Good call on the priority naming. Let's start doing that now. Although it's hard to tell what the
priority will be, so let's start in the middle.

This is a move and cleanup in preparation for doing it correctly. See the summary.

Rebase

Clearly this was possible with D7490 already and nothing really needs to be done. It can be argued that moving it to patch-recipes isn't even the right place anyway. Sorry for the noise.

The test plan does not test the supposed change, nor does it test the wildcard behavior that differs from the original PR.

Cleanup unneeded commit parent logic

I don't think you read the summary. All of these variants are re-written to appear as the first variant, ensuring that these are both greppable and linked.

Rebase

I noticed your test plan is rather complex and duplicates a lot of the logic in scripts/fetch_documentation.sh in the website repo. Consider improving it so it can take a local target instead of always pulling from the internet. This should come in a separate diff of course.

Match more cases

Add explicit check that HEAD has no changes from master

I don't think this makes sense. Even if markdown supports trailing whitespace as a line break, it doesn't mean it's a good idea. Consider:

• Not linting this will now allow accidental trailing whitespaces.
• Trailing whitespace does not render well in all IDEs/editors.
• Intentional line breaks should be obvious, which trailing whitespace are not.

Rebase

Rename the script to something more sensible

In D7476#177734, @deadalnix wrote:

I' not sure why you would revert back other changes, there were not bad per se.

Remove all the refactoring and just call the land patch script