Adding markdownlint-fix to precommit config?

[rrauenza]

Do you think it would be worth adding another hook for --fix?

I can do a pull request, but it's also a simple change I think. Just checking in first on your thoughts...

[DavidAnson]

If pre-commit supports parameters, I would suggest that, I think. Else, another hook seems reasonable.

[janosh]

See https://pre-commit.com/#config-args.

Simple as

- repo: local
  hooks:
    - id: some-hook
      args: [--first-arg, --second-arg]

[rrauenza]

I'd rather not use "local" so I can pin the repo via precommit and keep it in sync with the check mode...

[janosh]

@rrauenza I think you misunderstood. My example was just to demonstrate how to pass command line arguments to a hook. Put whatever you like for the repo.

[DavidAnson]

Are there concerns with #235 ? My limited understanding of pre-commit suggested it was a reasonable change.

[janosh]

@DavidAnson I don't see the need for #235. People who want to run with the --fix flag can can already do so like this:

  - repo: https://github.com/igorshubovych/markdownlint-cli
    rev: v0.29.0
    hooks:
      - id: markdownlint
        args: [--fix]

[DavidAnson]

And now they could do that by passing name "markdownlint-fix" and no args, right? I agree this is not necessary and I am normally against duplication, but I think it makes the idea of fixing during precommit more discoverable. I'm not expecting a lot of maintenance cost, so I'm inclined to leave it.

[DavidAnson]

Fixed by 2d2ba47

[janosh]

And now they could do that by passing name "markdownlint-fix" and no args, right? I agree this is not necessary and I am normally against duplication, but I think it makes the idea of fixing during precommit more discoverable. I'm not expecting a lot of maintenance cost, so I'm inclined to leave it.

That's right. I agree, little extra work from having two hooks.

[rrauenza]

@DavidAnson I don't see the need for #235. People who want to run with the --fix flag can can already do so like this:

  - repo: https://github.com/igorshubovych/markdownlint-cli
    rev: v0.29.0
    hooks:
      - id: markdownlint
        args: [--fix]

This doesn't work if you want to have both in the precommit file -- a fix and a check... the id's would duplicate.

I was assuming not every lint error could be fixed.

Maybe having two is redundant... I didn't look to see how --fix works when / if there are errors it can't fix.

Is this actually redundant in my precommit config?

  - repo: https://github.com/rrauenza/markdownlint-cli
    rev: 68b557874fcdd9aa19d0872180268c3b1b38053c
    hooks:
       - id: markdownlint-fix
       - id: markdownlint

[janosh]

This doesn't work if you want to have both in the precommit file -- a fix and a check... the id's would duplicate.

Why do you need both? markdownlint should complain about stuff it can't fix so it still acts like a checker in fix mode. Correct me if I'm wrong, this was just my assumption when I replied above.

[DavidAnson]

That's right, fix will do what it can and report on the things that remain. However, some people may only want to scan and not fix and so I think having both entries is reasonable so they can use the one that is best suited.