Remake of the Lexer

[Larsouille25]

This branch entirely remakes the lexer, now it's more clearer, clean and spans on Token are where they should go.

[Larsouille25]

Before merge make some benchmarks and try optimizing it if possible

[Larsouille25]

Benchmarks

keep in mind :

• the line "change" doesn't mean anything in this context, we only care about the actual time.
• and the benchmark is not fair because the remake can lex everything and the "actual" lexer is buggy and can miss lex some parts.

feat/remake-lexer -> 05332ea1dce5fc53e0ea8b9368b48ed1d9c9ffab before optimizations

lexer 1 + 1             time:   [175.45 ns 175.74 ns 175.96 ns]                        
                        change: [-55.168% -55.030% -54.893%] (p = 0.00 < 0.05)
                        Performance has improved.

lexer func(26 args) addition of the 26 args                                                                             
                        time:   [26.362 µs 26.380 µs 26.401 µs]
                        change: [+103.12% +106.23% +108.99%] (p = 0.00 < 0.05)
                        Performance has regressed.

main -> d9f05a8afa8ca29cca34068822e0e74528238ff5

lexer 1 + 1             time:   [327.50 ns 329.10 ns 331.03 ns]                        
                        change: [-3.1258% -1.5861% -0.0316%] (p = 0.05 < 0.05)
                        Change within noise threshold.

lexer func(26 args) addition of the 26 args                                                                             
                        time:   [6.1743 µs 6.1951 µs 6.2225 µs]
                        change: [-3.3750% -2.6583% -1.8773%] (p = 0.00 < 0.05)
                        Performance has improved.

I will for now not doing any kind of optimization(s) to focus on the implementation of a working compiler. But it needs to be optimized.