coderay 0.8.357 → 0.9.1

data/lib/coderay/helpers/plugin.rb

@@ -176,7 +176,7 @@ module PluginHost
   def inspect
     map = plugin_hash.dup
     map.each do |id, plugin|
-      map[id] = plugin.to_s[/(?>[\w_]+)$/]
+      map[id] = plugin.to_s[/(?>\w+)$/]
     end
     "#{name}[#{host_id}]#{map.inspect}"
   end
@@ -280,6 +280,16 @@ module Plugin
   def register_for *ids
     plugin_host.register self, *ids
   end
+  
+  # Returns the title of the plugin, or sets it to the
+  # optional argument +title+.
+  def title title = nil
+    if title
+      @title = title.to_s
+    else
+      @title ||= name[/([^:]+)$/, 1]
+    end
+  end
 
   # The host for this Plugin class.
   def plugin_host host = nil
@@ -301,15 +311,23 @@ module Plugin
   #
   # The above example loads the file myplugin/my_helper.rb relative to the
   # file in which MyPlugin was defined.
+  # 
+  # You can also load a helper from a different plugin:
+  # 
+  #  helper 'other_plugin/helper_name'
   def helper *helpers
     for helper in helpers
-      self::PLUGIN_HOST.require_helper plugin_id, helper.to_s
+      if helper.is_a?(String) && helper[/\//]
+        self::PLUGIN_HOST.require_helper $`, $'
+      else
+        self::PLUGIN_HOST.require_helper plugin_id, helper.to_s
+      end
     end
   end
 
   # Returns the pulgin id used by the engine.
   def plugin_id
-    name[/[\w_]+$/].downcase
+    name[/\w+$/].downcase
   end
 
 end

data/lib/coderay/helpers/word_list.rb

@@ -98,15 +98,22 @@ class CaseIgnoringWordList < WordList
   # Creates a new case-insensitive WordList with +default+ as default value.
   # 
   # You can activate caching to store the results for every [] request.
+  # This speeds up subsequent lookups for the same word, but also
+  # uses memory.
   def initialize default = false, caching = false
     if caching
       super(default, false) do |h, k|
         h[k] = h.fetch k.downcase, default
       end
     else
-      def self.[] key  # :nodoc:
-        super(key.downcase)
-      end
+      super(default, false)
+      extend Uncached
+    end
+  end
+  
+  module Uncached  # :nodoc:
+    def [] key
+      super(key.downcase)
     end
   end
 
@@ -120,4 +127,12 @@ class CaseIgnoringWordList < WordList
 
 end
 
-end
+end
+
+__END__
+# check memory consumption
+END {
+  ObjectSpace.each_object(CodeRay::CaseIgnoringWordList) do |wl|
+    p wl.inject(0) { |memo, key, value| memo + key.size + 24 }
+  end
+}

data/lib/coderay/scanner.rb

@@ -43,6 +43,7 @@ module CodeRay
     # You can also use +map+, +any?+, +find+ and even +sort_by+,
     # if you want.
     class Scanner < StringScanner
+      
       extend Plugin
       plugin_host Scanners
 
@@ -55,6 +56,8 @@ module CodeRay
       #
       # Define @default_options for subclasses.
       DEFAULT_OPTIONS = { :stream => false }
+      
+      KINDS_NOT_LOC = [:comment, :doctype]
 
       class << self
 
@@ -65,7 +68,14 @@ module CodeRay
 
         def normify code
           code = code.to_s
-          code.force_encoding 'binary' if code.respond_to? :force_encoding
+          if code.respond_to? :force_encoding
+            begin
+              code.force_encoding 'utf-8'
+              code[/\z/]  # raises an ArgumentError when code contains a non-UTF-8 char
+            rescue ArgumentError
+              code.force_encoding 'binary'
+            end
+          end
           code.to_unix
         end
         
@@ -104,9 +114,10 @@ module CodeRay
       #
       # Else, a Tokens object is used.
       def initialize code='', options = {}, &block
-        @options = self.class::DEFAULT_OPTIONS.merge options
         raise "I am only the basic Scanner class. I can't scan "\
           "anything. :( Use my subclasses." if self.class == Scanner
+        
+        @options = self.class::DEFAULT_OPTIONS.merge options
 
         super Scanner.normify(code)
 
@@ -121,6 +132,7 @@ module CodeRay
             "but :stream is #{@options[:stream]}" if block_given?
           @tokens ||= Tokens.new
         end
+        @tokens.scanner = self
 
         setup
       end
@@ -140,6 +152,11 @@ module CodeRay
       alias code string
       alias code= string=
 
+      # Returns the Plugin ID for this scanner.
+      def lang
+        self.class.plugin_id
+      end
+
       # Scans the code and returns all tokens in a Tokens object.
       def tokenize new_string=nil, options = {}
         options = @options.merge(options)
@@ -181,8 +198,21 @@ module CodeRay
       
       def column pos = self.pos
         return 0 if pos <= 0
+        string = string()
+        if string.respond_to?(:bytesize) && (defined?(@bin_string) || string.bytesize != string.size)
+          @bin_string ||= string.dup.force_encoding(:binary)
+          string = @bin_string
+        end
         pos - (string.rindex(?\n, pos) || 0)
       end
+      
+      def marshal_dump
+        @options
+      end
+      
+      def marshal_load options
+        @options = options
+      end
 
     protected
 
@@ -207,6 +237,7 @@ module CodeRay
       def reset_instance
         @tokens.clear unless @options[:keep_tokens]
         @cached_tokens = nil
+        @bin_string = nil if defined? @bin_string
       end
 
       # Scanner error with additional status information

data/lib/coderay/scanners/_map.rb

@@ -1,14 +1,20 @@
 module CodeRay
 module Scanners
 
-  map :cpp => :c,
-    :plain => :plaintext,
-    :pascal => :delphi,
+  map \
+    :h => :c,
+    :cplusplus => :cpp,
+    :'c++' => :cpp,
+    :ecma => :java_script,
+    :ecmascript => :java_script,
+    :ecma_script => :java_script,
     :irb => :ruby,
-    :xhtml => :nitro_xhtml,
     :javascript => :java_script,
     :js => :java_script,
     :nitro => :nitro_xhtml,
+    :pascal => :delphi,
+    :plain => :plaintext,
+    :xhtml => :html,
     :yml => :yaml
 
   default :plain

data/lib/coderay/scanners/c.rb

@@ -3,42 +3,50 @@ module Scanners
 
   class C < Scanner
 
-    register_for :c
-    
     include Streamable
+    
+    register_for :c
+    file_extension 'c'
 
     RESERVED_WORDS = [
-      'asm', 'break', 'case', 'continue', 'default', 'do', 'else',
-      'for', 'goto', 'if', 'return', 'switch', 'while',
-      'struct', 'union', 'enum', 'typedef',
-      'static', 'register', 'auto', 'extern',
-      'sizeof',
-      'volatile', 'const',  # C89
-      'inline', 'restrict', # C99
+      'asm', 'break', 'case', 'continue', 'default', 'do',
+      'else', 'enum', 'for', 'goto', 'if', 'return',
+      'sizeof', 'struct', 'switch', 'typedef', 'union', 'while',
+      'restrict',  # added in C99
     ]
 
     PREDEFINED_TYPES = [
-      'int', 'long', 'short', 'char', 'void',
+      'int', 'long', 'short', 'char',
       'signed', 'unsigned', 'float', 'double',
-      'bool', 'complex',  # C99
+      'bool', 'complex',  # added in C99
     ]
 
     PREDEFINED_CONSTANTS = [
       'EOF', 'NULL',
-      'true', 'false',  # C99
+      'true', 'false',  # added in C99
+    ]
+    DIRECTIVES = [
+      'auto', 'extern', 'register', 'static', 'void',
+      'const', 'volatile',  # added in C89
+      'inline',  # added in C99
     ]
 
     IDENT_KIND = WordList.new(:ident).
       add(RESERVED_WORDS, :reserved).
       add(PREDEFINED_TYPES, :pre_type).
+      add(DIRECTIVES, :directive).
       add(PREDEFINED_CONSTANTS, :pre_constant)
 
-    ESCAPE = / [rbfnrtv\n\\'"] | x[a-fA-F0-9]{1,2} | [0-7]{1,3} /x
+    ESCAPE = / [rbfntv\n\\'"] | x[a-fA-F0-9]{1,2} | [0-7]{1,3} /x
     UNICODE_ESCAPE =  / u[a-fA-F0-9]{4} | U[a-fA-F0-9]{8} /x
 
     def scan_tokens tokens, options
 
       state = :initial
+      label_expected = true
+      case_expected = false
+      label_expected_before_preproc_line = nil
+      in_preproc_line = false
 
       until eos?
 
@@ -49,8 +57,13 @@ module Scanners
 
         when :initial
 
-          if scan(/ \s+ | \\\n /x)
-            kind = :space
+          if match = scan(/ \s+ | \\\n /x)
+            if in_preproc_line && match != "\\\n" && match.index(?\n)
+              in_preproc_line = false
+              label_expected = label_expected_before_preproc_line
+            end
+            tokens << [match, :space]
+            next
 
           elsif scan(%r! // [^\n\\]* (?: \\. [^\n\\]* )* | /\* (?: .*? \*/ | .* ) !mx)
             kind = :comment
@@ -59,16 +72,32 @@ module Scanners
             match << scan_until(/ ^\# (?:elif|else|endif) .*? $ | \z /xm) unless eos?
             kind = :comment
 
-          elsif scan(/ [-+*\/=<>?:;,!&^|()\[\]{}~%]+ | \.(?!\d) /x)
+          elsif match = scan(/ [-+*=<>?:;,!&^|()\[\]{}~%]+ | \/=? | \.(?!\d) /x)
+            label_expected = match =~ /[;\{\}]/
+            if case_expected
+              label_expected = true if match == ':'
+              case_expected = false
+            end
             kind = :operator
 
           elsif match = scan(/ [A-Za-z_][A-Za-z_0-9]* /x)
             kind = IDENT_KIND[match]
-            if kind == :ident and check(/:(?!:)/)
-              match << scan(/:/)
+            if kind == :ident && label_expected && !in_preproc_line && scan(/:(?!:)/)
               kind = :label
+              match << matched
+            else
+              label_expected = false
+              if kind == :reserved
+                case match
+                when 'case', 'default'
+                  case_expected = true
+                end
+              end
             end
 
+          elsif scan(/\$/)
+            kind = :ident
+          
           elsif match = scan(/L?"/)
             tokens << [:open, :string]
             if match[0] == ?L
@@ -78,23 +107,30 @@ module Scanners
             state = :string
             kind = :delimiter
 
-          elsif scan(/#\s*(\w*)/)
-            kind = :preprocessor  # FIXME multiline preprocs
+          elsif scan(/#[ \t]*(\w*)/)
+            kind = :preprocessor
+            in_preproc_line = true
+            label_expected_before_preproc_line = label_expected
             state = :include_expected if self[1] == 'include'
 
           elsif scan(/ L?' (?: [^\'\n\\] | \\ #{ESCAPE} )? '? /ox)
+            label_expected = false
             kind = :char
 
           elsif scan(/0[xX][0-9A-Fa-f]+/)
+            label_expected = false
             kind = :hex
 
           elsif scan(/(?:0[0-7]+)(?![89.eEfF])/)
+            label_expected = false
             kind = :oct
 
-          elsif scan(/(?:\d+)(?![.eEfF])/)
+          elsif scan(/(?:\d+)(?![.eEfF])L?L?/)
+            label_expected = false
             kind = :integer
 
           elsif scan(/\d[fF]?|\d*\.\d+(?:[eE][+-]?\d+)?[fF]?|\d+[eE][+-]?\d+[fF]?/)
+            label_expected = false
             kind = :float
 
           else
@@ -110,6 +146,7 @@ module Scanners
             tokens << ['"', :delimiter]
             tokens << [:close, :string]
             state = :initial
+            label_expected = false
             next
           elsif scan(/ \\ (?: #{ESCAPE} | #{UNICODE_ESCAPE} ) /mox)
             kind = :char
@@ -117,6 +154,7 @@ module Scanners
             tokens << [:close, :string]
             kind = :error
             state = :initial
+            label_expected = false
           else
             raise_inspect "else case \" reached; %p not handled." % peek(1), tokens
           end
@@ -131,8 +169,8 @@ module Scanners
             state = :initial if match.index ?\n
 
           else
-            getch
-            kind = :error
+            state = :initial
+            next
 
           end
 

data/lib/coderay/scanners/cpp.rb

@@ -0,0 +1,228 @@
+module CodeRay
+module Scanners
+
+  class CPlusPlus < Scanner
+
+    include Streamable
+    
+    register_for :cpp
+    file_extension 'cpp'
+    title 'C++'
+    
+    # http://www.cppreference.com/wiki/keywords/start
+    RESERVED_WORDS = [
+      'and', 'and_eq', 'asm', 'bitand', 'bitor', 'break',
+      'case', 'catch', 'class', 'compl', 'const_cast',
+      'continue', 'default', 'delete', 'do', 'dynamic_cast', 'else',
+      'enum', 'export', 'for', 'goto', 'if', 'namespace', 'new',
+      'not', 'not_eq', 'or', 'or_eq', 'reinterpret_cast', 'return',
+      'sizeof', 'static_cast', 'struct', 'switch', 'template',
+      'throw', 'try', 'typedef', 'typeid', 'typename', 'union',
+      'while', 'xor', 'xor_eq'
+    ]
+
+    PREDEFINED_TYPES = [
+      'bool', 'char', 'double', 'float', 'int', 'long',
+      'short', 'signed', 'unsigned', 'wchar_t', 'string'
+    ]
+    PREDEFINED_CONSTANTS = [
+      'false', 'true',
+      'EOF', 'NULL',
+    ]
+    PREDEFINED_VARIABLES = [
+      'this'
+    ]
+    DIRECTIVES = [
+      'auto', 'const', 'explicit', 'extern', 'friend', 'inline', 'mutable', 'operator',
+      'private', 'protected', 'public', 'register', 'static', 'using', 'virtual', 'void',
+      'volatile'
+    ]
+
+    IDENT_KIND = WordList.new(:ident).
+      add(RESERVED_WORDS, :reserved).
+      add(PREDEFINED_TYPES, :pre_type).
+      add(PREDEFINED_VARIABLES, :local_variable).
+      add(DIRECTIVES, :directive).
+      add(PREDEFINED_CONSTANTS, :pre_constant)
+
+    ESCAPE = / [rbfntv\n\\'"] | x[a-fA-F0-9]{1,2} | [0-7]{1,3} /x
+    UNICODE_ESCAPE =  / u[a-fA-F0-9]{4} | U[a-fA-F0-9]{8} /x
+
+    def scan_tokens tokens, options
+
+      state = :initial
+      label_expected = true
+      case_expected = false
+      label_expected_before_preproc_line = nil
+      in_preproc_line = false
+
+      until eos?
+
+        kind = nil
+        match = nil
+        
+        case state
+
+        when :initial
+
+          if match = scan(/ \s+ | \\\n /x)
+            if in_preproc_line && match != "\\\n" && match.index(?\n)
+              in_preproc_line = false
+              label_expected = label_expected_before_preproc_line
+            end
+            tokens << [match, :space]
+            next
+
+          elsif scan(%r! // [^\n\\]* (?: \\. [^\n\\]* )* | /\* (?: .*? \*/ | .* ) !mx)
+            kind = :comment
+
+          elsif match = scan(/ \# \s* if \s* 0 /x)
+            match << scan_until(/ ^\# (?:elif|else|endif) .*? $ | \z /xm) unless eos?
+            kind = :comment
+
+          elsif match = scan(/ [-+*=<>?:;,!&^|()\[\]{}~%]+ | \/=? | \.(?!\d) /x)
+            label_expected = match =~ /[;\{\}]/
+            if case_expected
+              label_expected = true if match == ':'
+              case_expected = false
+            end
+            kind = :operator
+
+          elsif match = scan(/ [A-Za-z_][A-Za-z_0-9]* /x)
+            kind = IDENT_KIND[match]
+            if kind == :ident && label_expected && !in_preproc_line && scan(/:(?!:)/)
+              kind = :label
+              match << matched
+            else
+              label_expected = false
+              if kind == :reserved
+                case match
+                when 'class'
+                  state = :class_name_expected
+                when 'case', 'default'
+                  case_expected = true
+                end
+              end
+            end
+
+          elsif scan(/\$/)
+            kind = :ident
+          
+          elsif match = scan(/L?"/)
+            tokens << [:open, :string]
+            if match[0] == ?L
+              tokens << ['L', :modifier]
+              match = '"'
+            end
+            state = :string
+            kind = :delimiter
+
+          elsif scan(/#[ \t]*(\w*)/)
+            kind = :preprocessor
+            in_preproc_line = true
+            label_expected_before_preproc_line = label_expected
+            state = :include_expected if self[1] == 'include'
+
+          elsif scan(/ L?' (?: [^\'\n\\] | \\ #{ESCAPE} )? '? /ox)
+            label_expected = false
+            kind = :char
+
+          elsif scan(/0[xX][0-9A-Fa-f]+/)
+            label_expected = false
+            kind = :hex
+
+          elsif scan(/(?:0[0-7]+)(?![89.eEfF])/)
+            label_expected = false
+            kind = :oct
+
+          elsif scan(/(?:\d+)(?![.eEfF])L?L?/)
+            label_expected = false
+            kind = :integer
+
+          elsif scan(/\d[fF]?|\d*\.\d+(?:[eE][+-]?\d+)?[fF]?|\d+[eE][+-]?\d+[fF]?/)
+            label_expected = false
+            kind = :float
+
+          else
+            getch
+            kind = :error
+
+          end
+
+        when :string
+          if scan(/[^\\"]+/)
+            kind = :content
+          elsif scan(/"/)
+            tokens << ['"', :delimiter]
+            tokens << [:close, :string]
+            state = :initial
+            label_expected = false
+            next
+          elsif scan(/ \\ (?: #{ESCAPE} | #{UNICODE_ESCAPE} ) /mox)
+            kind = :char
+          elsif scan(/ \\ | $ /x)
+            tokens << [:close, :string]
+            kind = :error
+            state = :initial
+            label_expected = false
+          else
+            raise_inspect "else case \" reached; %p not handled." % peek(1), tokens
+          end
+
+        when :include_expected
+          if scan(/<[^>\n]+>?|"[^"\n\\]*(?:\\.[^"\n\\]*)*"?/)
+            kind = :include
+            state = :initial
+
+          elsif match = scan(/\s+/)
+            kind = :space
+            state = :initial if match.index ?\n
+
+          else
+            state = :initial
+            next
+
+          end
+        
+        when :class_name_expected
+          if scan(/ [A-Za-z_][A-Za-z_0-9]* /x)
+            kind = :class
+            state = :initial
+
+          elsif match = scan(/\s+/)
+            kind = :space
+
+          else
+            getch
+            kind = :error
+            state = :initial
+
+          end
+          
+        else
+          raise_inspect 'Unknown state', tokens
+
+        end
+
+        match ||= matched
+        if $DEBUG and not kind
+          raise_inspect 'Error token %p in line %d' %
+            [[match, kind], line], tokens
+        end
+        raise_inspect 'Empty token', tokens unless match
+
+        tokens << [match, kind]
+
+      end
+
+      if state == :string
+        tokens << [:close, :string]
+      end
+
+      tokens
+    end
+
+  end
+
+end
+end