@kyp44, I'd like to solicit your help testing these examples on a pygamer if you don't mind.

@jbeaurivage I'll definitely test these when I get a chance. Do we not need to wait for this PR to be merged though?

Thanks! And nope, these examples work with the mainstream release of ws2812-spi. That PR is to help with DMA SPI transactions, whereas these examples don't need DMA to work. Just compiling with optimizations enabled.

I've just had a play with this branch on my pygamer and can confirm all the neopixel examples work (except the battery one, as I don't have a battery handy).

My memory is fuzzy but think I found the ehal SPI traits don't provide strong enough guarantees to meet the neopixel timing requirements. Since we control the SPI implementation here too, maybe that's not a problem, but longer term it might be good to use neopixels as an example of how to write a peripheral driver outside of atsamd-rs - using the TC/TCC or similar to control the timing.

Since hardware tests are conclusive, this is ready to be merged.

@jbeaurivage Sorry, I was busy with life/kids stuff all weekend, but I just tested all the examples. As @ianrrees found, they all work really well, including the battery example. My PyGamer battery happened to be low so I watched the NeoPixels meter go up as I let it charge. The button example was especially neat!

Once the PR over at ws2812-spi-rs is merged, I think it would be great to add or convert an example to use DMA. I re-incorporated neopixels into my stress test example and, interestingly, even without using DMA, the neopixels now seem to work much better even when there are a lot of interrupts going on, though I still observed the colors glitching from time to time, presumably when the SPI write gets interrupted. I am hoping using DMA will help with this. In any case, this is generally a big improvement over what we had before so kudos!