require 'sass/scss/rx' module Sass module Script # The lexical analyzer for SassScript. # It takes a raw string and converts it to individual tokens # that are easier to parse. class Lexer include Sass::SCSS::RX # A struct containing information about an individual token. # # `type`: \[`Symbol`\] # : The type of token. # # `value`: \[`Object`\] # : The Ruby object corresponding to the value of the token. # # `source_range`: \[`Sass::Source::Range`\] # : The range in the source file in which the token appeared. # # `pos`: \[`Integer`\] # : The scanner position at which the SassScript token appeared. Token = Struct.new(:type, :value, :source_range, :pos) # The line number of the lexer's current position. # # @return [Integer] def line return @line unless @tok @tok.source_range.start_pos.line end # The number of bytes into the current line # of the lexer's current position (1-based). # # @return [Integer] def offset return @offset unless @tok @tok.source_range.start_pos.offset end # A hash from operator strings to the corresponding token types. OPERATORS = { '+' => :plus, '-' => :minus, '*' => :times, '/' => :div, '%' => :mod, '=' => :single_eq, ':' => :colon, '(' => :lparen, ')' => :rparen, ',' => :comma, 'and' => :and, 'or' => :or, 'not' => :not, '==' => :eq, '!=' => :neq, '>=' => :gte, '<=' => :lte, '>' => :gt, '<' => :lt, '#{' => :begin_interpolation, '}' => :end_interpolation, ';' => :semicolon, '{' => :lcurly, '...' => :splat, } OPERATORS_REVERSE = Sass::Util.map_hash(OPERATORS) {|k, v| [v, k]} TOKEN_NAMES = Sass::Util.map_hash(OPERATORS_REVERSE) {|k, v| [k, v.inspect]}.merge( :const => "variable (e.g. $foo)", :ident => "identifier (e.g. middle)") # A list of operator strings ordered with longer names first # so that `>` and `<` don't clobber `>=` and `<=`. OP_NAMES = OPERATORS.keys.sort_by {|o| -o.size} # A sub-list of {OP_NAMES} that only includes operators # with identifier names. IDENT_OP_NAMES = OP_NAMES.select {|k, _v| k =~ /^\w+/} PARSEABLE_NUMBER = /(?:(\d*\.\d+)|(\d+))(?:[eE]([+-]?\d+))?(#{UNIT})?/ # A hash of regular expressions that are used for tokenizing. REGULAR_EXPRESSIONS = { :whitespace => /\s+/, :comment => COMMENT, :single_line_comment => SINGLE_LINE_COMMENT, :variable => /(\$)(#{IDENT})/, :ident => /(#{IDENT})(\()?/, :number => PARSEABLE_NUMBER, :unary_minus_number => /-#{PARSEABLE_NUMBER}/, :color => HEXCOLOR, :id => /##{IDENT}/, :selector => /&/, :ident_op => /(#{Regexp.union(*IDENT_OP_NAMES.map do |s| Regexp.new(Regexp.escape(s) + "(?!#{NMCHAR}|\Z)") end)})/, :op => /(#{Regexp.union(*OP_NAMES)})/, } class << self private def string_re(open, close) /#{open}((?:\\.|\#(?!\{)|[^#{close}\\#])*)(#{close}|#\{)/m end end # A hash of regular expressions that are used for tokenizing strings. # # The key is a `[Symbol, Boolean]` pair. # The symbol represents which style of quotation to use, # while the boolean represents whether or not the string # is following an interpolated segment. STRING_REGULAR_EXPRESSIONS = { :double => { false => string_re('"', '"'), true => string_re('', '"') }, :single => { false => string_re("'", "'"), true => string_re('', "'") }, :uri => { false => /url\(#{W}(#{URLCHAR}*?)(#{W}\)|#\{)/, true => /(#{URLCHAR}*?)(#{W}\)|#\{)/ }, # Defined in https://developer.mozilla.org/en/CSS/@-moz-document as a # non-standard version of http://www.w3.org/TR/css3-conditional/ :url_prefix => { false => /url-prefix\(#{W}(#{URLCHAR}*?)(#{W}\)|#\{)/, true => /(#{URLCHAR}*?)(#{W}\)|#\{)/ }, :domain => { false => /domain\(#{W}(#{URLCHAR}*?)(#{W}\)|#\{)/, true => /(#{URLCHAR}*?)(#{W}\)|#\{)/ } } # @param str [String, StringScanner] The source text to lex # @param line [Integer] The 1-based line on which the SassScript appears. # Used for error reporting and sourcemap building # @param offset [Integer] The 1-based character (not byte) offset in the line in the source. # Used for error reporting and sourcemap building # @param options [{Symbol => Object}] An options hash; # see {file:SASS_REFERENCE.md#sass_options the Sass options documentation} def initialize(str, line, offset, options) @scanner = str.is_a?(StringScanner) ? str : Sass::Util::MultibyteStringScanner.new(str) @line = line @offset = offset @options = options @interpolation_stack = [] @prev = nil @tok = nil @next_tok = nil end # Moves the lexer forward one token. # # @return [Token] The token that was moved past def next @tok ||= read_token @tok, tok = nil, @tok @prev = tok tok end # Returns whether or not there's whitespace before the next token. # # @return [Boolean] def whitespace?(tok = @tok) if tok @scanner.string[0...tok.pos] =~ /\s\Z/ else @scanner.string[@scanner.pos, 1] =~ /^\s/ || @scanner.string[@scanner.pos - 1, 1] =~ /\s\Z/ end end # Returns the given character. # # @return [String] def char(pos = @scanner.pos) @scanner.string[pos, 1] end # Returns the next token without moving the lexer forward. # # @return [Token] The next token def peek @tok ||= read_token end # Rewinds the underlying StringScanner # to before the token returned by \{#peek}. def unpeek! return unless @tok @scanner.pos = @tok.pos @line = @tok.source_range.start_pos.line @offset = @tok.source_range.start_pos.offset end # @return [Boolean] Whether or not there's more source text to lex. def done? return if @next_tok whitespace unless after_interpolation? && !@interpolation_stack.empty? @scanner.eos? && @tok.nil? end # @return [Boolean] Whether or not the last token lexed was `:end_interpolation`. def after_interpolation? @prev && @prev.type == :end_interpolation end # Raise an error to the effect that `name` was expected in the input stream # and wasn't found. # # This calls \{#unpeek!} to rewind the scanner to immediately after # the last returned token. # # @param name [String] The name of the entity that was expected but not found # @raise [Sass::SyntaxError] def expected!(name) unpeek! Sass::SCSS::Parser.expected(@scanner, name, @line) end # Records all non-comment text the lexer consumes within the block # and returns it as a string. # # @yield A block in which text is recorded # @return [String] def str old_pos = @tok ? @tok.pos : @scanner.pos yield new_pos = @tok ? @tok.pos : @scanner.pos @scanner.string[old_pos...new_pos] end private def read_token if (tok = @next_tok) @next_tok = nil return tok end return if done? start_pos = source_position value = token return unless value type, val = value Token.new(type, val, range(start_pos), @scanner.pos - @scanner.matched_size) end def whitespace nil while scan(REGULAR_EXPRESSIONS[:whitespace]) || scan(REGULAR_EXPRESSIONS[:comment]) || scan(REGULAR_EXPRESSIONS[:single_line_comment]) end def token if after_interpolation? && (interp = @interpolation_stack.pop) interp_type, interp_value = interp if interp_type == :special_fun return special_fun_body(interp_value) else raise "[BUG]: Unknown interp_type #{interp_type}" unless interp_type == :string return string(interp_value, true) end end variable || string(:double, false) || string(:single, false) || number || id || color || selector || string(:uri, false) || raw(UNICODERANGE) || special_fun || special_val || ident_op || ident || op end def variable _variable(REGULAR_EXPRESSIONS[:variable]) end def _variable(rx) return unless scan(rx) [:const, @scanner[2]] end def ident return unless scan(REGULAR_EXPRESSIONS[:ident]) [@scanner[2] ? :funcall : :ident, @scanner[1]] end def string(re, open) line, offset = @line, @offset return unless scan(STRING_REGULAR_EXPRESSIONS[re][open]) if @scanner[0] =~ /([^\\]|^)\n/ filename = @options[:filename] Sass::Util.sass_warn <