Blaž Kristan

@MakerMatrix please report back as I can't test ATM.

I2S0 is no longer used. It is reserved for AR. On S2 there is no I2S1 so I2S0 is used instead.

Seen here: https://github.com/blazoncek/WLED/blob/esp-now-search/wled00/bus_wrapper.h Commit: blazoncek/WLED@e214bd01731313a9ff08b337a5160e345a93fd2b

Please re-base this PR to 0_15 branch.

IDK. I have very little knowledge about realtime apart from DDP implementation. If you can, do some testing and make appropriate changes if necessary. When you think it is ok...

Yes, parallel busses may be challenging as they share some buffers but that could be taken into account when specifying length of 0. In such case you could report only...

The issue arises from WLED not getting correct NTP response. Not strictly WLED issue, but network related. Ensure clear path (and low latency) for NTP responses, possibly by using alternate...

WLED requests NTP time approximately twice per day, retrying unsuccessful request every few seconds. If the NTP library fails to process retries correctly is is an issue of the library,...

`int` and `unsigned` will compile to *native* size (almost always) while providing (at minimum) 16 bit resolution. Their use in these `for()` loops has no effect on portability/compatibility but has...

> We had a major bug as we assumed our compiler Assumption was the cause of bug, not compiler. Compilers behave on a number of rules. Assumptions do not.