Missed bitshift optimization

[BoannWasHere]

Hello! I really love how tidy and efficient FAB LED is, but I found a missed optimization in the core loops, where it loops over the individual bits of each color byte.

For example, here, in onePortSoftwareSendBytes:

const uint8_t val = array[c];
for(int8_t b=7; b>=0; b--) {
    const bool bit = (val>>b) & 0x1;

The problem is that AVR has no instruction to shift by more than 1 bit at a time, so val >> b compiles to an additional loop that executes for every bit to be sent.

The extra loop can be eliminated by changing to:

uint8_t val = array[c];
for(uint8_t b=0; b<8; b++, val <<= 1) {
    const bool bit = val & 0x80;