The latest version of the TCP Sampler opens one socket for each identical host:port combination in each thread. [Different thread sockets to the same destination can be achieved by using e.g. localhost:8888 and LocalHost:8888.]
The current implementation assumes a request/response protocol, but it should be quite easy to extend that to provide separate send and receive samplers. Either as a new Sampler, or perhaps better by adding some options to the existing one. I can do that part. It should also be fairly simple to implement your own protocol - see TCPClientImpl.java for the default. I can help with that if required. S. On 11/08/05, Vinod Panicker <[EMAIL PROTECTED]> wrote: > On 8/11/05, sebb <[EMAIL PROTECTED]> wrote: > > I'm not convinced that you need separate threads to run this test. > > > > If you want to send a message on one connection, and then wait until a > > message arrives on another connection, then it seems to me that these > > can be in the same thread. > > Any pointers on how this can be achieved or should I RTFM? Its not > necessary that the two connections are independent execution units. I > could make do with just one thread. > > Regards, > Vinod. > > > On 11/08/05, Vinod Panicker <[EMAIL PROTECTED]> wrote: > > > On 8/10/05, Peter Lin <[EMAIL PROTECTED]> wrote: > > > > I don't think that is possible in jmeter. The threads are independent > > > > of each other and do not thread.join(thread) to the best of my > > > > knowledge. > > > > > > But this kind of functional testing becomes essential when trying to > > > test asynchronous non-request-response based protocols. Any idea if > > > support for this is planned? > > > > > > > On 8/10/05, Vinod Panicker <[EMAIL PROTECTED]> wrote: > > > > > Let me explain in more detail. > > > > > > > > > > There are two threads, X and Y. Both run simultaneously to reach > > > > > checkpoint A. X sends a request to the server which is sent to Y. X > > > > > needs to know that Y received what was sent. > > > > > > > > > > To implement the above I was asking for something like dependency > > > > > configuration / join implementation. > > > > > > > > > > Regards, > > > > > Vinod. > > > > > > > > > > On 8/10/05, Christian Baumgartner <[EMAIL PROTECTED]> wrote: > > > > > > Am I right that you want to run both Threads at the same time, but > > > > > > one > > > > > > should wait for the other finishing? > > > > > > > > > > > > If you click TestPlan there is an Option that runs ThreadsGroups > > > > > > one after > > > > > > another. But maybe it's more practicable to manage it with the > > > > > > Controllers > > > > > > and Timers. > > > > > > > > > > > > > > > > > > Chris > > > > > > > > > > > > > > > > > > -----Ursprüngliche Nachricht----- > > > > > > Von: Vinod Panicker [mailto:[EMAIL PROTECTED] > > > > > > Gesendet: Mittwoch, 10. August 2005 10:09 > > > > > > An: [email protected] > > > > > > Betreff: Dependency support for threads > > > > > > > > > > > > > > > > > > Hi, > > > > > > > > > > > > Is there any way by which I can specify that Thread x needs to wait > > > > > > till > > > > > > Thread y completes successfully? > > > > > > > > > > > > I'm trying to test an asynchronous TCP based IM protocol. So I > > > > > > want to be > > > > > > able to test if a message sent by x reached y. x and y will be > > > > > > independent > > > > > > sessions. > > > > > > > > > > > > Regards, > > > > > > Vinod. > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
