Peter Barker

> Why isn't this PR be deleting the whole hwdef/CubeOrange-periph-heavy folder? ... my recollection is that you were the only person speaking up in defence of CubeOrange-periph-heavy. Happy to see...

> I think we need a little more thought than a global replace. In some cases this is correct, in other cases this will build but it won't necessarily do...

@robustini the two PRs have been merged. Are we OK to close this issue?

I tested this in SITL, both MAVProxy and MissionPlanner, and on CUAVv5, both MAVProxy and MissionPlanner.

@khancyr did a whole lot of work to not run tests that we don't need to. I think we can close this, @ES-Alexander ?

@maciek252 I'd be surprised if it didn't work at this point. usart3 is not the same thing as `SERIAL3` - tridge's PR would put usart3 as `SERIAL7_`. Did you eventually...

I should note that if we didn't want to do the multiple-inheritance thing then we could do something like ``` parser = new NMEAParser(protocol, protocol_instance, functor_reset, functor_good_sentence_type, functor_term, functor_decode_complete); ```...

@rmackay9 sounds good - I'll probably try to convert AP_GPS_NMEA - happen to have a GPS that produces that on my desk ATM.

I think you're right, and it's pretty easy to fix. Is it a worthwhile patch if the thing is generated? We'll save a dozen bytes or so at the risk...

Still need to do parameter conversion.