From: lcargill99@comcast.com   
      
   m.sambin@gmail.com wrote:   
   > Hi.   
   >   
   > I have a native Visual C++ application which acts as a socket server:   
   > it accepts incoming TCP connections, and it reads/write some data   
   > from/to those sockets. Once a new TCP connection is received /   
   > accepted by the server, a new thread is spawn, and all reads (i.e.,   
   > calls to recv()) on that connected socket happen in that separate   
   > thread. All communication is done according to a well-defined   
   > standard application-level protocol, implemented by several clients   
   > by different developers/vendors. Hence, my server application is able   
   > to communicate with clients by different developers/vendors, as long   
   > as they comply with this well-defined application-level protocol.   
   >   
   > Now, my server application works fine with 99.9% of the client   
   > applications. Nevertheless, I am experiencing strange delays with   
   > just a few clients.   
   >   
   > In particular, with these "problematic" clients, once the client   
   > connection has been accepted by my server application, the recv()   
   > calls in my dedicated "receiver" thread hang for exactly 5 seconds   
   > before returning, even though data shall be available, since the   
   > client has already completed several send() calls.   
   >   
   > The clients connecting to my server application run on hosts residing   
   > on the same LAN as the host upon which my server application runs.   
   >   
   > Please notice the following facts:   
   >   
   > 1 - With these specific "problematic" clients, **ALL** recv() calls   
   > in my server application hang for 5 seconds before returning.   
   >   
   > 2 - My server application works fine (i.e., recv() show no relevant   
   > delays) with other clients (actually, the vast majority of them).   
   >   
   > 3 - On the other side, those "problematic" clients seem to work fine   
   > (i.e., no particular delays) with other third-party server   
   > applications implementing the same application-level protocol as my   
   > own server application. Some of those third-party server applications   
   > are open source: I've inspected the source code of those third-party   
   > server applications, but I could not find any relevant difference in   
   > their implementation which could apparently justify the 5-seconds   
   > delay that I am experiencing in my server application.   
   >   
   > I realize that I described the issue in quite generic terms, but   
   > maybe this "5 exact seconds" delay may tell something to some TCP/IP   
   > expert...   
   >   
   > I would really appreciate any hint or suggestion to troubleshoot this   
   > problem.   
   >   
   > Thanks in advance and best regards,   
   >   
   > Marco   
   >   
      
   You might consider looking into nonblocking sockets.   
      
   --   
   Les Cargill   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)