Experimental Physics and
Industrial Control System

"Jeff Hill" <[email protected]> · Mon, 7 Jul 2003 18:38:21 -0600

Bernhard,

This problem appears to be isolated to situations where the log client is
running on Microsoft Windows. There appear to be portability problems when
socket(), _open_osfhandle(), and _fdopen() are used to create a standard IO file
pointer. I committed changes to both R3.14 and R3.15 so that the log client
library does not use fprintf() to indirectly write into the socket.

The log client now appears to work correctly on win32. During testing the
unmodified R3.14 log server also appears to work w/o problems on win32.

I can send you a copy of the two files that were changed if you are interested.

Sorry about the inconvenience.

Best Regards,

Jeff

> -----Original Message-----
> From: kuner [mailto:[email protected]]
> Sent: Friday, July 04, 2003 4:47 AM
> To: tech-talk
> Subject: How to use the errlog with epics 3.14 on a windows Pc?
> 
> Hi all,
> 
> I tried this for initialisation: Set in shell environment:
> 
>   export  EPICS_IOC_LOG_INET=193.149.12.137
> 
> the variable is shown with epicsEnvShow from the iocShell
> 
> init iocLog in iocShellMain.cpp:
> 
>   int main(int argc,char *argv[])
>   {
>     iocLogInit();
>     if(argc>=2)
>     {
>         iocsh(argv[1]);
>         epicsThreadSleep(.2);
>     }
>   }
> 
> and used errlogPrintf. The messages are written on the shell, but not to
> the iocLogServer (the server doesn't run on a PC). There is a cyclic
> connection lost an reconnect to the server:
> 
>   log client: lost contact with error message log server at
> "193.149.12.137:7004"
>   ....
>   log client: connected to error message log server at
> "193.149.12.137:7004"
> 
> have a look on the server machine (ocean) shows for my client (hazel):
> 
>   ocean: .../1-3/caPutLogApp > netstat -a |grep 7004
>   tcp        0      0  ocean.7004             hazel.1769
> FIN_WAIT_2
>   tcp        0      0  ocean.7004             hazel.1766
> TIME_WAIT
>   tcp        0      0  *.7004                 *.*
> LISTEN
> 
> Whats going wrong there?
> 
> --
> -------------------------------------------------------------------
> -- [email protected] ----------- Bessy Gmbh -----------------
> -- +49 30 6392 4981 ------------------ Albert Einstein Strasse 15 -
> -------------------------------------- 12489 Berlin ---------------

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System