Introducing pre-commit and markdownlint-cli2 #701
Conversation
Closes #700 Adds [pre-commit](https://pre-commit.com) to the repository (and configures but doesn't yet enable [pre-commit.ci](https://pre-commit.ci)). Some basic `pre-commit` hooks are included but the meat of the linting is done via [markdownlint-cli2](https://github.com/DavidAnson/markdownlint-cli2) which is configured via `.markdownlint-cli2.yaml`. Retrospective linting of the `.md` and `.sh` files included in this PR has been undertaken (mostly line-length issues, but a typo has been corrected in `_posts/2022-12-22-git-blame.md` too). `README.md` has a new section on setting up and using `pre-commit` for this work and includes details of why retrospective linting has not been applied more widely (basically I didn't want to break any pages) and how to ignore revisions should anyone wish to undertake this task. `markdownlint-cli2` does have an option to `fix: true` which will where possible fix things automatically, but I had a go at doing this and it wasn't applying things very well (e.g. line lengths weren't being sorted). I have written a short script ([`wrap`](https://gitlab.com/nshephard/dotfiles/-/blob/master/bin/wrap)) that uses `fmt` to sensibly wrap lines for all files of a particular type at a specified length (and in parallel too) if anyone wanted to have a go at sorting things out. One thing to discuss would be whether we enable [pre-commit.ci](https://pre-commit.ci/). I didn't want to foist this upon people without consensus being reached.
|
This table is written in html with custom CSS because the website's theme does not show table borders. Should be corrected tbh. |
|
Not sure I agree with this rule.
|
One of the external images linked by Hacktoberfest 2018 was 404, so replace with image from the bottom of the page and removed the bottom image. Resolved missing alt-text in assets dir, but this directory should probably be ignored.
These are 3rd party.
Most cases where console output, so marked as text, others I attempted to specify the most suitable language.
|
Could it not be re-written as ...which makes more sense to me since its... ...and doesn't require another level of heading. |
|
Shall we merge this and use the linting going forward and leave the older posts as is? |
|
🤷♂️ |
|
It doesn't require that old code be linted unless a file is touched/modified and included in a commit since It would of course require contributors to have pre-commit installed locally though (no need to install the |
|
Any love for this? Should I spend time resolving the conflicts that have arisen? I guess the main question is whether people would use the hooks locally (which provide a much quicker feedback loop). |
Closes #852 More than just those changes though as I still have [pre-commit](https://pre-commit.com) configured and so linted the page with [Markdownlint-cli2](https://github.com/DavidAnson/markdownlint-cli2) (see #701).
Closes #700
Adds pre-commit to the repository (and configures but doesn't yet enable pre-commit.ci). Some basic
pre-commithooks are included but the meat of the linting is done via markdownlint-cli2 which is configured via.markdownlint-cli2.yaml.Retrospective linting of the
.mdand.shfiles included in this PR has been undertaken (mostly line-length issues, but a typo has been corrected in_posts/2022-12-22-git-blame.mdtoo).README.mdhas a new section on setting up and usingpre-commitfor this work and includes details of why retrospective linting has not been applied more widely (basically I didn't want to break any pages) and how to ignore revisions should anyone wish to undertake this task.markdownlint-cli2does have an option tofix: truewhich will where possible fix things automatically. I had a go at doing this and fixing all existing code withpre-commit run --all-filesbut found it wasn't applying things very well (e.g. line lengths weren't being corrected[1]) and I was very wary of messing up the site.One thing to discuss would be whether we enable pre-commit.ci. I didn't want to foist this upon people without consensus being reached.
[1] This led me to writing a short script (
wrap) that usesfmtto sensibly wrap lines for all files of a particular type at a specified length (and in parallel too) if anyone wanted to have a go at sorting things out.