3/11: feat(grammar): add benchmarks

[juleswritescode]

Summary

• Add criterion dependency
• Add parsing benchmarks
• Add lookahead iterator benchmark

Part of stacked PRs for #629

This is PR 3/11 in the keyword completion feature stack.

[psteinroe]

Doesn't criterion support writing benchmark results to a file vs adding comments?

[juleswritescode]

Doesn't criterion support writing benchmark results to a file vs adding comments?

it per default writes benchmarks to the /target directory.

I think the recommended approach isn't to store any results, tbh, because it'll differ so much on varying machines.

What we could do is running benchmarks on the same gh-runner machine and look for regressions.

In this case, it was only meant as a POC that updating the tree-sitter tree once for every checkable keyword is feasible 👍🏻