Add benchmarking

[InnocentZero]

No description provided.

[InnocentZero]

I'll be adding more, but I wanted initial feedback if any.

[kubouch]

Nice, new benchmarks are good. It would be good to comment on what each benchmark is supposed to measure. For example, the combined benchmark is supposed to give a general run-down of the parser's features.

Another thing I've wondered is that instead of having a 10x/100x/1000x copy of each benchmark, there could be some generator code that that takes the 1x version and multiplies it Nx, incrementing some number with search and replace, for example.

[InnocentZero]

I'll add the comments, sure. As for the 100x/1000x variants, I was checking out nushell's main repo and it seems that they are benchmarking it differently. Maybe we can just copy 1x file contents to the buffer and repeat it 10x/100x times to generate the testcases internally?

[kubouch]

Maybe we can just copy 1x file contents to the buffer and repeat it 10x/100x times to generate the testcases internally?

Yeah, that's what I meant. Copy the 1x version into a buffer in memory and parse that. But there needs to be some search and replace because sometimes duplicated content is not allowed (e.g., the same command defined twice in the same file, there needs to be foo1, foo2, ..., foo100 -- this is not yet implemented in this parser, but it's that way in Nushell).

[kubouch]

@InnocentZero Do you plan continuing this?

[InnocentZero]

Hi, really sorry for this getting stagnated. Personally I'd like to continue but RL is getting a bit hectic rn so I can't commit to it properly. I did try looking around for solutions, but wasn't able to write any significant code. If you want to, feel free to close the issue, and I'll open a new one when I get back to this.

[kubouch]

No worries, you can continue whenever you have time.