[Fix #9962] Update Style/WordArray to register an offense in percent style if any values contain spaces

changelog/change_update_stylewordarray_to_register_an.md

@@ -0,0 +1 @@
+* [#9962](https://github.com/rubocop/rubocop/issues/9962): Update `Style/WordArray` to register an offense in `percent` style if any values contain spaces. ([@dvandersluis][])

config/default.yml

@@ -4921,7 +4921,7 @@ Style/WordArray:
   StyleGuide: '#percent-w'
   Enabled: true
   VersionAdded: '0.9'
-  VersionChanged: '0.36'
+  VersionChanged: '<<next>>'
   EnforcedStyle: percent
   SupportedStyles:
     # percent style: %w(word1 word2)

lib/rubocop/cop/mixin/percent_array.rb

@@ -18,15 +18,16 @@ def invalid_percent_array_context?(node)
           !parent.parenthesized? && parent&.block_literal?
       end
 
+      # Override to determine values that are invalid in a percent array
+      def invalid_percent_array_contents?(_node)
+        false
+      end
+
       def allowed_bracket_array?(node)
         comments_in_array?(node) || below_array_length?(node) ||
           invalid_percent_array_context?(node)
       end
 
-      def message(_node)
-        style == :percent ? self.class::PERCENT_MSG : self.class::ARRAY_MSG
-      end
-
       def comments_in_array?(node)
         line_span = node.source_range.first_line...node.source_range.last_line
         processed_source.each_comment_in_lines(line_span).any?
@@ -35,9 +36,11 @@ def comments_in_array?(node)
       def check_percent_array(node)
         array_style_detected(:percent, node.values.size)
 
-        return unless style == :brackets
+        return unless style == :brackets || invalid_percent_array_contents?(node)
 
-        add_offense(node) { |corrector| correct_bracketed(corrector, node) }
+        add_offense(node, message: self.class::ARRAY_MSG) do |corrector|
+          correct_bracketed(corrector, node)
+        end
       end
 
       def check_bracketed_array(node, literal_prefix)
@@ -47,7 +50,7 @@ def check_bracketed_array(node, literal_prefix)
 
         return unless style == :percent
 
-        add_offense(node) do |corrector|
+        add_offense(node, message: self.class::PERCENT_MSG) do |corrector|
           percent_literal_corrector = PercentLiteralCorrector.new(@config, @preferred_delimiters)
           percent_literal_corrector.correct(corrector, node, literal_prefix)
         end

lib/rubocop/cop/style/word_array.rb

@@ -9,6 +9,9 @@ module Style
       # Alternatively, it can check for uses of the %w() syntax, in projects
       # which do not want to include that syntax.
       #
+      # NOTE: When using the `percent` style, %w() arrays containing a space
+      # will be registered as offenses.
+      #
       # Configuration option: MinSize
       # If set, arrays with fewer elements than this value will not trigger the
       # cop. For example, a `MinSize` of `3` will not enforce a style on an
@@ -21,12 +24,18 @@ module Style
       #   # bad
       #   ['foo', 'bar', 'baz']
       #
+      #   # bad (contains spaces)
+      #   %w[foo\ bar baz\ quux]
+      #
       # @example EnforcedStyle: brackets
       #   # good
       #   ['foo', 'bar', 'baz']
       #
       #   # bad
       #   %w[foo bar baz]
+      #
+      #   # good (contains spaces)
+      #   ['foo bar', 'baz quux']
       class WordArray < Base
         include ArrayMinSize
         include ArraySyntax
@@ -53,13 +62,22 @@ def on_array(node)
 
         private
 
-        def complex_content?(strings)
+        def complex_content?(strings, complex_regex: word_regex)
           strings.any? do |s|
+            next unless s.str_content
+
             string = s.str_content.dup.force_encoding(::Encoding::UTF_8)
-            !string.valid_encoding? || !word_regex.match?(string) || / /.match?(string)
+            !string.valid_encoding? ||
+              (complex_regex && !complex_regex.match?(string)) ||
+              / /.match?(string)
           end
         end
 
+        def invalid_percent_array_contents?(node)
+          # Disallow %w() arrays that contain invalid encoding or spaces
+          complex_content?(node.values, complex_regex: false)
+        end
+
         def word_regex
           Regexp.new(cop_config['WordRegex'])
         end

spec/rubocop/cop/style/word_array_spec.rb

@@ -317,6 +317,34 @@
       _investigate(cop2, parse_source('%w(a b c)'))
       expect(cop2.config_to_allow_offenses).to eq('EnforcedStyle' => 'percent', 'MinSize' => 3)
     end
+
+    it 'registers an offense for a %w() array containing spaces' do
+      expect_offense(<<~'RUBY')
+        %w(one\ two three\ four)
+        ^^^^^^^^^^^^^^^^^^^^^^^^ Use `[]` for an array of words.
+      RUBY
+
+      expect_correction(<<~RUBY)
+        ['one two', 'three four']
+      RUBY
+    end
+
+    it 'does not register an offense for a %w() array containing non word characters' do
+      expect_no_offenses(<<~RUBY)
+        %w(% %i %I %q %Q %r %s %w %W %x)
+      RUBY
+    end
+
+    it 'corrects properly when there is an extra trailing comma' do
+      expect_offense(<<~RUBY)
+        A = ["one", "two",]
+            ^^^^^^^^^^^^^^^ Use `%w` or `%W` for an array of words.
+      RUBY
+
+      expect_correction(<<~RUBY)
+        A = %w(one two)
+      RUBY
+    end
   end
 
   context 'when EnforcedStyle is array' do
@@ -338,6 +366,10 @@
       expect_no_offenses("['foo', 'bar', 'foo-bar']")
     end
 
+    it 'does not register an offense for arrays of strings with spaces' do
+      expect_no_offenses("['foo bar', 'baz quux']")
+    end
+
     it 'registers an offense for a %w() array' do
       expect_offense(<<~RUBY)
         %w(one two three)