In this case just download the axis2/C kit and build axis2C, then use a debugger and debug the axis2c_engine.
Very interesting when it starts up; go once through and you will understand axis much better. i.e. see when things take place and how they come into effect. Look for a directory IDES and find various VC VC6 VC9 Projects prepared as well as a Makefile project for your OS. But just be made aware, also it is documented on many places; the simple servers axis2/J and axis2/C are not for production! Josef Von: zack Hasit [mailto:zackha...@gmail.com] Gesendet: Donnerstag, 21. Februar 2013 01:48 An: Apache AXIS C User List Betreff: Re: Axis 2C embedded server Hi, How can I get more information on Inbuilt HTTP server called simple axis server. I am trying to find out about its architecture. I need to understand how threads are created and deleted (does it have a thread pool with max and min threads which are instantiated at start-up ?) and if its scalable. Word "simple" is making me think there could be limitations for its use in production scenario ? Thanks On Wed, Feb 20, 2013 at 10:41 AM, Stadelmann Josef <josef.stadelm...@axa-winterthur.ch> wrote: It is possible using addressing and the proper addressing header information and configuration records both at client and at server. i.e. in the header you need to tell the service which client is calling (request-id) and about the listener/port to contact as well as the call-back-routine to be called. The server once finished needs to know to which request-id this response-id belongs to; but that is not all. it needs to know clinet-host:port and client-call-back-routine processing the response. It is not so simple; hence I encourage you to go through the various echo examples AND OTHERS, client and server side, unless you understand the asynchronous technology field. Keep in mind you might have an asynchronous working transport AND you may have an asynchronous working client/server. This is not the same. your client might send a request, the server receives it, and allows the thread to return before it starts working your client may hand over a request to the transport and continue, the transport sends the request, the server accepts it and allows the thread to return, the transport at the client side is happy and maybe signals that o the client. Later on the server does the same with the response in reverse direction. This are two scenarios with many wrinkle and challenge to master. basically, the service, when it has the result ready after some time, needs to know the host:port numbers and a call-back-routine in the client to contact; and there is a need to know for the infrastructure axis2-engine to which request this response belongs to, given your client has multiple outstanding request belonging to request sent some time ago. Again: starts simple, keep subsystems small, test it, and integrate when all subsystems works. Unless you master your asynchronously working axis2 client talking to your asynchronous working web service, you don't need more than one later 2 pc's, and yes: long lasting sessions might be another issue to look at. Josef Von: zack Hasit [mailto:zackha...@gmail.com] Gesendet: Dienstag, 19. Februar 2013 18:23 An: Apache AXIS C User List Betreff: Re: Axis 2C embedded server Hi, I looked at "echo_non_blocking_dual.c". It is a client implementation to call the service. I was looking for server side. Ill try to explain the scenario.... When the echo service gets called following function gets called on server side : axis2_echo_echo. In this function the node "axiom_node_t *ret_node" is returned. What I really want to do is that I want to modify axis2_echo_echo to send the text message (with additional header info) to another server (lets call this server SERVER-A) using activeMQ and then go and service next echo request from client and forget about this one for now. Meanwhile I'll spawn another thread (in my application) to read responses from SERVER-A asynchronously (via activeMQ). Since responses would be out of order ill identify them based on headerinfo and ill know where this response needs to be sent to. I want to then send this to correct client. Please let me know if this is possible. On Tue, Feb 19, 2013 at 4:47 AM, Stadelmann Josef <josef.stadelm...@axa-winterthur.ch> wrote: seek this file in your axis2/c kit "echo_non_blocking_dual.c" not under samples but user guide clients ... J And Axis2/C is at version 1.6.0, and the trunk sources have entries from January 2013, so there is work going on. i.e. http://svn.apache.org/repos/asf/axis/axis2/c/core/trunk/src/core/transpo rt/http is brand new. Josef Von: zack Hasit [mailto:zackha...@gmail.com] Gesendet: Dienstag, 19. Februar 2013 02:45 An: c-user@axis.apache.org Betreff: Axis 2C embedded server Hi, I am planning to use Axis 2C. Could someone please help answer this. 1. Is the standalone http server (simple http server) good for production situation? I ask because it maybe for for testing purposes only. Any performance concerns with it ? 2. Is Axis being maintained actively ? I ask because there hasn't been a new release since 2009 ? 3. I need to break request and response in two separate functions. Is that possible. In samples I see echo implementation where one method implements service and returns the node to send response. I don't want to send response on return but at a later time after I do do some other processing aynchronously. Is there any sample for that ? Please let me know if there is another place to ask this. Thanks
