Wren Etheredge

Got it switched over to a typable command. I've named it `:character` (aliased to `:char`) since it should support any encoding, even if it's not ASCII compatible.

Rebased to fix conflicts. Let me know if I should write tests, otherwise it should be ready

Thanks for the pointer on the tests. I added that and a few more cases to fully cover the possible output formats.

@archseer, sorry for the ping, ~~but is there any chance of this making it into 22.12?~~ I can rebase if that would be helpful. Oops too late there on my...

Should I resolve the conflict? Looks like it's just accepting both the test I added & a couple new ones from `master`.

> Imo the best approach would be to have the macro require that the type of the statics are ["plain old data"](https://docs.rs/bytemuck/latest/bytemuck/trait.Pod.html#safety) to ensure that all bit patterns are valid...

Seems like that should work. I can think of 3 options: 1. Have the `#[ram(uninitialized)]` macro wrap the static in a `MaybeUninit` with a safety comment documenting what had to...

Maybe `persistent`? I was thinking of an option to allow reducing the list of reset reasons that skip initialization. Perhaps `#[ram(…, persistent)]` to use the default list and `#[ram(…, persistent(CoreSw)]`...

It will now only zero persistent ram after initial boot, since, I think[^1] that's the only time the RTC ram is not preserved. IMO, this should be expanded to all...

Great. My last concern is with the edge case of an externally triggered RTC RAM-preserving reset (brown out, power glitch, jtag reset, etc.) occurring *very* early in the first boot...