Merge pull request #8206 from owst/fix_8205_false_positive_in_redunda…

…ntregexpcharacterclass

[Fix #8205] Avoid false positive in Style/RedundantRegexpCharacterClass

CHANGELOG.md

@@ -15,6 +15,7 @@
 
 * [#8195](https://github.com/rubocop-hq/rubocop/issues/8195): Fix an error for `Style/RedundantFetchBlock` when using `#fetch` with empty block. ([@koic][])
 * [#8193](https://github.com/rubocop-hq/rubocop/issues/8193): Fix a false positive for `Style/RedundantRegexpCharacterClass` when using `[\b]`. ([@owst][])
+* [#8205](https://github.com/rubocop-hq/rubocop/issues/8205): Fix a false positive for `Style/RedundantRegexpCharacterClass` when using a leading escaped `]`. ([@owst][])
 
 ## 0.86.0 (2020-06-22)
 

lib/rubocop/cop/style/redundant_regexp_character_class.rb

@@ -38,6 +38,7 @@ class RedundantRegexpCharacterClass < Cop
              [^.*+?{}()|$] |  # or one that doesn't require escaping outside the character class
              \\[upP]\{[^}]+\} # or a unicode code-point or property
             )
+            (?<!\\)           # No \-prefix (i.e. not escaped)
             \]                # Literal ]
           )
         /x.freeze

spec/rubocop/cop/style/redundant_regexp_character_class_spec.rb

@@ -43,6 +43,19 @@
     end
   end
 
+  context 'with a character class containing an escaped [' do
+    it 'registers an offense and corrects' do
+      expect_offense(<<~'RUBY')
+        foo = /[\[]/
+               ^^^^ Redundant single-element character class, `[\[]` can be replaced with `\[`.
+      RUBY
+
+      expect_correction(<<~'RUBY')
+        foo = /\[/
+      RUBY
+    end
+  end
+
   context 'with a character class containing a space meta-character' do
     it 'registers an offense and corrects' do
       expect_offense(<<~'RUBY')
@@ -163,6 +176,12 @@
     end
   end
 
+  context 'with a character class with first element an escaped ]' do
+    it 'does not register an offense' do
+      expect_no_offenses('foo = /[\])]/')
+    end
+  end
+
   context 'with a negated character class with a single element' do
     it 'does not register an offense' do
       expect_no_offenses('foo = /[^x]/')