Thales

I would like to propose these traits for feedback: https://github.com/ra-kete/dma-poc/blob/43c148f7c0fc1389010c49c2259eb4077a754bc6/src/traits.rs They are an iteration over a [prototype](https://github.com/ra-kete/dma-poc/blob/b2330233ed20173c9ea683df82cbc70d1d44993d/src/traits.rs) proposed by @ra-kete. I added an associated type for allow implementations to be...

Thanks for the insight @adamgreig , like @ra-kete I also think that this is too microcontroller specific, so I don't think that there's a way to do it generically on...

@ra-kete ~~The problem with `[B; 2]: DmaWriteBuffer + 'static` is when you have wrapper types that implement `DmaWriteBuffer`, i.e. there is nothing enforcing that the real buffer location for the...

Yes, we can't just do ```rust unsafe impl DmaWriteBuffer for [B; 2] {} ``` But since the trait will be unsafe, the implementer will be responsible for checking this, and...

Yes, it should be doable. The firmware expects the address to read/write to be passed and the cli passes it starting from zero, but there is nothing in the design...

Why do you mention actix's context wait queue ? The wait queue will indeed block the actor from doing pretty much any other processing, that's its intention, but where on...

I might have found a clue. The future of the `WebSocketContext` isn't spawned directly on the runtime as the usual `Actor::start`, instead it's returned as a streaming `HttpResponse` to be...

@finnbear What OS are you using ? `poll_write` on `TcpStream` will eventually return an error on a severed connection, but how long it takes will vary with your OS and...

@tkaitchuck You can already define your own certificate verifiers and not use any webpki code for a while now.

@jbg They specifically mentioned that it was blocking specific implementations, it isn't. And, someone will have to "manually parse the certificate and implement their own verification logic", and with that,...