lexers/groovy.lua

   1 -- Copyright 2006-2015 Mitchell mitchell.att.foicica.com. See LICENSE.
   2 -- Groovy 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 = 'groovy'}
   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("'")
  20 local dq_str = l.delimited_range('"')
  21 local triple_sq_str = "'''" * (l.any - "'''")^0 * P("'''")^-1
  22 local triple_dq_str = '"""' * (l.any - '"""')^0 * P('"""')^-1
  23 local regex_str = #P('/') * l.last_char_includes('=~|!<>+-*?&,:;([{') *
  24                   l.delimited_range('/', true)
  25 local string = token(l.STRING, triple_sq_str + triple_dq_str + sq_str +
  26                                dq_str) +
  27                token(l.REGEX, regex_str)
  28
  29 -- Numbers.
  30 local number = token(l.NUMBER, l.float + l.integer)
  31
  32 -- Keywords.
  33 local keyword = token(l.KEYWORD, word_match{
  34   'abstract', 'break', 'case', 'catch', 'continue', 'default', 'do', 'else',
  35   'extends', 'final', 'finally', 'for', 'if', 'implements', 'instanceof',
  36   'native', 'new', 'private', 'protected', 'public', 'return', 'static',
  37   'switch', 'synchronized', 'throw', 'throws', 'transient', 'try', 'volatile',
  38   'while', 'strictfp', 'package', 'import', 'as', 'assert', 'def', 'mixin',
  39   'property', 'test', 'using', 'in',
  40   'false', 'null', 'super', 'this', 'true', 'it'
  41 })
  42
  43 -- Functions.
  44 local func = token(l.FUNCTION, word_match{
  45   'abs', 'any', 'append', 'asList', 'asWritable', 'call', 'collect',
  46   'compareTo', 'count', 'div', 'dump', 'each', 'eachByte', 'eachFile',
  47   'eachLine', 'every', 'find', 'findAll', 'flatten', 'getAt', 'getErr', 'getIn',
  48   'getOut', 'getText', 'grep', 'immutable', 'inject', 'inspect', 'intersect',
  49   'invokeMethods', 'isCase', 'join', 'leftShift', 'minus', 'multiply',
  50   'newInputStream', 'newOutputStream', 'newPrintWriter', 'newReader',
  51   'newWriter', 'next', 'plus', 'pop', 'power', 'previous', 'print', 'println',
  52   'push', 'putAt', 'read', 'readBytes', 'readLines', 'reverse', 'reverseEach',
  53   'round', 'size', 'sort', 'splitEachLine', 'step', 'subMap', 'times',
  54   'toInteger', 'toList', 'tokenize', 'upto', 'waitForOrKill', 'withPrintWriter',
  55   'withReader', 'withStream', 'withWriter', 'withWriterAppend', 'write',
  56   'writeLine'
  57 })
  58
  59 -- Types.
  60 local type = token(l.TYPE, word_match{
  61   'boolean', 'byte', 'char', 'class', 'double', 'float', 'int', 'interface',
  62   'long', 'short', 'void'
  63 })
  64
  65 -- Identifiers.
  66 local identifier = token(l.IDENTIFIER, l.word)
  67
  68 -- Operators.
  69 local operator = token(l.OPERATOR, S('=~|!<>+-/*?&.,:;()[]{}'))
  70
  71 M._rules = {
  72   {'whitespace', ws},
  73   {'keyword', keyword},
  74   {'function', func},
  75   {'type', type},
  76   {'identifier', identifier},
  77   {'comment', comment},
  78   {'string', string},
  79   {'number', number},
  80   {'operator', operator},
  81 }
  82
  83 M._foldsymbols = {
  84   _patterns = {'[{}]', '/%*', '%*/', '//'},
  85   [l.OPERATOR] = {['{'] = 1, ['}'] = -1},
  86   [l.COMMENT] = {['/*'] = 1, ['*/'] = -1, ['//'] = l.fold_line_comments('//')}
  87 }
  88
  89 return M