |
|
More...
|
about 16 years ago
|
| Removed unused dependency |
|
More...
|
about 16 years ago
|
| Removing empty dirs |
|
More...
|
about 16 years ago
|
| Switching to the alternative protobuf api |
|
More...
|
about 16 years ago
|
| - The wire protocol used in the staeful interface of the ProtoWireFormatFactory did not match exactly what the non-statefull version did. Matched them up. - Since to wireformat is doing the framing, swithced to using the toUnframedByteArray() methods to save a few bytes. |
|
More...
|
about 16 years ago
|
| Applying Colins patch at: https://issues.apache.org/activemq/browse/AMQ-2115 basically: |
|
More...
|
about 16 years ago
|
| - the ISinkController looks and smells like a IFlowSink.. so lets let it be one. - Can replace using the exclusive queue for connection outbound messages with just a FlowContoller |
|
More...
|
about 16 years ago
|
| Switching to a protobuf based wireformat |
|
More...
|
about 16 years ago
|
| Cleaned up the tests a little, and fixed a deadlock situation |
|
More...
|
about 16 years ago
|
| Applying colins patch https://issues.apache.org/activemq/browse/AMQ-2107 |
|
More...
|
about 16 years ago
|
| Adding a PipeTransport so we can bypass the IO layer in our testing. |
|
More...
|
about 16 years ago
|
| test case is now working against a real IO impl. |
|
More...
|
about 16 years ago
|
| Getting closer to being able to test real IO |
|
More...
|
about 16 years ago
|
| Brekaing out the nested classes of the MockBrokerTest in preperation of hooking in a version of the test the uses some real io. |
|
More...
|
about 16 years ago
|
| Fixing the 10_1_1 p2p case |
|
More...
|
about 16 years ago
|
| Initial spike of the flow module. This is basically an experiment of a different flow control and threading model to see if we can make a faster/simpler version of the broker. This code was jointly developed between Colin Macnaughton (yes, he has an icla on file) and me. |
|
More...
|
about 16 years ago
|
| Making a little sandbox area to play with some new flow control ideas. |
|
More...
|
about 16 years ago
|
