Create initial workflow for markdown tests #64
Conversation
Let's start with markdownlint-cli: https://github.com/marketplace/actions/markdownlint-cli
Update actions branch to current dev
|
Okay this works. Syntax is fine now, the checks run through. Results are: Only README.md is scanned so far. Not sure if it breaks on the first file with errors or if my files glob does not span across sub directories. Will find that out.
You see it enforces certain style standards, string max line length, consistent list syntax etc. Most make sense, IMO, the 80 characters per line should be the limit for a documentation page that might be seen on 4k wide screen monitors is a different question. HTML line breaks automatically, so this would be something to disable via the actions config file.
The rules are default markdownlint rules: https://github.com/DavidAnson/markdownlint/blob/main/doc/Rules.md So e.g. I didn't understand what problems it has with the last three lines and MD034, so I check the list: https://github.com/DavidAnson/markdownlint/blob/main/doc/Rules.md#md034---bare-url-used Implementing this the first time will for sure raise a large number of "errors", so it should be gone through these and decided whether those make sense in our case or we want to exclude them in our case, like the 80 character per line rule (IMO). |
+ push + PR trigger seems to be doubled, use push only.
Enforce ATX header style, which means using prefixing headers with hashtags only, like: # h1 header ## h2 header ### h3 header Disable max line length rule, which defaults to 80 characters per line which is not reasonable for HTML documents, seen on 4k wide screens as well.
+ Set repo and edit URLs to match current owner and branch + Build MkDocs into a docs dir and set root dir accordingly
|
@MichaIng - The idea is excellent - before we commit, we need to make successful the checks. Let me review the rules too and start applying the changes. |
|
Jep, I'm still playing around, thinking about merging the tests together to avoid overhead and switching spellcheck to be done on the final html files instead of the initial markdown files, where it as well can be customised a bid better to e.g. exclude code. Link checking as well needs some solution to not check the same host too often and receive 429 response: raviqqe/liche#42 But there are already some invalid/outdated link finds, some actual typos and when we apply wrap links into angle brackets, those should be excluded from spellcheck, which covers most of the false alarms there. So basically we can merge before starting to solve things and then do that step by step instead. |
|
Spellcheck action: https://github.com/rojopolis/spellcheck-github-actions Link checker action: https://github.com/peter-evans/link-checker |
|
@MichaIng - to be able to work on it, I plan to merge the pull request and start focusing first on check_spelling. |
|
@fpetru |
|
@fpetru I tend to write/develop on GitHub directly, hence there are a lot of commits, back and forth, typos, syntax etc. This messes the commit history unnecessary, so having them as a single clean final commit enhances beauty and readability of the commit history. @StephanStS same for you, as long as there is no good reason to have each commit separated => "Squash and merge" Ah, and respect when the PR opener marks it as "draft" which basically means, "I'm still working on it", so do not merge until opener marked it to be ready 😉. |
Let's start with markdownlint-cli: https://github.com/marketplace/actions/markdownlint-cli