Using an Interrupt from a method inside a class

[gfvalvo]

For me the code compile without that bad change.

Bulldog's original code compiled for me too. I was just wondering if my use of casting:
(void (*)())(&fan2->handleInterrupt)
would function properly, not just compile without error.

There are some warnings...

The warnings depend on what processor I try to compile for. No warnings for Uno. Yes warning for Teensy. IDE 1.82.

Anyway, the warnings are just sprint() complaining. They're unrelated to Bulldog's technique for the interrupt attachment.

[Whandall]

Bulldog's original code compiled for me too. I was just wondering if my use of casting:
(void (*)())(&fan2->handleInterrupt)
would function properly, not just compile without error.
The warnings depend on what processor I try to compile for. No warnings for Uno. Yes warning for Teensy. IDE 1.82.

No, your method can not work.

I got the warnings for a Mega.

[BulldogLowell]

I got the warnings for a Mega.

sprintf() warnings, as pointed out

[Whandall]

Still errors (on 8-bit Arduinos). The second %d should print the high part of the first.

[septillion]

Are there Arduino environments in which that matters?

Both variants compile (for a Mega and ESP32) without warnings.

The first notation will be at least legal for some years to come IMHO.

Might be, depends on the function you try to call.
have a look at https://www.cprogramming.com/c++11/c++11-nullptr-strongly-typed-enum-class.html

And its part of C++11 so it will be valid for a long long time as well

[sdmetzger]

Bulldog Lowell's answer #11 is still relevant today!   I stumbled on this while trying to encapsulate my stepper motor controller routines in a class for use on an ESP32.   I have found a lot of hackarounds and avoidance of the issue by doing the ISR callbacks outside a class because a member routine's address is kind-of tough to get to, but his solution did the trick.  Kudos.