Josh Goebel

My first thought was to slap #ifdef DEBUG around the calls but then I saw the only place that file is used is by one of the test programs... could...

What about this? ``` unsigned long start_time = 0; unsigned long end_time = 0; void hexbright::set_light(int start_level, int end_level, long time) { int current_level = get_light_level(); start_light_level = start_level ==...

> The main advantage with the current code is that it should be fairly obvious when the amount of user code is high enough to cause accelerometer sample loss. How...

I need to do some testing. I just got my first Hexbright. Do you know the typical cost of the update() routine itself off the top of your head? IE,...

Have you thought about moving the accelerometer stuff to a timer interrupt, since it seems to be the most time critical code? Then you'd never miss a 120hz tick and...

Or isn't there and accelerometer interrupt also? I haven't gotten around to looking into interrupts yet.

If we have those functions now isn't this completed?

Well, it's only a replacement if all the pins and values written are constants as determined by the compiler and therefor the dynamic code paths can be optimized out. So...

I also would be curious to know this. (though I'd guess this type of support [if possible] would be much slow than native regex engine)

Will add it to the frame loop after https://github.com/MLXXXp/Arduboy2/pull/2 is merged.