Central-system 1.6 example - "fatal error: concurrent map writes" when connecting multiple chargers rapidly

[mbbirkel]

While testing the OCPP 1.6 central-system using the provided example code, i sometimes get the error fatal error: concurrent map writes when connecting multiple mock-chargers within a short timeframe (around 1ms or lower).

I used similar code for handling connections to the 1.6 central-system example:

	// Add handlers for dis/connection of charge points
	centralSystem.SetNewChargePointHandler(func(chargePoint ocpp16.ChargePointConnection) {

		handler.chargePoints[chargePoint.ID()] = &ChargePointState{connectors: map[int]*ConnectorInfo{}, transactions: map[int]*TransactionInfo{}}
		log.WithField("client", chargePoint.ID()).Info("new charge point connected")
	})

However, this code will have errors with concurrent write to handler.chargePoints, like described.
We can fix this issue by adding a mutex like so:

var mux sync.Mutex         
        
   // ...

	// Add handlers for dis/connection of charge points
	centralSystem.SetNewChargePointHandler(func(chargePoint ocpp16.ChargePointConnection) {
		mux.Lock()
		defer mux.Unlock()
		handler.chargePoints[chargePoint.ID()] = &ChargePointState{connectors: map[int]*ConnectorInfo{}, transactions: map[int]*TransactionInfo{}}
		log.WithField("client", chargePoint.ID()).Info("new charge point connected")
	})

Should thread-safety of the chargePoints map be the responsibility of the library, or should the function that is passed to SetChargePointDisconnectedHandler be responsible for thread safety? If it is not the responsibility of the library, I think at least the example for central-system 1.6 should be updated with a mux like I showed here! 😄

[lorenzodonini]

Hey @mbbirkel, incoming callbacks from chargers (both on new connection and on messages) are fully concurrent, hence they need to be synchronized by the application.

I will update the examples to include the mux on the server side, hopefully making it clearer. Thanks for pointing it out.