lexers/idl.lua

   1 -- Copyright 2006-2016 Mitchell mitchell.att.foicica.com. See LICENSE.
   2 -- IDL 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 = 'idl'}
   9
  10 -- Whitespace.
  11 local ws = token(l.WHITESPACE, l.space^1)
  12
  13 -- Comments.
  14 local line_comment = '//' * l.nonnewline_esc^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 sq_str = l.delimited_range("'", true)
  20 local dq_str = l.delimited_range('"', true)
  21 local string = token(l.STRING, sq_str + dq_str)
  22
  23 -- Numbers.
  24 local number = token(l.NUMBER, l.float + l.integer)
  25
  26 -- Preprocessor.
  27 local preproc_word = word_match{
  28   'define', 'undef', 'ifdef', 'ifndef', 'if', 'elif', 'else', 'endif',
  29   'include', 'warning', 'pragma'
  30 }
  31 local preproc = token(l.PREPROCESSOR,
  32                       l.starts_line('#') * preproc_word * l.nonnewline^0)
  33
  34 -- Keywords.
  35 local keyword = token(l.KEYWORD, word_match{
  36   'abstract', 'attribute', 'case', 'const', 'context', 'custom', 'default',
  37   'exception', 'enum', 'factory', 'FALSE', 'in', 'inout', 'interface', 'local',
  38   'module', 'native', 'oneway', 'out', 'private', 'public', 'raises',
  39   'readonly', 'struct', 'support', 'switch', 'TRUE', 'truncatable', 'typedef',
  40   'union', 'valuetype'
  41 })
  42
  43 -- Types.
  44 local type = token(l.TYPE, word_match{
  45   'any', 'boolean', 'char', 'double', 'fixed', 'float', 'long', 'Object',
  46   'octet', 'sequence', 'short', 'string', 'unsigned', 'ValueBase', 'void',
  47   'wchar', 'wstring'
  48 })
  49
  50 -- Identifiers.
  51 local identifier = token(l.IDENTIFIER, l.word)
  52
  53 -- Operators.
  54 local operator = token(l.OPERATOR, S('!<>=+-/*%&|^~.,:;?()[]{}'))
  55
  56 M._rules = {
  57   {'whitespace', ws},
  58   {'keyword', keyword},
  59   {'type', type},
  60   {'identifier', identifier},
  61   {'string', string},
  62   {'comment', comment},
  63   {'number', number},
  64   {'preprocessor', preproc},
  65   {'operator', operator},
  66 }
  67
  68 return M