rom_funcs results output over USB instead of UART

[camrbuss]

• Added f64 tests for chips with rom_version>1
• Output results over USB Serial instead of hardware UART

[9names]

I like the idea, but just replacing the UART output with USB-UART output doesn't really improve the situation.
A better option would be to have a optional logger feature in the HAL where you can choose between uart, usb-cdc, rtt or defmt for getting textual output from your program.
That way all the examples can output data in the user's preferred format, and end-users can also opt in to get a simpler debug output.
Also: make it a separate PR to updating the example.

[jannic]

As far as I know, the default license of subscriber content published on stack overflow doesn't allow redistribution under MIT / Apache-2 license.

[camrbuss]

Oh my, I didn't realize that. How should I rectify the situation?

[jannic]

Perhaps just use https://docs.rs/heapless/0.7.16/heapless/struct.Vec.html or https://docs.rs/heapless/0.7.16/heapless/struct.String.html instead of implementing your own buffer?

[9names]

This might be a dumb question, but why are we buffering anyway?
Can't you push this straight into the SerialPort?

[camrbuss]

I don't believe usbd_serial implements write_str

Is there another way that I am not seeing?

[9names]

I was assuming it wouldn't be difficult to build a wrapper struct for usbd_serial and implement write_str on that, but I think the generics would be a pain.
Maybe it would be easier to make a PR to get usbd_serial to (conditionally) impl write_str

[9names]

Looking at rust-embedded-community/usbd-serial#16, seems this has already been proposed and declined since the blocking call would have to poll the USB device (which it can't).
So: would really require having some logger/stdin+stdout functionality, which would possibly also be part of how we could resolve #456