Add instruction encoding for "C" extension

When implementing the "C" extension set, I didn't implement instruction encoding like I did for the other instruction formats.

Be sure to add all the appropriate test cases, using `encoding_test` to ensure that decoding some bytes then re-encoding them again produces the original bytes. Here's an example test:

https://github.com/michaelmelanson/riscy/blob/2cee7487e1dbc7e65c2af8c7ef9ebc211fa9bf37/isa/src/lib.rs#L155-L162

## Instruction formats

These instruction formats are documented in Chapter 12 ("C” Standard Extension for Compressed Instructions, Version 2.0) of the RISC-V specification:

- [ ] CIW
- [ ] CS
- [ ] CL
- [ ] CI
- [ ] CNOP / CB
- [ ] CA
- [ ] CJ
- [ ] CR
- [ ] CSS

Start in isa/src/lib.rs, using the encoding_test example around the referenced lines to trace how other instruction formats decode and re-encode. Consult Chapter 12 of the RISC-V specification and cover CIW, CS, CL, CI, CNOP/CB, CA, CJ, CR, and CSS. Done means the listed formats encode correctly and decoding then re-encoding test cases reproduces the original bytes.

Written by the indexing model from the issue text.