It is clear that the RFX works with one TUS9x00 at a time. A send command from one TSU9x00 will block other TSU9x00's for the duration.

What happens if another serial device connected to the RFX9600 reports an unsolicited asynchronous event (perhaps a user changes an input from the other device's front panel) while the RFX is blocked processing the first unit's command? Will the event be lost? If so, how will I be able to determine this situation in my ProntoScript program?

Along the same lines. I'm working with Yamaha components that buffer "reports", waiting for the controller to fetch them. If the Yamaha's (unspecified length) buffer overflows (unspecified) difficulties will strike.

In the above RFX9600 unsolicited report scenario, how deep is a serial port's buffer?

---

Using the ProntoScript developer's guide suggestion of using zero timeout serial reads to poll for asynchronous events, I assume that I'll receive a serial timeout if the Receiver's buffer is empty. I'm looking for suggestions about how to descriminate between an empty asynchronous event poll and a problem with a command reply.

I don't believe the async input will be lost. The serial port buffer is, I believe, 512 bytes deep. I also believe that there is a different error depending on timeout vs conflict accessing the serial port.

As RS-232 is NOT truly asynchronous, my suggestion is, if at all possible, to write your modules with TCP/IP vs RS-232. In fact, Barry Gordon, in protest to get vendors to convert to TCP/IP, will likely NEVER write an RS-232 interface.