Use string literal names rule

[Dschungelabenteuer]

Issue: #111

What Changed

• Added ESLint rule to enforce string literals to override Story names.
• This supports both the following patterns:

export const MyStory = {…, name}

const MyStory = {…, name}
export { MyStory }

Documentation and any plain english content may need to be rephrased.

Checklist

Check the ones applicable to your change:

• [X] Ran yarn update-all
• [X] Tests are updated
• [X] Documentation is updated

Change Type

Indicate the type of change your pull request is:

• [ ] maintenance
• [ ] documentation
• [ ] patch
• [X] minor
• [ ] major

[yannbf]

This is fantastic @Dschungelabenteuer ! I made a few tweaks and added a few more test cases which we need to support. Can you take a look?

[Dschungelabenteuer]

@yannbf It looks good to me, ~~should I go green?~~ I'll work something out :)

[Dschungelabenteuer]

Sorry for being that long to address these cases @yannbf, I've been a bit busy lately!

While taking a look at some other issues (including #28 and #67 you mentioned on Discord), I ended up being a bit more ambitious than just addressing the original issue and its test cases.

I definitely wanted to have a generic, easy and re-usable way of restricting validation logic to stories and their properties. I came up with some wrapper function which abstracts the whole process of "figuring out whether we're in a Story context" and exposes a very simple API to get stories and story properties. You can even filter them, which can come handy as illustrated by the use-string-literal-names rule, which specifically targets both name and storyName properties. A few notes:

• It should work with the test cases you provided (including the "spread" scenario)
• Code mentions "properties" instead of "parameters" so that it does not create any confusion with the actual parameters property of a Story.

I believe this could make writing new rules much easier, and bring in a bit of confidence about what's going on under the hood. The rule I've wrote to address the original issue now just holds in <50 LOC! However, I could not come up with any good approach on how to unit test the wrapper, which kind of upsets me!

I'd love to get your insight!