Andrew Hlynskyi

I think yes, but instead of using a temporary place (if it exists on a separate fs it would have a cost to move a file), I would compile to...

@maxbrunsfeld I tried to recall what I thought about of possible fixes when I saw this issue at the first time. If to use compiling in a temporary location and...

> I'd like to **not** parallelize the `parse` or `query` commands. Let's leave them simple. They don't need maximal throughput. Users who need to do parallel parsing should probably use...

@maxbrunsfeld @amaanq Would like to hear your thoughts on this proposal.

I think this PR needs to have a clear explanation why _the superseded design was rejected due to focus on portability_ and the next issues would be closed as _not...

### Additional notes Practice of migration of many grammars to pure C scanners for better portability showed that even complex scanners don't require too much of facilities and mostly require...

What is good with ms resolution is that it's easy visually see all 1ms outliers.

I think all mentioned cases with catastrophic recursion actually have an issue in a grammars' composition and instead of attempts to handle unlimited recursions by queries it's better to reorganize...

I think @maxbrunsfeld needs to register there https://docsearch.algolia.com/apply and generate own `appId` and `containerId` that will be used in the code.

Thank you for reporting the issue, I confirm that it's reproducible on a [ad32b8a](https://github.com/irevoire/tree-sitter-numbat/blob/ad32b8a08aacd0ff8f40981776c0d571682fc0c9/grammar.js) commit.