Processing Multiple Messages in a Single Time Step
This example illustrates a common design pattern with messages: processing all the messages in the queue in a single time step.
The normal semantics of messages implies that only a single message is processed in every time step. This is because once a message is removed from the queue due to evaluating a message trigger, it is considered valid till the end of that time step. Further evaluations of the same message in the same time step reuse the already removed message instead of removing a new message. See the example Lifetime of Stateflow® Messages for more information.
However, in some cases, it might be necessary to process all the messages in the queue in the same time step.
The "Sender" chart sends three messages at time step t = 0 which all get queued in the input queue "M" of the "Receiver" chart.
The "Receiver" chart is in state "Init" at t = 0 and transitions to "ProcessAllMessages" at t = 5. In the entry action, it needs to get the sum of the data in all messages in the queue.
It accomplishes this by calling the graphical function "sumOfAllMessageData". This function has a while loop which evaluates the input message "M" in a queue. Normally, this would have resulted in the same message being returned every time, resulting in an infinite loop. However, after processing the message data:
yy = yy + M.data;
it immediately forwards the input message to an output "junk" queue.
This allows the next message evaluation in the same time step to remove the next message from the queue. Once all messages have been removed (and forwarded), the message evaluation returns false and the while loop is exited.
Simulating the model shows the output of the "Receiver" chart goes from 0 to 6 at time t = 5.