Thoroughly testing an RMT based implementation?

[papadeltasierra]

I've been playing with adding to add ESP32 RMT (Remote Control Transceiver) support to this project and was wondering what testing would you consider thorough? My current thoughts are:

1. All existing unit tests must pass and as many new unit tests written as sensible
2. All four combinations of old/RMT send to old/RMT recv should work using examples
3. "Dumping" an unknown remote must be possible using the old style receive as this doesn't need to know the actual parameters of the transmission protocol (unlike RMT)
4. Existing ESP8266 must continue to work (using the old stye of course)
5. What else?

Are there are any of the protocols that you feel might be particularly challenging? I've implemented such that, for example, Manchester encoding should work fully and I've not spotted any oddities in the other protocols (perhaps I've not studied hard enough)?

Just waiting for some new boards to arrive before I can start any testing. I've been using a combination of AI, me reading up and reviewing and some small by hand fixes and it looks like it should work, but the proof will be in the pudding ;-).