lua/lexers/protobuf.lua

   1 -- Copyright 2016-2017 David B. Lamkins <david@lamkins.net>. See LICENSE.
   2 -- Protocol Buffer IDL LPeg lexer.
   3 -- <https://developers.google.com/protocol-buffers/>
   4
   5 local l = require('lexer')
   6 local token, word_match = l.token, l.word_match
   7 local P, R, S = lpeg.P, lpeg.R, lpeg.S
   8
   9 local M = {_NAME = 'protobuf'}
  10
  11 -- Whitespace.
  12 local ws = token(l.WHITESPACE, l.space^1)
  13
  14 -- Comments.
  15 local line_comment = '//' * l.nonnewline_esc^0
  16 local block_comment = '/*' * (l.any - '*/')^0 * P('*/')^-1
  17 local comment = token(l.COMMENT, line_comment + block_comment)
  18
  19 -- Strings.
  20 local sq_str = P('L')^-1 * l.delimited_range("'", true)
  21 local dq_str = P('L')^-1 * l.delimited_range('"', true)
  22 local string = token(l.STRING, sq_str + dq_str)
  23
  24 -- Numbers.
  25 local number = token(l.NUMBER, l.float + l.integer)
  26
  27 -- Keywords.
  28 local keyword = token(l.KEYWORD, word_match{
  29   'contained', 'syntax', 'import', 'option', 'package', 'message', 'group',
  30   'oneof', 'optional', 'required', 'repeated', 'default', 'extend',
  31   'extensions', 'to', 'max', 'reserved', 'service', 'rpc', 'returns'
  32 })
  33
  34 -- Types.
  35 local type = token(l.TYPE, word_match{
  36   'int32', 'int64', 'uint32', 'uint64', 'sint32', 'sint64', 'fixed32',
  37   'fixed64', 'sfixed32', 'sfixed64', 'float', 'double', 'bool', 'string',
  38   'bytes', 'enum', 'true', 'false'
  39 })
  40
  41 -- Identifiers.
  42 local identifier = token(l.IDENTIFIER, l.word)
  43
  44 -- Operators.
  45 local operator = token(l.OPERATOR, S('<>=|;,.()[]{}'))
  46
  47 M._rules = {
  48   {'whitespace', ws},
  49   {'keyword', keyword},
  50   {'type', type},
  51   {'identifier', identifier},
  52   {'string', string},
  53   {'comment', comment},
  54   {'number', number},
  55   {'operator', operator},
  56 }
  57
  58 return M