The client code would only have gotten to where it is if the connection succeeded and the client was able to write data to the connection without error. If the client is able to get that far, and also have the connection still sitting waiting for data, then something is seriously messed up with the AIX socket support. The problem has got to be in the handling of the request on the rxapi daemon side. If something dies in the request (such as would be the case with the try\catch blocks I pointed you to), then the thread in question would shutdown. This should have closed the connetion and cause a failure in the client, which does not appear to have happened. However, the request should not happen to be dying either.
Another potential problem might be failure of some kind in creating the thread to service that connection in the first place. although nothing has changed in any of the threading code that I'm aware of that would suddenly start causing an error. In any event, the problem has got to be in the rxapi server, there should be a thread servicing that request, to it is necessary to figure out why that thread is no longer there. Rick On Mon, Jul 27, 2009 at 8:00 AM, Rainer Tammer<[email protected]> wrote: > Hello Rick, > > Rick McGuire wrote: >> Not much of a mystery there, the first send() method calls the second. >> That's a perfectly normal looking call stack. You're probably going >> to need to debug ApiServer::processMessages() method, which is what >> should be on the other end of this client connection. Pay particular >> attention to the catch() blocks to see if some sort of error occurred. >> >> > I think that the problem is on the client side. During the hang I can > see no thread in the rxapi server... > I think that the client never gets connected to the rxapi server... >> Also, you might want to check that you don't have an old version of >> rxapi running. The latest builds are using the new port number, so it >> is possible that something has gotten messed up there. >> >> > Killed and cleaned 10 times ... >> Rick >> > Bye > Rainer >> On Mon, Jul 27, 2009 at 7:40 AM, Rainer Tammer<[email protected]> wrote: >> >>> Hallo, >>> there is one strange thing in the call stack: >>> >>> ClientMessage.ClientMessage::send(SysClientStream*)(this = >>> 0x2ff20260, pipe = 0x30c13978), line 82 in "ClientMessage.cpp" >>> .. >>> ClientMessage.ClientMessage::send()(this = 0x2ff20260), line 59 in >>> "ClientMessage.cpp" >>> >>> I do not understand why both functions are active... >>> >>> I think somehow the initial connect does not go through and ooRexx does >>> not catch this. >>> Unfortunately my knowledge of the inner workings of the interpreter >>> (rxapi client part) is a bit to low .. :-(( >>> >>> >>> Bye >>> Rainer >>> >>> ----------------------------------------------------------------------------- >>> _______________________________________________ >>> Oorexx-devel mailing list >>> [email protected] >>> https://lists.sourceforge.net/lists/listinfo/oorexx-devel >>> >>> >> >> ------------------------------------------------------------------------------ >> _______________________________________________ >> Oorexx-devel mailing list >> [email protected] >> https://lists.sourceforge.net/lists/listinfo/oorexx-devel >> >> >> > > > ------------------------------------------------------------------------------ > _______________________________________________ > Oorexx-devel mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/oorexx-devel > ------------------------------------------------------------------------------ _______________________________________________ Oorexx-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/oorexx-devel
