Consider allowing running of programatic tasks via Node API about lint-staged HOT 5 OPEN

I really like it! Do you want to create a PR for that? @iiroj wdyt?

from lint-staged.

Is the functionality essentially the same as creating a separate Node.js script file, and having that be the linter? What benefit is there for defining it directly in the config?

It sounds doable, but is a separate pipeline to the existing thing being based on spawning processes with execa.

from lint-staged.

To answer myself: one benefit is that file names would be passed directly so there's less need for escaping etc.

from lint-staged.

This might be a good use of node:worker_threads so that we can run them in parallel and capture console output similar to execa.

from lint-staged.

Since we execute the current "function config" functions before running actual tasks and also validate the output to be string | string[], the only way to handle this in a backwards-compatible way would be to return more functions:

// lint-staged.config.js
export default {
  "*.js": (stagedFiles) => {
    return () => lintJsFiles(stagedFiles)
  }
}

the syntax of the returned functions could be:

type TaskFunction = (stagedMatchedFiles: string[]) => Promise<void>

but since they're returned from inside a config function, one could of course handle the matched filenames in there as well.

what do you think?

from lint-staged.