Matthijs Kooijman
Matthijs Kooijman
> I went to read the other feature request, which I'm all in favor, but it doesn't fix the fact that a SERCOM#_Hanlder() has been defined. It should, the idea...
Yeah, the would be moved to separate files per instance. Though I now see that indeed that these handlers are currently defined in variant.cpp, not in the core itself, so...
This seems like a reasonable approach, but I think we can do better. For AVR, we already set things up so that if `Serial` is not references, the entire object...
Good analysis, it seems that clearing the flag first is good for the reasons already given: 1) makes sure that the value read in the ISR callback will always (eventually)...
It is a bit of a breaking change, but given the difference between AVR and SAMD now, one could also think of it as a bugfix. I also wonder how...
> Yes it does. @coffeemachine This is only true because your PR does not just allow `clearInterrupt()`, it also *prevents* clearing the interrupt *after* the ISR if `clearInterrupt()` was called...
To keep the discussion clean, I created new issues for the "separate serial objects" implementation here: https://github.com/arduino/ArduinoCore-sam/issues/100 and https://github.com/arduino/ArduinoCore-samd/issues/489. Let's keep *this* issue about adding the serial interrupt API.
At first glance, seems like a bug to me. I thought that maybe `adc12b_get_latest_value` would return a zero padded value, but it seems to just return the `LCDR` register as-is:...
One problem with arguments to the constructor is that they cannot influence the static size of the class, so the only way to implement that is by using malloc for...
> and it is anyways very little code in the ring buffer implementation, so I suppose this would not come at a too large flash cost. Yeah, so that's probably...