Dario Nieuwenhuis

closing because my motivation for doing this is gone. https://github.com/knurling-rs/defmt/pull/879#issuecomment-3312522725 I understand there's other motivations, like making it Just Compile on linux/windows/mac. If someone wants to explore those, feel free...

I'm not sure we should do this. The only advantage AFAICT is better codegen when the contents are inmutable. I.e. `Mutex` and not `Mutex`. However, what's the use case for...

> contains some pointer - e.g. to a register. (That would also explain why it's wrapped in a Mutex in the first place, even though it's immutable: To avoid concurrent...

done in #54

Some concerns: - Deadline scheduling is not implemented in the critical-section run queue, only in the atomics one. - I don't think we should add `cordyceps` as a dep, I'd...

> Yep, this could be addressed in a future PR, right now we've only been using this on targets with atomics I would prefer everything to land together. Reasons: -...

#4608

Implementing blocking DMA won't allow sharing an i2c bus between an async driver and a blocking driver. That's a completely separate issue, that requires a Mutex that you can either...

the ideal "solution" to me is what's implemented now, ie completely ignore reset values. What's your use case? why does it depend so much on having the reset values? in...

The API sounds good, but it's a bit of boilerplate the user would have to write. I think it'd be even better if `embassy_stm32::init` would take care of everything automatically,...