class RubyLexer
The generated lexer RubyLexer
Attributes
Last comment lexed, or nil
The file name / path
Additional context surrounding tokens that both the lexer and grammar use.
The current line number.
The StringScanner for this lexer.
Additional context surrounding tokens that both the lexer and grammar use.
The previous position. Only available if the :column option is on.
The StringScanner for this lexer.
The position of the start of the current line. Only available if the :column option is on.
The current lexical state.
Last token read via next_token.
Public Class Methods
# File lib/ruby_lexer.rb, line 119 def initialize _ = nil @lex_state = nil # remove one warning under $DEBUG @lex_state = EXPR_NONE self.cond = RubyParserStuff::StackState.new(:cond, $DEBUG) self.cmdarg = RubyParserStuff::StackState.new(:cmdarg, $DEBUG) self.ss = RPStringScanner.new "" reset end
Public Instance Methods
Yields on the current action.
# File lib/ruby_lexer.rex.rb, line 72 def action yield end
# File lib/ruby_lexer.rb, line 130 def arg_ambiguous self.warning "Ambiguous first argument. make sure." end
# File lib/ruby_lexer.rb, line 134 def arg_state is_after_operator? ? EXPR_ARG : EXPR_BEG end
The current column, starting at 0. Only available if the :column option is on.
# File lib/ruby_lexer.rex.rb, line 90 def column old_pos - start_of_current_line_pos end
# File lib/ruby_lexer.rb, line 1119 def d o $stderr.puts o.inspect end
# File lib/ruby_lexer.rb, line 138 def debug n raise "debug #{n}" end
# File lib/ruby_lexer_strings.rb, line 4 def eat_whitespace r = scan(/\s+/) self.lineno += r.count("\n") if r r += eat_whitespace if eos? && in_heredoc? # forces heredoc pop r end
# File lib/ruby_lexer.rb, line 142 def expr_dot? lex_state =~ EXPR_DOT end
# File lib/ruby_lexer.rb, line 146 def expr_fname? # REFACTOR lex_state =~ EXPR_FNAME end
# File lib/ruby_lexer.rb, line 150 def expr_result token, text cond.push false cmdarg.push false result EXPR_BEG, token, text end
# File lib/ruby_lexer_strings.rb, line 13 def heredoc here # ../compare/parse30.y:7678 _, term, func, _indent_max, _lineno, range = here start_line = lineno eos = term # HACK indent = func =~ STR_FUNC_INDENT self.string_buffer = [] last_line = self.ss_string[range] if range eol = last_line && last_line.end_with?("\r\n") ? "\r\n" : "\n" # HACK expand = func =~ STR_FUNC_EXPAND # TODO? p->heredoc_line_indent == -1 indent_re = indent ? "[ \t]*" : nil eos_re = /#{indent_re}#{Regexp.escape eos}(?=\r?\n|\z)/ err_msg = "can't match #{eos_re.inspect} anywhere in " maybe_pop_stack rb_compile_error err_msg if end_of_stream? if beginning_of_line? && scan(eos_re) then scan(/\r?\n|\z/) self.lineno += 1 if matched =~ /\n/ heredoc_restore self.lex_strterm = nil self.lex_state = EXPR_END return :tSTRING_END, [term, func, range] end if expand then case when scan(/#(?=\$(-.|[a-zA-Z_0-9~\*\$\?!@\/\\;,\.=:<>\"\&\`\'+]))/) then # TODO: !ISASCII # ?! see parser_peek_variable_name return :tSTRING_DVAR, matched when scan(/#(?=\@\@?[a-zA-Z_])/) then # TODO: !ISASCII return :tSTRING_DVAR, matched when scan(/#[{]/) then self.command_start = true return :tSTRING_DBEG, [matched, lineno] when scan(/#/) then string_buffer << "#" end begin # NOTE: this visibly diverges from the C code but uses tokadd_string # to stay clean. str = tokadd_string func, eol, nil rb_compile_error err_msg if str == RubyLexer::EOF if str != eol then str = string_buffer.join string_buffer.clear return result nil, :tSTRING_CONTENT, str, start_line else string_buffer << scan(/\r?\n/) self.lineno += 1 # TODO: try to remove most scan(/\n/) and friends end end until check eos_re else until check(eos_re) do string_buffer << scan(/.*(\r?\n|\z)/) self.lineno += 1 rb_compile_error err_msg if end_of_stream? end end string_content = begin s = string_buffer.join s.b.force_encoding Encoding::UTF_8 s end string_buffer.clear result nil, :tSTRING_CONTENT, string_content, start_line end
# File lib/ruby_lexer_strings.rb, line 98 def heredoc_identifier # ../compare/parse30.y:7354 token = :tSTRING_BEG func = STR_FUNC_BORING term = nil indent = nil quote = nil char_pos = nil byte_pos = nil heredoc_indent_mods = "-" heredoc_indent_mods += '\~' if ruby23plus? case when scan(/([#{heredoc_indent_mods}]?)([\'\"\`])(.*?)\2/) then mods, quote, term = match[1], match[2], match[3] char_pos = ss.charpos byte_pos = ss.pos func |= STR_FUNC_INDENT unless mods.empty? func |= STR_FUNC_DEDENT if mods == "~" func |= case quote when "\'" then STR_SQUOTE when '"' then STR_DQUOTE when "`" then token = :tXSTRING_BEG STR_XQUOTE else debug 1 end when scan(/[#{heredoc_indent_mods}]?([\'\"\`])(?!\1*\Z)/) then rb_compile_error "unterminated here document identifier" when scan(/([#{heredoc_indent_mods}]?)(#{IDENT_CHAR}+)/) then mods, term = match[1], match[2] quote = '"' char_pos = ss.charpos byte_pos = ss.pos func |= STR_FUNC_INDENT unless mods.empty? func |= STR_FUNC_DEDENT if mods == "~" func |= STR_DQUOTE else return end old_lineno = self.lineno rest_of_line = scan(/.*(?:\r?\n|\z)/) self.lineno += rest_of_line.count "\n" char_pos_end = ss.charpos - 1 range = nil range = char_pos..char_pos_end unless rest_of_line.empty? self.lex_strterm = [:heredoc, term, func, indent, old_lineno, range, byte_pos] result nil, token, quote, old_lineno end
# File lib/ruby_lexer_strings.rb, line 158 def heredoc_restore # ../compare/parse30.y:7438 _, _term, _func, _indent, lineno, range, bytepos = lex_strterm new_ss = ss.class.new self.ss_string[0..range.max] new_ss.pos = bytepos lineno_push lineno ss_push new_ss nil end
# File lib/ruby_lexer.rb, line 156 def in_fname? # REFACTOR lex_state =~ EXPR_FNAME end
# File lib/ruby_lexer.rb, line 160 def int_with_base base rb_compile_error "Invalid numeric format" if matched =~ /__/ text = matched case when text.end_with?("ri") result EXPR_NUM, :tIMAGINARY, Complex(0, Rational(text.chop.chop.to_i(base))) when text.end_with?("r") result EXPR_NUM, :tRATIONAL, Rational(text.chop.to_i(base)) when text.end_with?("i") result EXPR_NUM, :tIMAGINARY, Complex(0, text.chop.to_i(base)) else result EXPR_NUM, :tINTEGER, text.to_i(base) end end
# File lib/ruby_lexer.rb, line 176 def is_after_operator? lex_state =~ EXPR_FNAME|EXPR_DOT end
# File lib/ruby_lexer.rb, line 180 def is_arg? lex_state =~ EXPR_ARG_ANY end
# File lib/ruby_lexer.rb, line 184 def is_beg? lex_state =~ EXPR_BEG_ANY || lex_state == EXPR_LAB # yes, == EXPR_LAB end
# File lib/ruby_lexer.rb, line 188 def is_end? lex_state =~ EXPR_END_ANY end
# File lib/ruby_lexer.rb, line 192 def is_label_possible? (lex_state =~ EXPR_LABEL|EXPR_ENDFN && !cmd_state) || is_arg? end
# File lib/ruby_lexer.rb, line 196 def is_label_suffix? check(/:(?!:)/) end
# File lib/ruby_lexer.rb, line 208 def is_local_id id # maybe just make this false for now self.parser.env[id.to_sym] == :lvar # HACK: this isn't remotely right end
# File lib/ruby_lexer.rb, line 200 def is_space_arg? c = "x" is_arg? and space_seen and c !~ /\s/ end
# File lib/ruby_lexer.rb, line 204 def lambda_beginning? lpar_beg && lpar_beg == paren_nest end
# File lib/ruby_lexer.rb, line 1125 def lex_strterm= o self.old_lex_strterm= o where = caller.first.split(/:/).first(2).join(":") $stderr.puts d :lex_strterm => [o, where] end
The current line number.
# File lib/ruby_lexer.rb, line 1145 def lineno= n self.old_lineno= n where = caller.first.split(/:/).first(2).join(":") $stderr.puts d :lineno => [n, where] end
The current location in the parse.
# File lib/ruby_lexer.rex.rb, line 126 def location [ (filename || "<input>"), lineno, column, ].compact.join(":") end
# File lib/ruby_lexer.rb, line 213 def lvar_defined? id # TODO: (dyna_in_block? && dvar_defined?(id)) || local_id?(id) self.parser.env[id.to_sym] == :lvar end
The match groups for the current scan.
# File lib/ruby_lexer.rex.rb, line 63 def matches m = (1..9).map { |i| ss[i] } m.pop until m[-1] or m.empty? m end
# File lib/ruby_lexer_strings.rb, line 170 def newtok string_buffer.clear end
Lex the next token.
# File lib/ruby_lexer.rex.rb, line 137 def next_token maybe_pop_stack return process_string_or_heredoc if lex_strterm self.cmd_state = self.command_start self.command_start = false self.space_seen = false # TODO: rename token_seen? self.last_state = lex_state token = nil until ss.eos? or token do if ss.check(/\n/) then self.lineno += 1 # line starts 1 position after the newline self.start_of_current_line_pos = ss.pos + 1 end self.old_pos = ss.pos token = case state when nil then case when ss.skip(/[\ \t\r\f\v]+/) then action { self.space_seen = true; next } when text = ss.scan(/\n|\#/) then process_newline_or_comment text when text = ss.scan(/[\]\)\}]/) then process_brace_close text when ss.match?(/\!/) then case when is_after_operator? && (text = ss.scan(/\!\@/)) then action { result EXPR_ARG, TOKENS[text], text } when text = ss.scan(/\![=~]?/) then action { result :arg_state, TOKENS[text], text } end # group /\!/ when ss.match?(/\./) then case when text = ss.scan(/\.\.\.?/) then process_dots text when ss.skip(/\.\d/) then action { rb_compile_error "no .<digit> floating literal anymore put 0 before dot" } when ss.skip(/\./) then action { self.lex_state = EXPR_BEG; result EXPR_DOT, :tDOT, "." } end # group /\./ when text = ss.scan(/\(/) then process_paren text when text = ss.scan(/\,/) then action { result EXPR_PAR, TOKENS[text], text } when ss.match?(/=/) then case when text = ss.scan(/\=\=\=|\=\=|\=~|\=>|\=(?!begin\b)/) then action { result arg_state, TOKENS[text], text } when bol? && (text = ss.scan(/\=begin(?=\s)/)) then process_begin text when text = ss.scan(/\=(?=begin\b)/) then action { result arg_state, TOKENS[text], text } end # group /=/ when ruby22_label? && (text = ss.scan(/\"#{SIMPLE_STRING}\":/o)) then process_label text when text = ss.scan(/\"(#{SIMPLE_STRING})\"/o) then process_simple_string text when text = ss.scan(/\"/) then action { string STR_DQUOTE, '"'; result nil, :tSTRING_BEG, text } when text = ss.scan(/\@\@?\d/) then action { rb_compile_error "`#{text}` is not allowed as a variable name" } when text = ss.scan(/\@\@?#{IDENT_CHAR}+/o) then process_ivar text when ss.match?(/:/) then case when not_end? && (text = ss.scan(/:([a-zA-Z_]#{IDENT_CHAR}*(?:[?]|[!](?!=)|=(?==>)|=(?![=>]))?)/o)) then process_symbol text when not_end? && (text = ss.scan(/\:\"(#{SIMPLE_STRING})\"/o)) then process_symbol text when not_end? && (text = ss.scan(/\:\'(#{SSTRING})\'/o)) then process_symbol text when text = ss.scan(/\:\:/) then process_colon2 text when text = ss.scan(/\:/) then process_colon1 text end # group /:/ when text = ss.scan(/->/) then action { result EXPR_ENDFN, :tLAMBDA, text } when text = ss.scan(/[+-]/) then process_plus_minus text when ss.match?(/[+\d]/) then case when ss.skip(/#{NUM_BAD}/o) then action { rb_compile_error "Invalid numeric format" } when ss.skip(/#{INT_DEC}/o) then action { int_with_base 10 } when ss.skip(/#{INT_HEX}/o) then action { int_with_base 16 } when ss.skip(/#{INT_BIN}/o) then action { int_with_base 2 } when ss.skip(/#{INT_OCT_BAD}/o) then action { rb_compile_error "Illegal octal digit." } when ss.skip(/#{INT_OCT}/o) then action { int_with_base 8 } when ss.skip(/#{FLOAT_BAD}/o) then action { rb_compile_error "Trailing '_' in number." } when text = ss.scan(/#{FLOAT}/o) then process_float text when ss.skip(/#{INT_DEC2}/o) then action { int_with_base 10 } when ss.skip(/[0-9]/) then action { rb_compile_error "Bad number format" } end # group /[+\d]/ when text = ss.scan(/\[/) then process_square_bracket text when was_label? && (text = ss.scan(/\'#{SSTRING}\':?/o)) then process_label_or_string text when text = ss.scan(/\'/) then action { string STR_SQUOTE, "'"; result nil, :tSTRING_BEG, text } when ss.match?(/\|/) then case when ss.skip(/\|\|\=/) then action { result EXPR_BEG, :tOP_ASGN, "||" } when ss.skip(/\|\|/) then action { result EXPR_BEG, :tOROP, "||" } when ss.skip(/\|\=/) then action { result EXPR_BEG, :tOP_ASGN, "|" } when ss.skip(/\|/) then action { state = is_after_operator? ? EXPR_ARG : EXPR_PAR; result state, :tPIPE, "|" } end # group /\|/ when text = ss.scan(/\{/) then process_brace_open text when ss.match?(/\*/) then case when ss.skip(/\*\*=/) then action { result EXPR_BEG, :tOP_ASGN, "**" } when ss.skip(/\*\*/) then action { result :arg_state, space_vs_beginning(:tDSTAR, :tDSTAR, :tPOW), "**" } when ss.skip(/\*\=/) then action { result EXPR_BEG, :tOP_ASGN, "*" } when ss.skip(/\*/) then action { result :arg_state, space_vs_beginning(:tSTAR, :tSTAR, :tSTAR2), "*" } end # group /\*/ when ss.match?(/</) then case when ss.skip(/\<\=\>/) then action { result :arg_state, :tCMP, "<=>" } when ss.skip(/\<\=/) then action { result :arg_state, :tLEQ, "<=" } when ss.skip(/\<\<\=/) then action { result EXPR_BEG, :tOP_ASGN, "<<" } when text = ss.scan(/\<\</) then process_lchevron text when ss.skip(/\</) then action { result :arg_state, :tLT, "<" } end # group /</ when ss.match?(/>/) then case when ss.skip(/\>\=/) then action { result :arg_state, :tGEQ, ">=" } when ss.skip(/\>\>=/) then action { result EXPR_BEG, :tOP_ASGN, ">>" } when ss.skip(/\>\>/) then action { result :arg_state, :tRSHFT, ">>" } when ss.skip(/\>/) then action { result :arg_state, :tGT, ">" } end # group />/ when ss.match?(/\`/) then case when expr_fname? && (ss.skip(/\`/)) then action { result EXPR_END, :tBACK_REF2, "`" } when expr_dot? && (ss.skip(/\`/)) then action { result((cmd_state ? EXPR_CMDARG : EXPR_ARG), :tBACK_REF2, "`") } when ss.skip(/\`/) then action { string STR_XQUOTE, '`'; result nil, :tXSTRING_BEG, "`" } end # group /\`/ when text = ss.scan(/\?/) then process_questionmark text when ss.match?(/&/) then case when ss.skip(/\&\&\=/) then action { result EXPR_BEG, :tOP_ASGN, "&&" } when ss.skip(/\&\&/) then action { result EXPR_BEG, :tANDOP, "&&" } when ss.skip(/\&\=/) then action { result EXPR_BEG, :tOP_ASGN, "&" } when ss.skip(/\&\./) then action { result EXPR_DOT, :tLONELY, "&." } when text = ss.scan(/\&/) then process_amper text end # group /&/ when text = ss.scan(/\//) then process_slash text when ss.match?(/\^/) then case when ss.skip(/\^=/) then action { result EXPR_BEG, :tOP_ASGN, "^" } when ss.skip(/\^/) then action { result :arg_state, :tCARET, "^" } end # group /\^/ when ss.skip(/\;/) then action { self.command_start = true; result EXPR_BEG, :tSEMI, ";" } when ss.match?(/~/) then case when is_after_operator? && (ss.skip(/\~@/)) then action { result :arg_state, :tTILDE, "~" } when ss.skip(/\~/) then action { result :arg_state, :tTILDE, "~" } end # group /~/ when ss.match?(/\\/) then case when ss.skip(/\\\r?\n/) then action { self.lineno += 1; self.space_seen = true; next } when ss.skip(/\\/) then action { rb_compile_error "bare backslash only allowed before newline" } end # group /\\/ when text = ss.scan(/\%/) then process_percent text when ss.match?(/\$/) then case when text = ss.scan(/\$_\w+/) then process_gvar text when text = ss.scan(/\$_/) then process_gvar text when text = ss.scan(/\$[~*$?!@\/\\;,.=:<>\"]|\$-\w?/) then process_gvar text when in_fname? && (text = ss.scan(/\$([\&\`\'\+])/)) then process_gvar text when text = ss.scan(/\$([\&\`\'\+])/) then process_backref text when in_fname? && (text = ss.scan(/\$([1-9]\d*)/)) then process_gvar text when text = ss.scan(/\$([1-9]\d*)/) then process_nthref text when text = ss.scan(/\$0/) then process_gvar text when text = ss.scan(/\$#{IDENT_CHAR}+/) then process_gvar text when text = ss.scan(/\$\W/) then process_gvar_oddity text end # group /\$/ when text = ss.scan(/\_/) then process_underscore text when text = ss.scan(/#{IDENT_CHAR}+/o) then process_token text when ss.skip(/\004|\032|\000|\Z/) then action { [RubyLexer::EOF, RubyLexer::EOF] } when text = ss.scan(/./) then action { rb_compile_error "Invalid char #{text.inspect} in expression" } else text = ss.string[ss.pos .. -1] raise ScanError, "can not match (#{state.inspect}) at #{location}: '#{text}'" end else raise ScanError, "undefined state at #{location}: '#{state}'" end # token = case state next unless token # allow functions to trigger redo w/ nil end # while raise LexerError, "bad lexical result at #{location}: #{token.inspect}" unless token.nil? || (Array === token && token.size >= 2) # auto-switch state self.state = token.last if token && token.first == :state token end
# File lib/ruby_lexer_strings.rb, line 174 def nextc # TODO: # if (UNLIKELY((p->lex.pcur == p->lex.pend) || p->eofp || RTEST(p->lex.nextline))) { # if (nextline(p)) return -1; # } maybe_pop_stack c = ss.getch if c == "\n" then ss.unscan c = nil end c end
# File lib/ruby_lexer.rb, line 218 def not_end? not is_end? end
Parse the given string.
# File lib/ruby_lexer.rex.rb, line 104 def parse str self.ss = scanner_class.new str self.lineno = 1 self.start_of_current_line_pos = 0 self.state ||= nil do_parse end
Read in and parse the file at path
.
# File lib/ruby_lexer.rex.rb, line 116 def parse_file path self.filename = path open path do |f| parse f.read end end
# File lib/ruby_lexer_strings.rb, line 192 def parse_string quote # ../compare/parse30.y:7273 _, func, term, paren = quote qwords = func =~ STR_FUNC_QWORDS regexp = func =~ STR_FUNC_REGEXP expand = func =~ STR_FUNC_EXPAND list = func =~ STR_FUNC_LIST termx = func =~ STR_FUNC_TERM # TODO: document wtf this means space = false term_re = regexp_cache[term] if termx then # self.nextc if qwords # delayed term self.lex_strterm = nil return result EXPR_END, regexp ? :tREGEXP_END : :tSTRING_END, term end space = true if qwords and eat_whitespace if list then debug 4 # quote[1] -= STR_FUNC_LIST # space = true end # TODO: move to quote.nest! if string_nest == 0 && scan(term_re) then if qwords then quote[1] |= STR_FUNC_TERM return :tSPACE, matched end return string_term func end return result nil, :tSPACE, " " if space newtok if expand && check(/#/) then t = self.scan_variable_name return t if t tokadd "#" end # TODO: add string_nest, enc, base_enc ? lineno = self.lineno if tokadd_string(func, term, paren) == RubyLexer::EOF then if qwords then rb_compile_error "unterminated list meets end of file" end if regexp then rb_compile_error "unterminated regexp meets end of file" else rb_compile_error "unterminated string meets end of file" end end result nil, :tSTRING_CONTENT, string_buffer.join, lineno end
# File lib/ruby_lexer.rb, line 222 def possibly_escape_string text, check content = match[1] if text =~ check then unescape_string content else content.gsub(/\\\\/, "\\").gsub(/\\\'/, "'") end end
# File lib/ruby_lexer.rb, line 232 def process_amper text token = if is_arg? && space_seen && !check(/\s/) then warning("`&' interpreted as argument prefix") :tAMPER elsif lex_state =~ EXPR_BEG|EXPR_MID then :tAMPER else :tAMPER2 end result :arg_state, token, "&" end
# File lib/ruby_lexer.rb, line 245 def process_backref text token = match[1].to_sym # TODO: can't do lineno hack w/ symbol result EXPR_END, :tBACK_REF, token end
# File lib/ruby_lexer.rb, line 251 def process_begin text self.comment ||= +"" self.comment << matched unless scan(/.*?\n=end( |\t|\f)*[^\n]*(\n|\z)/m) then self.comment = nil rb_compile_error("embedded document meets end of file") end self.comment << matched self.lineno += matched.count("\n") # HACK? nil # TODO end
TODO: make all tXXXX terminals include lexer.lineno … enforce it somehow?
# File lib/ruby_lexer.rb, line 268 def process_brace_close text case matched when "}" then self.brace_nest -= 1 return :tSTRING_DEND, matched if brace_nest < 0 end # matching compare/parse26.y:8099 cond.pop cmdarg.pop case matched when "}" then self.lex_state = ruby24minus? ? EXPR_ENDARG : EXPR_END return :tRCURLY, matched when "]" then self.paren_nest -= 1 self.lex_state = ruby24minus? ? EXPR_ENDARG : EXPR_END return :tRBRACK, matched when ")" then self.paren_nest -= 1 self.lex_state = EXPR_ENDFN return :tRPAREN, matched else raise "Unknown bracing: #{matched.inspect}" end end
# File lib/ruby_lexer.rb, line 296 def process_brace_open text # matching compare/parse23.y:8694 self.brace_nest += 1 if lambda_beginning? then self.lpar_beg = nil self.paren_nest -= 1 # close arg list when lambda opens body return expr_result(:tLAMBEG, "{") end token = case when lex_state =~ EXPR_LABELED then :tLBRACE # hash when lex_state =~ EXPR_ARG_ANY|EXPR_END|EXPR_ENDFN then :tLCURLY # block (primary) "{" in parse.y when lex_state =~ EXPR_ENDARG then :tLBRACE_ARG # block (expr) else :tLBRACE # hash end state = token == :tLBRACE_ARG ? EXPR_BEG : EXPR_PAR self.command_start = true if token != :tLBRACE cond.push false cmdarg.push false result state, token, text end
# File lib/ruby_lexer.rb, line 326 def process_colon1 text # ?: / then / when if is_end? || check(/\s/) then return result EXPR_BEG, :tCOLON, text end case when scan(/\'/) then string STR_SSYM, matched when scan(/\"/) then string STR_DSYM, matched end result EXPR_FNAME, :tSYMBEG, text end
# File lib/ruby_lexer.rb, line 342 def process_colon2 text if is_beg? || lex_state =~ EXPR_CLASS || is_space_arg? then result EXPR_BEG, :tCOLON3, text else result EXPR_DOT, :tCOLON2, text end end
# File lib/ruby_lexer.rb, line 350 def process_dots text # parse32.y:10216 is_beg = self.is_beg? self.lex_state = EXPR_BEG return result EXPR_ENDARG, :tBDOT3, text if parser.in_argdef && text == "..." # TODO: version check? tokens = ruby27plus? && is_beg ? BTOKENS : TOKENS result EXPR_BEG, tokens[text], text end
# File lib/ruby_lexer.rb, line 362 def process_float text rb_compile_error "Invalid numeric format" if text =~ /__/ case when text.end_with?("ri") result EXPR_NUM, :tIMAGINARY, Complex(0, Rational(text.chop.chop)) when text.end_with?("i") result EXPR_NUM, :tIMAGINARY, Complex(0, text.chop.to_f) when text.end_with?("r") result EXPR_NUM, :tRATIONAL, Rational(text.chop) else result EXPR_NUM, :tFLOAT, text.to_f end end
# File lib/ruby_lexer.rb, line 377 def process_gvar text if parser.class.version > 20 && text == "$-" then rb_compile_error "unexpected $undefined" end result EXPR_END, :tGVAR, text end
# File lib/ruby_lexer.rb, line 385 def process_gvar_oddity text rb_compile_error "#{text.inspect} is not allowed as a global variable name" end
# File lib/ruby_lexer.rb, line 389 def process_ivar text tok_id = text =~ /^@@/ ? :tCVAR : :tIVAR result EXPR_END, tok_id, text end
# File lib/ruby_lexer.rb, line 394 def process_label text symbol = possibly_escape_string text, /^\"/ result EXPR_LAB, :tLABEL, symbol end
# File lib/ruby_lexer.rb, line 400 def process_label_or_string text if @was_label && text =~ /:\Z/ then @was_label = nil return process_label text elsif text =~ /:\Z/ then self.pos -= 1 # put back ":" text = text[0..-2] end orig_line = lineno str = text[1..-2].gsub(/\\\\/, "\\").gsub(/\\\'/, "\'") self.lineno += str.count("\n") result EXPR_END, :tSTRING, str, orig_line end
# File lib/ruby_lexer.rb, line 416 def process_lchevron text if (lex_state !~ EXPR_DOT|EXPR_CLASS && !is_end? && (!is_arg? || lex_state =~ EXPR_LABELED || space_seen)) then tok = self.heredoc_identifier return tok if tok end if is_after_operator? then self.lex_state = EXPR_ARG else self.command_start = true if lex_state =~ EXPR_CLASS self.lex_state = EXPR_BEG end result lex_state, :tLSHFT, "\<\<" end
# File lib/ruby_lexer.rb, line 434 def process_newline_or_comment text # ../compare/parse30.y:9126 ish c = matched if c == "#" then self.pos -= 1 while scan(/\s*\#.*(\n+|\z)/) do self.lineno += matched.count "\n" self.comment ||= +"" self.comment << matched.gsub(/^ +#/, "#").gsub(/^ +$/, "") end return nil if end_of_stream? end c = (lex_state =~ EXPR_BEG|EXPR_CLASS|EXPR_FNAME|EXPR_DOT && lex_state !~ EXPR_LABELED) if c || self.lex_state == EXPR_LAB then # yes, == EXPR_LAB # ignore if !fallthrough? if !c && parser.in_kwarg then # normal newline self.command_start = true return result EXPR_BEG, :tNL, nil else maybe_pop_stack return # goto retry end end if scan(/[\ \t\r\f\v]+/) then self.space_seen = true end if check(/#/) then return # goto retry elsif check(/&\.|\.(?!\.)/) then # C version is a hellish obfuscated xnor return # goto retry end self.command_start = true result EXPR_BEG, :tNL, nil end
# File lib/ruby_lexer.rb, line 478 def process_nthref text # TODO: can't do lineno hack w/ number result EXPR_END, :tNTH_REF, match[1].to_i end
# File lib/ruby_lexer.rb, line 483 def process_paren text token = if is_beg? then :tLPAREN elsif !space_seen then # foo( ... ) => method call, no ambiguity :tLPAREN2 elsif is_space_arg? then :tLPAREN_ARG elsif lex_state =~ EXPR_ENDFN && !lambda_beginning? then # TODO: # warn("parentheses after method name is interpreted as " \ # "an argument list, not a decomposed argument") :tLPAREN2 else :tLPAREN2 # plain "(" in parse.y end self.paren_nest += 1 cond.push false cmdarg.push false result EXPR_PAR, token, text end
# File lib/ruby_lexer.rb, line 507 def process_percent text case when is_beg? then process_percent_quote when scan(/\=/) result EXPR_BEG, :tOP_ASGN, "%" when is_space_arg?(check(/\s/)) || (lex_state =~ EXPR_FITEM && check(/s/)) process_percent_quote else result :arg_state, :tPERCENT, "%" end end
called from process_percent
# File lib/ruby_lexer_strings.rb, line 260 def process_percent_quote # ../compare/parse30.y:8645 c = getch # type %<type><term>...<term> long_hand = !!(c =~ /[QqWwIixrs]/) if end_of_stream? || c !~ /\p{Alnum}/ then term = c # TODO? PERCENT_END[c] || c debug 2 if c && c !~ /\p{ASCII}/ c = "Q" else term = getch debug 3 if term =~ /\p{Alnum}|\P{ASCII}/ end if end_of_stream? or c == RubyLexer::EOF or term == RubyLexer::EOF then rb_compile_error "unterminated quoted string meets end of file" end # "\0" is special to indicate beg=nnd and that no nesting? paren = term term = PERCENT_END[term] term, paren = paren, "\0" if term.nil? # TODO: "\0" -> nil text = long_hand ? "%#{c}#{paren}" : "%#{term}" current_line = self.lineno token_type, string_type = case c when "Q" then [:tSTRING_BEG, STR_DQUOTE] when "q" then [:tSTRING_BEG, STR_SQUOTE] when "W" then eat_whitespace [:tWORDS_BEG, STR_DQUOTE | STR_FUNC_QWORDS] when "w" then eat_whitespace [:tQWORDS_BEG, STR_SQUOTE | STR_FUNC_QWORDS] when "I" then eat_whitespace [:tSYMBOLS_BEG, STR_DQUOTE | STR_FUNC_QWORDS] when "i" then eat_whitespace [:tQSYMBOLS_BEG, STR_SQUOTE | STR_FUNC_QWORDS] when "x" then [:tXSTRING_BEG, STR_XQUOTE] when "r" then [:tREGEXP_BEG, STR_REGEXP] when "s" then self.lex_state = EXPR_FNAME [:tSYMBEG, STR_SSYM] else rb_compile_error "unknown type of %string. Expected [QqWwIixrs], found '#{c}'." end string string_type, term, paren result nil, token_type, text, current_line end
# File lib/ruby_lexer.rb, line 520 def process_plus_minus text sign = matched utype, type = if sign == "+" then [:tUPLUS, :tPLUS] else [:tUMINUS, :tMINUS] end if is_after_operator? then if scan(/@/) then return result(EXPR_ARG, utype, "#{sign}@") else return result(EXPR_ARG, type, sign) end end return result(EXPR_BEG, :tOP_ASGN, sign) if scan(/\=/) if is_beg? || (is_arg? && space_seen && !check(/\s/)) then arg_ambiguous if is_arg? if check(/\d/) then return nil if utype == :tUPLUS return result EXPR_BEG, :tUMINUS_NUM, sign end return result EXPR_BEG, utype, sign end result EXPR_BEG, type, sign end
# File lib/ruby_lexer.rb, line 552 def process_questionmark text if is_end? then return result EXPR_BEG, :tEH, "?" end if end_of_stream? then rb_compile_error "incomplete character syntax: parsed #{text.inspect}" end if check(/\s|\v/) then unless is_arg? then c2 = { " " => "s", "\n" => "n", "\t" => "t", "\v" => "v", "\r" => "r", "\f" => "f" }[matched] if c2 then warning("invalid character syntax; use ?\\" + c2) end end # ternary return result EXPR_BEG, :tEH, "?" elsif check(/\w(?=\w)/) then # ternary, also return result EXPR_BEG, :tEH, "?" end c = if scan(/\\/) then self.read_escape else getch end result EXPR_END, :tSTRING, c end
# File lib/ruby_lexer.rb, line 590 def process_simple_string text orig_line = lineno self.lineno += text.count("\n") str = unescape_string text[1..-2] result EXPR_END, :tSTRING, str, orig_line end
# File lib/ruby_lexer.rb, line 599 def process_slash text if is_beg? then string STR_REGEXP, matched return result nil, :tREGEXP_BEG, "/" end if scan(/\=/) then return result(EXPR_BEG, :tOP_ASGN, "/") end if is_arg? && space_seen then unless scan(/\s/) then arg_ambiguous string STR_REGEXP, "/" return result(nil, :tREGEXP_BEG, "/") end end result :arg_state, :tDIVIDE, "/" end
# File lib/ruby_lexer.rb, line 621 def process_square_bracket text self.paren_nest += 1 token = nil if is_after_operator? then case when scan(/\]\=/) then self.paren_nest -= 1 # HACK? I dunno, or bug in MRI return result EXPR_ARG, :tASET, "[]=" when scan(/\]/) then self.paren_nest -= 1 # HACK? I dunno, or bug in MRI return result EXPR_ARG, :tAREF, "[]" else rb_compile_error "unexpected '['" end elsif is_beg? then token = :tLBRACK elsif is_arg? && (space_seen || lex_state =~ EXPR_LABELED) then token = :tLBRACK else token = :tLBRACK2 end cond.push false cmdarg.push false result EXPR_PAR, token, text end
# File lib/ruby_lexer_strings.rb, line 323 def process_string_or_heredoc # ../compare/parse30.y:9075 if lex_strterm[0] == :heredoc then self.heredoc lex_strterm else self.parse_string lex_strterm end end
# File lib/ruby_lexer.rb, line 650 def process_symbol text symbol = possibly_escape_string text, /^:\"/ # stupid emacs result EXPR_LIT, :tSYMBOL, symbol end
# File lib/ruby_lexer.rb, line 656 def process_token text # matching: parse_ident in compare/parse23.y:7989 # FIX: remove: self.last_state = lex_state token = self.token = text token << matched if scan(/[\!\?](?!=)/) tok_id = case when token =~ /[!?]$/ then :tFID when lex_state =~ EXPR_FNAME && scan(/=(?:(?![~>=])|(?==>))/) then # ident=, not =~ => == or followed by => # TODO test lexing of a=>b vs a==>b token << matched :tIDENTIFIER when token =~ /^[A-Z]/ then :tCONSTANT else :tIDENTIFIER end if is_label_possible? and is_label_suffix? then scan(/:/) return result EXPR_LAB, :tLABEL, token end # TODO: mb == ENC_CODERANGE_7BIT && lex_state !~ EXPR_DOT if lex_state !~ EXPR_DOT then # See if it is a reserved word. keyword = RubyParserStuff::Keyword.keyword token return process_token_keyword keyword if keyword end # matching: compare/parse32.y:9031 state = if lex_state =~ EXPR_BEG_ANY|EXPR_ARG_ANY|EXPR_DOT then cmd_state ? EXPR_CMDARG : EXPR_ARG elsif lex_state =~ EXPR_FNAME then EXPR_ENDFN else EXPR_END end self.lex_state = state tok_id = :tIDENTIFIER if tok_id == :tCONSTANT && is_local_id(token) if last_state !~ EXPR_DOT|EXPR_FNAME and (tok_id == :tIDENTIFIER) and # not EXPR_FNAME, not attrasgn lvar_defined?(token) then state = EXPR_END|EXPR_LABEL end result state, tok_id, token end
# File lib/ruby_lexer.rb, line 712 def process_token_keyword keyword # matching MIDDLE of parse_ident in compare/parse32.y:9695 state = lex_state return result(EXPR_ENDFN, keyword.id0, token) if lex_state =~ EXPR_FNAME self.lex_state = keyword.state self.command_start = true if lex_state =~ EXPR_BEG case when keyword.id0 == :kDO then # parse32.y line 9712 case when lambda_beginning? then self.lpar_beg = nil # lambda_beginning? == FALSE in the body of "-> do ... end" self.paren_nest -= 1 # TODO: question this? result lex_state, :kDO_LAMBDA, token when cond.is_in_state then result lex_state, :kDO_COND, token when cmdarg.is_in_state && state != EXPR_CMDARG then result lex_state, :kDO_BLOCK, token else result lex_state, :kDO, token end when state =~ EXPR_PAD then result lex_state, keyword.id0, token when keyword.id0 != keyword.id1 then result EXPR_PAR, keyword.id1, token else result lex_state, keyword.id1, token end end
# File lib/ruby_lexer.rb, line 744 def process_underscore text self.unscan # put back "_" if beginning_of_line? && scan(/\__END__(\r?\n|\Z)/) then ss.terminate [RubyLexer::EOF, RubyLexer::EOF] elsif scan(/#{IDENT_CHAR}+/) then process_token matched end end
# File lib/ruby_lexer.rb, line 755 def rb_compile_error msg msg += ". near line #{self.lineno}: #{self.rest[/^.*/].inspect}" raise RubyParser::SyntaxError, msg end
# File lib/ruby_lexer_strings.rb, line 331 def read_escape flags = nil # ../compare/parse30.y:6712 case when scan(/\\/) then # Backslash '\\' when scan(/n/) then # newline "\n" when scan(/t/) then # horizontal tab "\t" when scan(/r/) then # carriage-return "\r" when scan(/f/) then # form-feed "\f" when scan(/v/) then # vertical tab "\13" when scan(/a/) then # alarm(bell) "\007" when scan(/e/) then # escape "\033" when scan(/[0-7]{1,3}/) then # octal constant (matched.to_i(8) & 0xFF).chr.force_encoding Encoding::UTF_8 when scan(/x([0-9a-fA-F]{1,2})/) then # hex constant # TODO: force encode everything to UTF-8? match[1].to_i(16).chr.force_encoding Encoding::UTF_8 when scan(/b/) then # backspace "\010" when scan(/s/) then # space " " when check(/M-\\u/) then debug 5 when scan(/M-\\(?=.)/) then c = read_escape c[0] = (c[0].ord | 0x80).chr c when scan(/M-(\p{ASCII})/) then # TODO: ISCNTRL(c) -> goto eof c = match[1] c[0] = (c[0].ord | 0x80).chr c when check(/(C-|c)\\u/) then debug 6 when scan(/(C-|c)\\?\?/) then 127.chr when scan(/(C-|c)\\/) then c = read_escape c[0] = (c[0].ord & 0x9f).chr c when scan(/(?:C-|c)(.)/) then c = match[1] c[0] = (c[0].ord & 0x9f).chr c when scan(/^[89]/i) then # bad octal or hex... MRI ignores them :( matched when scan(/u(\h{4})/) then [match[1].to_i(16)].pack("U") when scan(/u(\h{1,3})/) then debug 7 rb_compile_error "Invalid escape character syntax" when scan(/u\{(\h+(?: +\h+)*)\}/) then match[1].split.map { |s| s.to_i(16) }.pack("U*") when scan(/[McCx0-9]/) || end_of_stream? then rb_compile_error("Invalid escape character syntax") else getch end.dup end
# File lib/ruby_lexer_strings.rb, line 397 def regx_options # ../compare/parse30.y:6914 newtok options = scan(/\p{Alpha}+/) || "" rb_compile_error("unknown regexp options: %s" % [options]) if options =~ /[^ixmonesu]/ options end
# File lib/ruby_lexer.rb, line 760 def reset self.lineno = 1 self.brace_nest = 0 self.command_start = true self.comment = nil self.lex_state = EXPR_NONE self.lex_strterm = nil self.lpar_beg = nil self.paren_nest = 0 self.space_seen = false self.string_nest = 0 self.token = nil self.string_buffer = [] self.old_ss = nil self.old_lineno = nil self.cond.reset self.cmdarg.reset end
# File lib/ruby_lexer.rb, line 787 def ruby22_label? ruby22plus? and is_label_possible? end
# File lib/ruby_lexer.rb, line 791 def ruby22plus? parser.class.version >= 22 end
# File lib/ruby_lexer.rb, line 795 def ruby23plus? parser.class.version >= 23 end
# File lib/ruby_lexer.rb, line 799 def ruby24minus? parser.class.version <= 24 end
# File lib/ruby_lexer.rb, line 803 def ruby27plus? parser.class.version >= 27 end
# File lib/ruby_lexer_strings.rb, line 408 def scan_variable_name # ../compare/parse30.y:7208 case when scan(/#(?=\$(-.|[a-zA-Z_0-9~\*\$\?!@\/\\;,\.=:<>\"\&\`\'+]))/) then # TODO: !ISASCII return :tSTRING_DVAR, matched when scan(/#(?=\@\@?[a-zA-Z_])/) then # TODO: !ISASCII return :tSTRING_DVAR, matched when scan(/#[{]/) then self.command_start = true return :tSTRING_DBEG, [matched, lineno] when scan(/#/) then # do nothing but swallow end # if scan(/\P{ASCII}|_|\p{Alpha}/) then # TODO: fold into above DVAR cases # # if (!ISASCII(c) || c == '_' || ISALPHA(c)) # # return tSTRING_DVAR; # end nil end
The current scanner class. Must be overridden in subclasses.
# File lib/ruby_lexer.rex.rb, line 97 def scanner_class StringScanner end
# File lib/ruby_lexer.rb, line 807 def space_vs_beginning space_type, beg_type, fallback if is_space_arg? check(/./m) then warning "`**' interpreted as argument prefix" space_type elsif is_beg? then beg_type else # TODO: warn_balanced("**", "argument prefix"); fallback end end
# File lib/ruby_lexer_strings.rb, line 431 def string type, beg, nnd = nil # label = (IS_LABEL_POSSIBLE() ? str_label : 0); # p->lex.strterm = NEW_STRTERM(str_dquote | label, '"', 0); # p->lex.ptok = p->lex.pcur-1; type |= STR_FUNC_LABEL if is_label_possible? self.lex_strterm = [:strterm, type, beg, nnd || "\0"] end
# File lib/ruby_lexer_strings.rb, line 440 def string_term func # ../compare/parse30.y:7254 self.lex_strterm = nil return result EXPR_END, :tREGEXP_END, self.regx_options if func =~ STR_FUNC_REGEXP if func =~ STR_FUNC_LABEL && is_label_suffix? then self.getch self.lex_state = EXPR_BEG|EXPR_LABEL return :tLABEL_END, string_buffer.join end self.lex_state = EXPR_END return :tSTRING_END, [self.matched, func] end
# File lib/ruby_lexer_strings.rb, line 458 def tokadd c # ../compare/parse30.y:6548 string_buffer << c end
# File lib/ruby_lexer_strings.rb, line 462 def tokadd_escape # ../compare/parse30.y:6840 case when scan(/\\\n/) then # just ignore when scan(/\\([0-7]{1,3}|x[0-9a-fA-F]{1,2})/) then tokadd matched when scan(/\\([MC]-|c)(?=\\)/) then tokadd matched self.tokadd_escape when scan(/\\([MC]-|c)(.)/) then tokadd matched self.tokadd_escape if check(/\\/) # recurse if continued! when scan(/\\[McCx]/) then # all unprocessed branches from above have failed rb_compile_error "Invalid escape character syntax" when scan(/\\(.)/m) then chr, = match[1] tokadd "\\" tokadd chr else rb_compile_error "Invalid escape character syntax: %p" % [self.rest.lines.first] end end
# File lib/ruby_lexer_strings.rb, line 487 def tokadd_string func, term, paren # ../compare/parse30.y:7020 qwords = func =~ STR_FUNC_QWORDS escape = func =~ STR_FUNC_ESCAPE expand = func =~ STR_FUNC_EXPAND regexp = func =~ STR_FUNC_REGEXP paren_re = regexp_cache[paren] if paren != "\0" term_re = if term == "\n" /\r?\n/ else regexp_cache[term] end until end_of_stream? do case when paren_re && scan(paren_re) then self.string_nest += 1 when scan(term_re) then if self.string_nest == 0 then self.pos -= 1 # TODO: ss.unscan 665 errors #$ HACK: why do we depend on this so hard? break # leave eos loop, go parse term in caller (heredoc or parse_string) else self.lineno += matched.count("\n") self.string_nest -= 1 end when expand && check(/#[\$\@\{]/) then # do nothing since we used `check` break # leave eos loop when check(/\\/) then case when scan(/\\\n/) then self.lineno += 1 case when qwords then tokadd "\n" next when expand then next if func !~ STR_FUNC_INDENT if term == "\n" then unscan # rollback scan(/\\/) # and split scan(/\n/) # this is `matched` break end tokadd "\\" debug 9 else unscan # rollback scan(/\\/) # this is `matched` end when check(/\\\\/) then tokadd '\\' if escape nextc # ignore 1st \\ nextc # for tokadd ss.matched, below when scan(/\\u/) then unless expand then tokadd "\\" next end tokadd_utf8 term, func, regexp next else scan(/\\/) # eat it, we know it's there return RubyLexer::EOF if end_of_stream? if scan(/\P{ASCII}/) then tokadd "\\" unless expand tokadd self.matched next end case when regexp then if term !~ SIMPLE_RE_META && scan(term_re) then tokadd matched next end self.pos -= 1 # TODO: ss.unscan 15 errors # HACK? decide whether to eat the \\ above if _esc = tokadd_escape && end_of_stream? then debug 10 end next # C's continue = Ruby's next when expand then tokadd "\\" if escape tokadd read_escape next when qwords && scan(/\s/) then # ignore backslashed spaces in %w when !check(term_re) && !(paren_re && check(paren_re)) then tokadd "\\" next else getch # slurp it too for matched below end end # inner case for /\\/ when scan(/\P{ASCII}/) then # not currently checking encoding stuff -- drops to tokadd below when qwords && check(/\s/) then break # leave eos loop else t = Regexp.escape term == "\n" ? "\r\n" : term x = Regexp.escape paren if paren && paren != "\000" q = "\\s" if qwords re = /[^#{t}#{x}\#\\#{q}]+/ scan re or getch self.lineno += matched.count "\n" if matched end # big case tokadd self.matched end # until end_of_stream? if self.matched then self.matched elsif end_of_stream? then RubyLexer::EOF end end
# File lib/ruby_lexer_strings.rb, line 616 def tokadd_utf8 term, func, regexp_literal # ../compare/parse30.y:6646 tokadd "\\u" if regexp_literal case when scan(/\h{4}/) then codepoint = [matched.to_i(16)].pack("U") tokadd regexp_literal ? matched : codepoint when scan(/\{\s*(\h{1,6}(?:\s+\h{1,6})*)\s*\}/) then codepoints = match[1].split.map { |s| s.to_i 16 }.pack("U") if regexp_literal then tokadd "{" tokadd match[1].split.join(" ") tokadd "}" else tokadd codepoints end else rb_compile_error "unterminated Unicode escape" end end
# File lib/ruby_lexer.rb, line 828 def unescape s r = ESCAPES[s] return r if r x = case s when /^[0-7]{1,3}/ then ($&.to_i(8) & 0xFF).chr when /^x([0-9a-fA-F]{1,2})/ then $1.to_i(16).chr when /^M-(.)/ then ($1[0].ord | 0x80).chr when /^(C-|c)(.)/ then ($2[0].ord & 0x9f).chr when /^[89a-f]/i then # bad octal or hex... ignore? that's what MRI does :( s when /^[McCx0-9]/ then rb_compile_error("Invalid escape character syntax") when /u(\h{4})/ then [$1.delete("{}").to_i(16)].pack("U") when /u(\h{1,3})/ then rb_compile_error("Invalid escape character syntax") when /u\{(\h+(?:\s+\h+)*)\}/ then $1.split.map { |cp| cp.to_i(16) }.pack("U*") else s end x end
# File lib/ruby_lexer.rb, line 819 def unescape_string str str = str.gsub(ESC) { unescape($1).b.force_encoding Encoding::UTF_8 } if str.valid_encoding? str else str.b end end
# File lib/ruby_lexer.rb, line 858 def warning s # do nothing for now end
# File lib/ruby_lexer.rb, line 862 def was_label? @was_label = ruby22_label? true end