`additional_code_points` for `invisible_character` is logically unsound as documented

[rgoldberg]

New Issue Checklist

• I've Updated SwiftLint to the latest version.
• I've searched for existing GitHub issues.

Bug Description

additional_code_points for invisible_character is logically unsound as documented.

Does additional_code_points contain additional code points beyond a baseline, or all the code points?

If the former, then, from the description, its default should be []; if the latter, then it should be renamed code_points.

Overall, code_points is more flexible, even if the current functionality is actually for additional code points.

If switching naming and/or functionality to code_points, documentation should become:

Disallows characters specified in `code_points` in string literals (which should be replaced by escapes).

By default, `code_points` contains the following invisible characters, as they can cause hard-to-debug issues: zero-width space (U+200B), zero-width non-joiner (U+200C), and FEFF formatting character (U+FEFF).

Environment

• SwiftLint version 0.6.3

[SimplyDanny]

It's called additional_code_points because it would be code points beyond the default set which is always active.

[rgoldberg]

Thanks for the info. Then shouldn't the default for additional_code_points be [], not [“200B”, “200C”, “FEFF”]?

[SimplyDanny]

The rule will always check for 200B, 200C and FEFF. Everything users add into additional_code_points will be checked in addition, so [“200B”, “200C”, “FEFF”] + additional_code_points will be the effective list of characters to lint against.

[rgoldberg]

Sorry, maybe I'm not explaining my current issue properly:

The current documented default (I don't know what the code does) for additional_code_points is ["200B", "200C", "FEFF"].

Given that the rule will always report violations for those code points, they aren't "additional" code points; they're hardcoded code points.

Specifying them as default "additional" code points is redundant (or incorrect, depending on how you view "additional") & confusing.

My suggestion is to change the documentation for the default for additional_code_points from ["200B", "200C", "FEFF"] to []. I doubt that the code must change, but you can update that, too, if it unnecessarily includes redundant code points in an array.