Incorrect fix suggestion for Style/RedundantRegexpCharacterClass

[patrickdavey]

I have a string where I am trying to replace some unicode chars and rubocop is saying that it's a redundant single-element character class when I don't believe it is.

In this example I am going to try to replace all a and b (unicode 0061 0062) with the letter F

 s = "aaabbbccc"
 s.gsub(/[\u{0061 0062}]/, "F") 
# this correctly outputs "FFFFFFccc" which is correct.

Rubocop however suggests I should remove the square brackets with :

Style/RedundantRegexpCharacterClass: Redundant single-element character class, `[\u{0061 0062}]` can be replaced with `\u{0061 0062}`

However, if you remove the braces you get an incorrect value

 s = "aaabbbccc"
 s.gsub(/\u{0061 0062}/, "F") 
# this incorrectly outputs "aaFbbccc" which is incorrect.

---

Expected behavior

/[\u{0061 0062}]/

should not suggest Style/RedundantRegexpCharacterClass

Actual behavior

/[\u{0061 0062}]/

does suggest Style/RedundantRegexpCharacterClass

Steps to reproduce the problem

1. create the line s = /[\u{0061 0062}]/ in a ruby file, expect no violations.

RuboCop version

1.27.0