Design screen API

[thejpster]

It should be broadly like the Python API to makes things simple for new users.

See https://pythonhosted.org/sense-hat/api/.

WIP in the https://github.com/thejpster/sensehat-rs/tree/matrix_api branch.

[thejpster]

cc: @saibatizoku

[saibatizoku]

Cool. FYI - I'm still testing a basic 8x8 rotation directly on the FrameLine, and will be adding that as as another feature of the sensehat-screen crate. I'm also prototyping a scrolling-frame feature... all of this without external dependencies.

I'm all about the zen of zero-copy and keeping things light... sometimes I'm successful at it. :)

[thejpster]

Excellent! It is right that your crate be as efficient as possible, while mine is a substitute for Python usable by complete Rust novices. I'm therefore happy to put a layer over your crate which is simpler, but slower.

I have to keep my own buffer unfortunately because I need to read and write RGB triplets individually but that's no real bother.

[saibatizoku]

Agreed! That's also the reason why I didn't implement Screen::open like the Python package (by searching for the fb by name), I think that's more up your crate's alley.

[saibatizoku]

FYI, I just bumped the version for sensehat-screen, for your WIP branch.

[thejpster]

As an FYI for myself we should use https://github.com/kornelski/imgref/blob/master/README.md here.

[thejpster]

I've added text and clear methods to the SenseHat struct, as per your examples, as I needed it for a tutorial.

I have some thoughts about the sensehat-screen crate API but that shouldn't this very high-level API.

[saibatizoku]

Thanks! I didn't have time these past months to work on those methods, I'll sync with you repo.

I'd love to hear your thoughts on the API.

My process has been to write the code from the bottom up, mainly to learn Rust, that's the reason for adding structs/enums for new features. I think that I'm at a point in which I can refactor for a better user experience.