(v14): Make character class sorting in regex-shorthand optional or at least documented

[brettz9]

Hi,

With the latest 14.0.0, I'm getting the following false positive for regex-shorthand (besides those reported).

The rule is expecting sorting within character classes, though this is not advertised as being part of the rule in the docs. E.g., /#([^)'"]+)/ can be optimized to /#([^"')]+)/ . I think this should be either advertised as such, or, better yet, made into an option, since it is sometimes clearer to make one's own groupings without sorting, e.g., [AaQqTt].

Thanks!

[fisker]

We are using regexp-tree to optimize them since 14.0.0, unfortunately seems there is no option to avoid sorting character class https://github.com/DmitrySoshnikov/regexp-tree/blob/master/src/optimizer/README.md#transforms

I guess we can document this.

[sindresorhus]

Yes, let's document it.

[brettz9]

@fisker : Thank you for the background info.

I've filed DmitrySoshnikov/regexp-tree#199 in the hopes it may be possible to configure in the future.

[voxpelli]

I agree that this is problematic.

One part of it is that different languages sort characters in different ways. Eg. in Swedish åäö is last in the alphabet and in that order and this rule eg. wants to resort åä to äå. To makes matter even more complicated, German sorts those letters after their dot-less equivalents, so ä after a and ö after o – so one universal true rule is not really possible when sorting purely on letters, unless one wants to enforce a "true" sorting upon everyone else.

I've now disabled this rule on our projects, which is a pity, because it has lots of other nice things to it than the sorting.

[voxpelli]

I did find a way to deactivate character sorting @fisker, you can find it in #468

[fungiboletus]

May I suggest to disable character class sorting by default ? The optimisation seems unnecessary and it makes regexps less readable.