lexers/antlr.lua

   1 -- Copyright 2006-2015 Mitchell mitchell.att.foicica.com. See LICENSE.
   2 -- ANTLR LPeg lexer.
   3
   4 local l = require('lexer')
   5 local token, word_match = l.token, l.word_match
   6 local P, R, S = lpeg.P, lpeg.R, lpeg.S
   7
   8 local M = {_NAME = 'antlr'}
   9
  10 -- Whitespace.
  11 local ws = token(l.WHITESPACE, l.space^1)
  12
  13 -- Comments.
  14 local line_comment = '//' * l.nonnewline^0
  15 local block_comment = '/*' * (l.any - '*/')^0 * P('*/')^-1
  16 local comment = token(l.COMMENT, line_comment + block_comment)
  17
  18 -- Strings.
  19 local string = token(l.STRING, l.delimited_range("'", true))
  20
  21 -- Keywords.
  22 local keyword = token(l.KEYWORD, word_match{
  23   'abstract', 'break', 'case', 'catch', 'continue', 'default', 'do', 'else',
  24   'extends', 'final', 'finally', 'for', 'if', 'implements', 'instanceof',
  25   'native', 'new', 'private', 'protected', 'public', 'return', 'static',
  26   'switch', 'synchronized', 'throw', 'throws', 'transient', 'try', 'volatile',
  27   'while', 'package', 'import', 'header', 'options', 'tokens', 'strictfp',
  28   'false', 'null', 'super', 'this', 'true'
  29 })
  30
  31 -- Types.
  32 local type = token(l.TYPE, word_match{
  33   'boolean', 'byte', 'char', 'class', 'double', 'float', 'int', 'interface',
  34   'long', 'short', 'void'
  35 })
  36
  37 -- Functions.
  38 local func = token(l.FUNCTION, 'assert')
  39
  40 -- Identifiers.
  41 local identifier = token(l.IDENTIFIER, l.word)
  42
  43 -- Operators.
  44 local operator = token(l.OPERATOR, S('$@:;|.=+*?~!^>-()[]{}'))
  45
  46 -- Actions.
  47 local action = #P('{') * operator * token('action', (1 - P('}'))^0) *
  48                (#P('}') * operator)^-1
  49
  50 M._rules = {
  51   {'whitespace', ws},
  52   {'keyword', keyword},
  53   {'type', type},
  54   {'function', func},
  55   {'identifier', identifier},
  56   {'string', string},
  57   {'comment', comment},
  58   {'action', action},
  59   {'operator', operator},
  60 }
  61
  62 M._tokenstyles = {
  63   action = l.STYLE_NOTHING
  64 }
  65
  66 M._foldsymbols = {
  67   _patterns = {'[:;%(%){}]', '/%*', '%*/', '//'},
  68   [l.OPERATOR] = {
  69     [':'] = 1, [';'] = -1, ['('] = 1, [')'] = -1, ['{'] = 1, ['}'] = -1
  70   },
  71   [l.COMMENT] = {['/*'] = 1, ['*/'] = -1, ['//'] = l.fold_line_comments('//')}
  72 }
  73
  74 return M