Jack Grigg

There are other places we also need callback functionality (e.g. my [WIP thread wrapping PR](https://github.com/flipperzero-rs/flipperzero/pull/49/files#diff-7e18767e7236ac3bfd2d79c9bdc557bcc534e00f0e7bbd0c107b5b6afff95e8bR64-R79), where I box twice to get a thin pointer), so we should probably standardise on...

My without-review inclination is to merge this PR as-is and make further improvements in subsequent PRs. I'm going to attempt to review the PR sometime in the next week or...

Re: the callback approach, I think we should work with what is implemented now, and then a subsequent PR could completely change how the callbacks are implemented. We don't need...

> By the way, I suggest that we use root crate (the on with workspaces) to host `flipperzero` crate since this would simplify project structure a bit also allowing us...

Re: this PR, the summary indicates that there are a lot of changes that are maybe used by the GUI changes, but are themselves unrelated to the GUI. To keep...

Yep. People could still use one of those crates, but I believe it would get reported as a memory leak by the FZ tooling.

We can't use `once_cell` because it requires `std` for `Mutex`. I'm instead planning on using `generic_once_cell`, which works for any mutex that implements `lock_api::RawMutex`. I've implemented the latter locally, and...

That's a very clever way of going about it! It's also probably not necessary before upstream reaches 1.0, because they are making changes quite frequently at the moment, and I'm...

> That's a very clever way of going about it! It's also probably not necessary before upstream reaches 1.0, because they are making changes quite frequently at the moment, and...

Using $n = 21$ gives a data complexity of $2^{126}$ for the distinguisher, which is close enough to the 128-bit security level that I think it is sufficiently close to...