Escape sequences should work the same with both backends

[aboeglin]

How it works within the JS backend and how we should make it work for the llvm backend as well:

\x[2 hexa digits]
\u[4 hexa digits]
\u{whatever amount of digits}

On the last one, we would check that the value contained inside is actually still within unicode boundaries, that is: 10FFFF as upper bound.