Chris@909: module CodeRay
Chris@909: module Scanners
Chris@909:   
Chris@909:   # Scanner for JavaScript.
Chris@909:   # 
Chris@909:   # Aliases: +ecmascript+, +ecma_script+, +javascript+
Chris@909:   class JavaScript < Scanner
Chris@909:     
Chris@909:     register_for :java_script
Chris@909:     file_extension 'js'
Chris@909:     
Chris@909:     # The actual JavaScript keywords.
Chris@909:     KEYWORDS = %w[
Chris@909:       break case catch continue default delete do else
Chris@909:       finally for function if in instanceof new
Chris@909:       return switch throw try typeof var void while with
Chris@909:     ]  # :nodoc:
Chris@909:     PREDEFINED_CONSTANTS = %w[
Chris@909:       false null true undefined NaN Infinity
Chris@909:     ]  # :nodoc:
Chris@909:     
Chris@909:     MAGIC_VARIABLES = %w[ this arguments ]  # :nodoc: arguments was introduced in JavaScript 1.4
Chris@909:     
Chris@909:     KEYWORDS_EXPECTING_VALUE = WordList.new.add %w[
Chris@909:       case delete in instanceof new return throw typeof with
Chris@909:     ]  # :nodoc:
Chris@909:     
Chris@909:     # Reserved for future use.
Chris@909:     RESERVED_WORDS = %w[
Chris@909:       abstract boolean byte char class debugger double enum export extends
Chris@909:       final float goto implements import int interface long native package
Chris@909:       private protected public short static super synchronized throws transient
Chris@909:       volatile
Chris@909:     ]  # :nodoc:
Chris@909:     
Chris@909:     IDENT_KIND = WordList.new(:ident).
Chris@909:       add(RESERVED_WORDS, :reserved).
Chris@909:       add(PREDEFINED_CONSTANTS, :predefined_constant).
Chris@909:       add(MAGIC_VARIABLES, :local_variable).
Chris@909:       add(KEYWORDS, :keyword)  # :nodoc:
Chris@909:     
Chris@909:     ESCAPE = / [bfnrtv\n\\'"] | x[a-fA-F0-9]{1,2} | [0-7]{1,3} /x  # :nodoc:
Chris@909:     UNICODE_ESCAPE =  / u[a-fA-F0-9]{4} | U[a-fA-F0-9]{8} /x  # :nodoc:
Chris@909:     REGEXP_ESCAPE =  / [bBdDsSwW] /x  # :nodoc:
Chris@909:     STRING_CONTENT_PATTERN = {
Chris@909:       "'" => /[^\\']+/,
Chris@909:       '"' => /[^\\"]+/,
Chris@909:       '/' => /[^\\\/]+/,
Chris@909:     }  # :nodoc:
Chris@909:     KEY_CHECK_PATTERN = {
Chris@909:       "'" => / (?> [^\\']* (?: \\. [^\\']* )* ) ' \s* : /mx,
Chris@909:       '"' => / (?> [^\\"]* (?: \\. [^\\"]* )* ) " \s* : /mx,
Chris@909:     }  # :nodoc:
Chris@909:     
Chris@909:   protected
Chris@909:     
Chris@909:     def scan_tokens encoder, options
Chris@909:       
Chris@909:       state = :initial
Chris@909:       string_delimiter = nil
Chris@909:       value_expected = true
Chris@909:       key_expected = false
Chris@909:       function_expected = false
Chris@909:       
Chris@909:       until eos?
Chris@909:         
Chris@909:         case state
Chris@909:           
Chris@909:         when :initial
Chris@909:           
Chris@909:           if match = scan(/ \s+ | \\\n /x)
Chris@909:             value_expected = true if !value_expected && match.index(?\n)
Chris@909:             encoder.text_token match, :space
Chris@909:             
Chris@909:           elsif match = scan(%r! // [^\n\\]* (?: \\. [^\n\\]* )* | /\* (?: .*? \*/ | .* ) !mx)
Chris@909:             value_expected = true
Chris@909:             encoder.text_token match, :comment
Chris@909:             
Chris@909:           elsif check(/\.?\d/)
Chris@909:             key_expected = value_expected = false
Chris@909:             if match = scan(/0[xX][0-9A-Fa-f]+/)
Chris@909:               encoder.text_token match, :hex
Chris@909:             elsif match = scan(/(?>0[0-7]+)(?![89.eEfF])/)
Chris@909:               encoder.text_token match, :octal
Chris@909:             elsif match = scan(/\d+[fF]|\d*\.\d+(?:[eE][+-]?\d+)?[fF]?|\d+[eE][+-]?\d+[fF]?/)
Chris@909:               encoder.text_token match, :float
Chris@909:             elsif match = scan(/\d+/)
Chris@909:               encoder.text_token match, :integer
Chris@909:             end
Chris@909:             
Chris@909:           elsif value_expected && match = scan(/<([[:alpha:]]\w*) (?: [^\/>]*\/> | .*?<\/\1>)/xim)
Chris@909:             # TODO: scan over nested tags
Chris@909:             xml_scanner.tokenize match, :tokens => encoder
Chris@909:             value_expected = false
Chris@909:             next
Chris@909:             
Chris@909:           elsif match = scan(/ [-+*=<>?:;,!&^|(\[{~%]+ | \.(?!\d) /x)
Chris@909:             value_expected = true
Chris@909:             last_operator = match[-1]
Chris@909:             key_expected = (last_operator == ?{) || (last_operator == ?,)
Chris@909:             function_expected = false
Chris@909:             encoder.text_token match, :operator
Chris@909:             
Chris@909:           elsif match = scan(/ [)\]}]+ /x)
Chris@909:             function_expected = key_expected = value_expected = false
Chris@909:             encoder.text_token match, :operator
Chris@909:             
Chris@909:           elsif match = scan(/ [$a-zA-Z_][A-Za-z_0-9$]* /x)
Chris@909:             kind = IDENT_KIND[match]
Chris@909:             value_expected = (kind == :keyword) && KEYWORDS_EXPECTING_VALUE[match]
Chris@909:             # TODO: labels
Chris@909:             if kind == :ident
Chris@909:               if match.index(?$)  # $ allowed inside an identifier
Chris@909:                 kind = :predefined
Chris@909:               elsif function_expected
Chris@909:                 kind = :function
Chris@909:               elsif check(/\s*[=:]\s*function\b/)
Chris@909:                 kind = :function
Chris@909:               elsif key_expected && check(/\s*:/)
Chris@909:                 kind = :key
Chris@909:               end
Chris@909:             end
Chris@909:             function_expected = (kind == :keyword) && (match == 'function')
Chris@909:             key_expected = false
Chris@909:             encoder.text_token match, kind
Chris@909:             
Chris@909:           elsif match = scan(/["']/)
Chris@909:             if key_expected && check(KEY_CHECK_PATTERN[match])
Chris@909:               state = :key
Chris@909:             else
Chris@909:               state = :string
Chris@909:             end
Chris@909:             encoder.begin_group state
Chris@909:             string_delimiter = match
Chris@909:             encoder.text_token match, :delimiter
Chris@909:             
Chris@909:           elsif value_expected && (match = scan(/\//))
Chris@909:             encoder.begin_group :regexp
Chris@909:             state = :regexp
Chris@909:             string_delimiter = '/'
Chris@909:             encoder.text_token match, :delimiter
Chris@909:             
Chris@909:           elsif match = scan(/ \/ /x)
Chris@909:             value_expected = true
Chris@909:             key_expected = false
Chris@909:             encoder.text_token match, :operator
Chris@909:             
Chris@909:           else
Chris@909:             encoder.text_token getch, :error
Chris@909:             
Chris@909:           end
Chris@909:           
Chris@909:         when :string, :regexp, :key
Chris@909:           if match = scan(STRING_CONTENT_PATTERN[string_delimiter])
Chris@909:             encoder.text_token match, :content
Chris@909:           elsif match = scan(/["'\/]/)
Chris@909:             encoder.text_token match, :delimiter
Chris@909:             if state == :regexp
Chris@909:               modifiers = scan(/[gim]+/)
Chris@909:               encoder.text_token modifiers, :modifier if modifiers && !modifiers.empty?
Chris@909:             end
Chris@909:             encoder.end_group state
Chris@909:             string_delimiter = nil
Chris@909:             key_expected = value_expected = false
Chris@909:             state = :initial
Chris@909:           elsif state != :regexp && (match = scan(/ \\ (?: #{ESCAPE} | #{UNICODE_ESCAPE} ) /mox))
Chris@909:             if string_delimiter == "'" && !(match == "\\\\" || match == "\\'")
Chris@909:               encoder.text_token match, :content
Chris@909:             else
Chris@909:               encoder.text_token match, :char
Chris@909:             end
Chris@909:           elsif state == :regexp && match = scan(/ \\ (?: #{ESCAPE} | #{REGEXP_ESCAPE} | #{UNICODE_ESCAPE} ) /mox)
Chris@909:             encoder.text_token match, :char
Chris@909:           elsif match = scan(/\\./m)
Chris@909:             encoder.text_token match, :content
Chris@909:           elsif match = scan(/ \\ | $ /x)
Chris@909:             encoder.end_group state
Chris@909:             encoder.text_token match, :error
Chris@909:             key_expected = value_expected = false
Chris@909:             state = :initial
Chris@909:           else
Chris@909:             raise_inspect "else case \" reached; %p not handled." % peek(1), encoder
Chris@909:           end
Chris@909:           
Chris@909:         else
Chris@909:           raise_inspect 'Unknown state', encoder
Chris@909:           
Chris@909:         end
Chris@909:         
Chris@909:       end
Chris@909:       
Chris@909:       if [:string, :regexp].include? state
Chris@909:         encoder.end_group state
Chris@909:       end
Chris@909:       
Chris@909:       encoder
Chris@909:     end
Chris@909:     
Chris@909:   protected
Chris@909:     
Chris@909:     def reset_instance
Chris@909:       super
Chris@909:       @xml_scanner.reset if defined? @xml_scanner
Chris@909:     end
Chris@909:     
Chris@909:     def xml_scanner
Chris@909:       @xml_scanner ||= CodeRay.scanner :xml, :tokens => @tokens, :keep_tokens => true, :keep_state => false
Chris@909:     end
Chris@909:     
Chris@909:   end
Chris@909:   
Chris@909: end
Chris@909: end