lexers/bash.lua

   1 -- Copyright 2006-2015 Mitchell mitchell.att.foicica.com. See LICENSE.
   2 -- Shell 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 = 'bash'}
   9
  10 -- Whitespace.
  11 local ws = token(l.WHITESPACE, l.space^1)
  12
  13 -- Comments.
  14 local comment = token(l.COMMENT, '#' * l.nonnewline^0)
  15
  16 -- Strings.
  17 local sq_str = l.delimited_range("'", false, true)
  18 local dq_str = l.delimited_range('"')
  19 local ex_str = l.delimited_range('`')
  20 local heredoc = '<<' * P(function(input, index)
  21   local s, e, _, delimiter =
  22     input:find('%-?(["\']?)([%a_][%w_]*)%1[\n\r\f;]+', index)
  23   if s == index and delimiter then
  24     local _, e = input:find('[\n\r\f]+'..delimiter, e)
  25     return e and e + 1 or #input + 1
  26   end
  27 end)
  28 local string = token(l.STRING, sq_str + dq_str + ex_str + heredoc)
  29
  30 -- Numbers.
  31 local number = token(l.NUMBER, l.float + l.integer)
  32
  33 -- Keywords.
  34 local keyword = token(l.KEYWORD, word_match({
  35   'if', 'then', 'elif', 'else', 'fi', 'case', 'in', 'esac', 'while', 'for',
  36   'do', 'done', 'continue', 'local', 'return', 'select',
  37   -- Operators.
  38   '-a', '-b', '-c', '-d', '-e', '-f', '-g', '-h', '-k', '-p', '-r', '-s', '-t',
  39   '-u', '-w', '-x', '-O', '-G', '-L', '-S', '-N', '-nt', '-ot', '-ef', '-o',
  40   '-z', '-n', '-eq', '-ne', '-lt', '-le', '-gt', '-ge'
  41 }, '-'))
  42
  43 -- Identifiers.
  44 local identifier = token(l.IDENTIFIER, l.word)
  45
  46 -- Variables.
  47 local variable = token(l.VARIABLE,
  48                        '$' * (S('!#?*@$') + l.digit^1 + l.word +
  49                               l.delimited_range('{}', true, true)))
  50
  51 -- Operators.
  52 local operator = token(l.OPERATOR, S('=!<>+-/*^&|~.,:;?()[]{}'))
  53
  54 M._rules = {
  55   {'whitespace', ws},
  56   {'keyword', keyword},
  57   {'identifier', identifier},
  58   {'string', string},
  59   {'comment', comment},
  60   {'number', number},
  61   {'variable', variable},
  62   {'operator', operator},
  63 }
  64
  65 M._foldsymbols = {
  66   _patterns = {'[a-z]+', '[{}]', '#'},
  67   [l.KEYWORD] = {
  68     ['if'] = 1, fi = -1, case = 1, esac = -1, ['do'] = 1, done = -1
  69   },
  70   [l.OPERATOR] = {['{'] = 1, ['}'] = -1},
  71   [l.COMMENT] = {['#'] = l.fold_line_comments('#')}
  72 }
  73
  74 return M