Open reneherrero opened 6 months ago
the nb
crate is tiny, it compiles in milliseconds. If you don't use the nb stuff, code from nb
doesnt end up at all on the final binary (the linker doesn't link unused code).
IMO the split is not worth the effort.
Then why create an embedded-hal-nb crate? Async for CAN is also in the works #585 ...
According to the stm32 bxcan project, the embedded-can crate is deprecated.
Is this the case? Understanding the state of affairs is "complicated"... even more so as far as CAN FD is concerned.
Any insight would be much appreciated,
Hi,
I noticed that the embedded-can crate is always pulling the nb crate even if it's not being used.
One option would be to add features. Another would be to refactor the code and follow the same pattern as the embedded-hal crates (better in my opinion).
Thoughts?