Pull C0 escape sequences into its own file

[Tyriar]

Part of #460

---

This is the plan for #460 but to apply it to CSI and OSC as well. Once we move the relevant code to TS we'll be able to use template literals to make the code very readable. Example:

result.key = '\x1b[1;' + (modifiers + 1) + 'C';

Would become something like:

result.key = `${CSI.START}1;${modifiers + 1}${CSI.CUF}`;

Or potentially:

result.key = csiCode(CSI.CUF, '1', modifiers + 1);

We can also look into having TS inline the constants: microsoft/TypeScript#3922 microsoft/TypeScript#3192

@parisk let me know what you think of the structure.

[Tyriar]

I'm beginning to reconsider this after spending some time redoing the parser in #462

Having C0 codes are definitely useful as they're nonsensical without it (eg. \x08 is backspace). CSI codes however are more difficult to name as they have multiple names, the information on them are probably best located within the InputHandler and the Parser itself. The CSI final letter is the best way to identify the behavior.

[parisk]

What if we use methods that will get parameters instead of attributes for CSI?

Or what if each code had multiple "children" and acted like a map?

Would such a thing work?

[Tyriar]

I think that's probably overkill. Check out the parser PR for what I'm thinking now, the descriptive comments and csi code names move with the implementations to InputHandler.

Also remember if it's a map they're not inlined which could cause some significant bloat when dealing with so many codes.