REF: refactor FB_Logger based on FB_ConnectionlessSocket

[klauer]

As best as I can tell, the current master is based partly on old code that used FB_SocketUdpCreate, which required much more in the way of state tracking. That's been cleaned up in this PR and drastically simplified.

As this is my first bit of non-trivial PLC code, you might want to look over the changes closely...

Changes / Notes

• Remove eState and rely on FB_ConnectionlessSocket to handle socket creation/deletion for us
• Only grab from the log message queue when a socket exists
• Do not require local port to be the same as syslog - it should just be assigned by the OS
• Retry creating the socket if previously failed
• Useful socket error codes, which map onto winsock error codes: https://infosys.beckhoff.com/english.php?content=../content/1033/tf6310_tc3_tcpip/18014398593621899.html&id

Testing

if xInit = 1 then
    xInit := 0;
    logger(i_xReset:=TRUE);
END_IF

iCycle := iCycle + 1;

IF (iCycle MOD 500) = 0 THEN
    GVL_Logger.fbLogMessage(
             i_sMsg:='message',
             i_eSevr:=E_MesgSevr.Debug,
             i_eSubsystem:=E_Subsystem.MOTION
             );
END_IF
logger();

And using nc as a syslog stand-in /usr/bin/nc -u -l 59999 -b <interface>, I periodically see:

<167> 1 2019-07-15T20:22:19.145Z MOTION - - BOMmessage

Requirements

• TF6310 must be installed and licensed on the PLC.
  • The version from 2014 in /reg/g/pcds/plc-common/InstallFiles/Beckhoff does not work with 4022.

TODO

• Add the above test information and requirements to documentation
• Copy over TF6310 to the shared location

cc @teddyrendahl

[klauer]

@ZLLentz fair enough!

[slacAWallace]

Working on this now.

[slacAWallace]

I was able to confirm the PLC is sending out syslog formatted messages using wireshark so it seems changing to the connectionless socket fb worked.

[klauer]

👏