lexers/pascal.lua

   1 -- Copyright 2006-2015 Mitchell mitchell.att.foicica.com. See LICENSE.
   2 -- Pascal 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 = 'pascal'}
   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 bblock_comment = '{' * (l.any - '}')^0 * P('}')^-1
  16 local pblock_comment = '(*' * (l.any - '*)')^0 * P('*)')^-1
  17 local comment = token(l.COMMENT, line_comment + bblock_comment + pblock_comment)
  18
  19 -- Strings.
  20 local string = token(l.STRING, S('uUrR')^-1 *
  21                                l.delimited_range("'", true, true))
  22
  23 -- Numbers.
  24 local number = token(l.NUMBER, (l.float + l.integer) * S('LlDdFf')^-1)
  25
  26 -- Keywords.
  27 local keyword = token(l.KEYWORD, word_match({
  28   'and', 'array', 'as', 'at', 'asm', 'begin', 'case', 'class', 'const',
  29   'constructor', 'destructor', 'dispinterface', 'div', 'do', 'downto', 'else',
  30   'end', 'except', 'exports', 'file', 'final', 'finalization', 'finally', 'for',
  31   'function', 'goto', 'if', 'implementation', 'in', 'inherited',
  32   'initialization', 'inline', 'interface', 'is', 'label', 'mod', 'not',
  33   'object', 'of', 'on', 'or', 'out', 'packed', 'procedure', 'program',
  34   'property', 'raise', 'record', 'repeat', 'resourcestring', 'set', 'sealed',
  35   'shl', 'shr', 'static', 'string', 'then', 'threadvar', 'to', 'try', 'type',
  36   'unit', 'unsafe', 'until', 'uses', 'var', 'while', 'with', 'xor',
  37   'absolute', 'abstract', 'assembler', 'automated', 'cdecl', 'contains',
  38   'default', 'deprecated', 'dispid', 'dynamic', 'export', 'external', 'far',
  39   'forward', 'implements', 'index', 'library', 'local', 'message', 'name',
  40   'namespaces', 'near', 'nodefault', 'overload', 'override', 'package',
  41   'pascal', 'platform', 'private', 'protected', 'public', 'published', 'read',
  42   'readonly', 'register', 'reintroduce', 'requires', 'resident', 'safecall',
  43   'stdcall', 'stored', 'varargs', 'virtual', 'write', 'writeln', 'writeonly',
  44   'false', 'nil', 'self', 'true'
  45 }, nil, true))
  46
  47 -- Functions.
  48 local func = token(l.FUNCTION, word_match({
  49   'chr', 'ord', 'succ', 'pred', 'abs', 'round', 'trunc', 'sqr', 'sqrt',
  50   'arctan', 'cos', 'sin', 'exp', 'ln', 'odd', 'eof', 'eoln'
  51 }, nil, true))
  52
  53 -- Types.
  54 local type = token(l.TYPE, word_match({
  55   'shortint', 'byte', 'char', 'smallint', 'integer', 'word', 'longint',
  56   'cardinal', 'boolean', 'bytebool', 'wordbool', 'longbool', 'real', 'single',
  57   'double', 'extended', 'comp', 'currency', 'pointer'
  58 }, nil, true))
  59
  60 -- Identifiers.
  61 local identifier = token(l.IDENTIFIER, l.word)
  62
  63 -- Operators.
  64 local operator = token(l.OPERATOR, S('.,;^@:=<>+-/*()[]'))
  65
  66 M._rules = {
  67   {'whitespace', ws},
  68   {'keyword', keyword},
  69   {'function', func},
  70   {'type', type},
  71   {'string', string},
  72   {'identifier', identifier},
  73   {'comment', comment},
  74   {'number', number},
  75   {'operator', operator},
  76 }
  77
  78 return M