checker.rb in yard-spellcheck-0.1.1

- old
+ new

@@ -1,8 +1,7 @@
 require 'ffi/hunspell'
 require 'set'
-require 'yard'
 
 module YARD
   module Spellcheck
     #
     # Handles loading and spellchecking YARD Documentation.
@@ -11,24 +10,27 @@
 
       # YARD tags to not spellcheck.
       SKIP_TAGS = Set['example', 'since', 'see', 'api']
 
       # The Regexp to use for scanning in words.
-      WORD_REGEXP = /[^\W_][[^\W_]'-]*[^\W_]/
+      WORD_REGEXP = /[^\W_][\w'-]*[^\W_]/
 
+      # The Regexp to filter out Acronyms.
+      ACRONYM_REGEXP = /(([A-Z]\.){2,}|[A-Z]{2,})/
+
+      # The Regexp to filter out CamelCase words.
+      CAMEL_CASE_REGEXP = /([A-Z][a-z]+){2,}/
+
       # The language to spellcheck against.
       attr_accessor :lang
 
       # The words to ignore.
       attr_reader :ignore
 
       # The words to add to the dictionary.
       attr_reader :added
 
-      # The known words from the documentation.
-      attr_reader :known
-
       # The misspelled words.
       attr_reader :misspelled
 
       #
       # Initializes the spellchecker.
@@ -56,11 +58,10 @@
 
         if options[:add]
           @added += options[:add]
         end
 
-        @known = Set[]
         @misspelled = Hash.new { |hash,key| hash[key] = 0 }
       end
 
       #
       # Spellchecks the YARD Documentation.
@@ -85,18 +86,12 @@
 
         # load the YARD cache
         YARD::Registry.load!
 
         # clear any statistics from last run
-        @known.clear
         @misspelled.clear
 
-        # load known Class and Module names
-        YARD::Registry.all(:class, :module).each do |obj|
-          obj.path.split('::').each { |name| @known << name }
-        end
-
         FFI::Hunspell.dict(@lang) do |dict|
           # add user specified words
           @added.each { |word| dict.add(word) }
 
           unless names.empty?
@@ -165,10 +160,14 @@
       #
       def spellcheck(text,dict)
         typos = Set[]
 
         text.scan(WORD_REGEXP).each do |word|
-          next if (@known.include?(word) || @ignore.include?(word))
+          # ignore all underscored names
+          next if word.include?('_')
+
+          # ignore all acronyms and CamelCase words
+          next if (word =~ ACRONYM_REGEXP || word =~ CAMEL_CASE_REGEXP)
 
           if (@misspelled.has_key?(word) || !dict.valid?(word))
             @misspelled[word] += 1
 
             typos << word