Merge pull request #1982 from rubocop/improve-cwc

Change `RSpec/ContextWording` cop to always report an offense when both `Prefixes` and `AllowedPatterns` are empty

CHANGELOG.md

@@ -2,6 +2,8 @@
 
 ## Master (Unreleased)
 
+- Change `RSpec/ContextWording` cop to always report an offense when both `Prefixes` and `AllowedPatterns` are empty. ([@ydah])
+
 ## 3.1.0 (2024-10-01)
 
 - Add `RSpec/StringAsInstanceDoubleConstant` to check for and correct strings used as instance_doubles. ([@corsonknowles])

docs/modules/ROOT/pages/cops_rspec.adoc

@@ -720,6 +720,9 @@ the configuration to meet project needs. Other acceptable prefixes may
 include `if`, `unless`, `for`, `before`, `after`, or `during`.
 They may consist of multiple words if desired.
 
+If both `Prefixes` and `AllowedPatterns` are empty, this cop will always
+report an offense. So you need to set at least one of them.
+
 This cop can be customized allowed context description pattern
 with `AllowedPatterns`. By default, there are no checking by pattern.
 

lib/rubocop/cop/rspec/context_wording.rb

@@ -12,6 +12,9 @@ module RSpec
       #
       # @see http://www.betterspecs.org/#contexts
       #
+      # If both `Prefixes` and `AllowedPatterns` are empty, this cop will always
+      # report an offense. So you need to set at least one of them.
+      #
       # @example `Prefixes` configuration
       #   # .rubocop.yml
       #   # RSpec/ContextWording:
@@ -58,7 +61,9 @@ module RSpec
       class ContextWording < Base
         include AllowedPattern
 
-        MSG = 'Context description should match %<patterns>s.'
+        MSG_MATCH = 'Context description should match %<patterns>s.'
+        MSG_ALWAYS = 'Current settings will always report an offense. Please ' \
+                     'add allowed words to `Prefixes` or `AllowedPatterns`.'
 
         # @!method context_wording(node)
         def_node_matcher :context_wording, <<~PATTERN
@@ -67,8 +72,7 @@ class ContextWording < Base
 
         def on_block(node) # rubocop:disable InternalAffairs/NumblockHandler
           context_wording(node) do |context|
-            if bad_pattern?(context)
-              message = format(MSG, patterns: expect_patterns)
+            unless matches_allowed_pattern?(description(context))
               add_offense(context, message: message)
             end
           end
@@ -84,12 +88,6 @@ def prefix_regexes
           @prefix_regexes ||= prefixes.map { |pre| /^#{Regexp.escape(pre)}\b/ }
         end
 
-        def bad_pattern?(node)
-          return false if allowed_patterns.empty?
-
-          !matches_allowed_pattern?(description(node))
-        end
-
         def description(context)
           if context.xstr_type?
             context.value.value
@@ -98,6 +96,14 @@ def description(context)
           end
         end
 
+        def message
+          if allowed_patterns.empty?
+            MSG_ALWAYS
+          else
+            format(MSG_MATCH, patterns: expect_patterns)
+          end
+        end
+
         def expect_patterns
           inspected = allowed_patterns.map do |pattern|
             pattern.inspect.gsub(/\A"|"\z/, '/')

spec/rubocop/cop/rspec/context_wording_spec.rb

@@ -220,9 +220,10 @@
       { 'Prefixes' => [], 'AllowedPatterns' => [] }
     end
 
-    it 'skips any description' do
-      expect_no_offenses(<<~RUBY)
-        context 'arbitrary text' do
+    it 'always registers an offense' do
+      expect_offense(<<~RUBY)
+        context 'this is an incorrect context' do
+                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Current settings will always report an offense. Please add allowed words to `Prefixes` or `AllowedPatterns`.
         end
       RUBY
     end