Awesome! Thanks you. I was meaning to look at this for a while now, but haven't haven't been able to find the time.

Your guess is correct: We had just merged in the oldtokio-socketcan crate, which hadn't been updated in a few years. So it definitely needed an update.

I'll test this out ASAP.

I have seen that there is a std::io::{Read, Write} implementation for the sync CanSocket. Would it be beneficial to also have the corresponding AsyncRead and AsyncWrite implementation for tokio?

Futhermore I can move the tests for the tokio module to the integration tests folder like it is done for the async_io implementation. Is this ok?

This is great!

It's a breaking change and if it gets considered I might as well make my change breaking.

To that end, I've included these commits in my PR so I can base my changes off of them.

Also:

I have seen that there is a std::io::{Read, Write} implementation for the sync CanSocket. Would it be beneficial to also have the corresponding AsyncRead and AsyncWrite implementation for tokio?

This would be wonderful as it would support doing atokio::io::copy and would make my life easier.

👍

Happy to keep rebasing my changes on top of anything which happens in this PR.

I just pushed the implementation ofAsyncRead andAsyncWrite. I also added the missing implementation ofRead andWrite for the synchronousCanFdSocket as a separate commit. However, there are no tests for the sync implementation. But it is just a pass-through to the innersocket2::Socket, maybe its not necessary.

The tests for the tokio module could need some refactoring, and maybe moving to the tests folder because these are technically integration test.