[Feature Request] Add benchmarks for long file lexing + parsing

[Razican]

Current situation
We currently don't have any benchmark checking how fast we lex and parse big files, but our current approach of reading the whole file into memory and only when we finish start lexing, and only when we finish lexing start parsing, all in one thread is far from optimal, especially for big files.

Benchmark explanation
We need a benchmark where we load a big JS file from disk (at least 5kB) and we lex + parse it.

Note that this file must only contain syntax that Boa understands. It can be a nonsense file, though, copy-pasting the same code multiple times.

It's important that the whole process is inside of the benchmark function, including reading the file from the disk, lexing it and parsing it.