-- Copyright 2006-2024 Mitchell. See LICENSE. -- YAML LPeg lexer. -- It does not keep track of indentation perfectly. local lexer = lexer local word_match = lexer.word_match local P, S, B = lpeg.P, lpeg.S, lpeg.B local lex = lexer.new(..., {fold_by_indentation = true}) -- Distinguish between horizontal and vertical space so indenting tabs can be marked as errors. local tab_indent = lex:tag(lexer.ERROR .. '.indent', lexer.starts_line('\t', true)) lex:modify_rule('whitespace', tab_indent + lex:tag(lexer.WHITESPACE, S(' \r\n')^1 + P('\t')^1)) -- Document boundaries. lex:add_rule('doc_bounds', lex:tag(lexer.OPERATOR, lexer.starts_line(P('---') + '...'))) -- Keys. local word = (lexer.alnum + '-')^1 lex:add_rule('key', -P('- ') * lex:tag(lexer.STRING, word * (S(' \t_')^1 * word^-1)^0) * #P(':' * lexer.space)) -- Collections. lex:add_rule('collection', lex:tag(lexer.OPERATOR, lexer.after_set('?-:\n', S('?-') * #P(' '), ' \t') + ':' * #P(lexer.space) + S('[]{}') + ',' * #P(' '))) -- Alias indicators. local anchor = lex:tag(lexer.OPERATOR, '&') * lex:tag(lexer.LABEL, word) local alias = lex:tag(lexer.OPERATOR, '*') * lex:tag(lexer.LABEL, word) lex:add_rule('alias', anchor + alias) -- Tags. local explicit_tag = '!!' * word_match{ 'map', 'omap', 'pairs', 'set', 'seq', -- collection 'binary', 'bool', 'float', 'int', 'merge', 'null', 'str', 'timestamp', 'value', 'yaml' -- scalar } local verbatim_tag = '!' * lexer.range('<', '>', true) local short_tag = '!' * word * ('!' * (1 - lexer.space)^1)^-1 lex:add_rule('tag', lex:tag(lexer.TYPE, explicit_tag + verbatim_tag + short_tag)) -- Comments. lex:add_rule('comment', lex:tag(lexer.COMMENT, lexer.to_eol('#'))) -- Reserved. lex:add_rule('reserved', B(S(':,') * ' ') * lex:tag(lexer.ERROR, S('@`') + lexer.starts_line(S('@`')))) -- Constants. local scalar_end = #(S(' \t')^0 * lexer.newline + S(',]}') + -1) lex:add_rule('constant', lex:tag(lexer.CONSTANT_BUILTIN, word_match('null true false', true)) * scalar_end) -- Strings. local sq_str = lexer.range("'") local dq_str = lexer.range('"') lex:add_rule('string', lex:tag(lexer.STRING, sq_str + dq_str) * (scalar_end + #P(':' * lexer.space))) -- Timestamps. local year = lexer.digit * lexer.digit * lexer.digit * lexer.digit local month = lexer.digit * lexer.digit^-1 local day = lexer.digit * lexer.digit^-1 local date = year * '-' * month * '-' * day local hours = lexer.digit * lexer.digit^-1 local minutes = lexer.digit * lexer.digit local seconds = lexer.digit * lexer.digit local fraction = '.' * lexer.digit^0 local time = hours * ':' * minutes * ':' * seconds * fraction^-1 local zone = 'Z' + S(' \t')^-1 * S('-+') * hours * (':' * minutes)^-1 lex:add_rule('timestamp', lex:tag(lexer.NUMBER .. '.timestamp', date * (S('tT \t') * time * zone^-1)^-1) * scalar_end) -- Numbers. local special_num = S('+-')^-1 * '.' * word_match('inf nan', true) local number = lexer.number + special_num lex:add_rule('number', (B(lexer.alnum) * lex:tag(lexer.DEFAULT, number) + lex:tag(lexer.NUMBER, number)) * scalar_end) -- Scalars. local block_indicator = S('|>') * (S('-+') * lexer.digit^-1 + lexer.digit * S('-+')^-1)^-1 local block = lpeg.Cmt(lpeg.C(block_indicator * lexer.newline), function(input, index, indicator) local indent = lexer.indent_amount[lexer.line_from_position(index - #indicator)] for s, i, j in input:gmatch('()\n()[ \t]*()[^ \t\r\n]', index) do -- ignore blank lines if s >= index then -- compatibility for Lua < 5.4, which doesn't have init for string.gmatch() if j - i <= indent then return s end end end return #input + 1 end) local seq = B('- ') * lexer.nonnewline^1 local csv = B(', ') * (lexer.nonnewline - S(',]}'))^1 local stop_chars, LF = {[string.byte('{')] = true, [string.byte('\n')] = true}, string.byte('\n') local map = B(': ') * lexer.nonnewline * P(function(input, index) local pos = index while pos > 1 and not stop_chars[input:byte(pos)] do pos = pos - 1 end local s = input:find(input:byte(pos) ~= LF and '[\n,}]' or '\n', index) return s or #input + 1 end) lex:add_rule('scalar', lex:tag(lexer.DEFAULT, block + seq + csv + map)) -- Directives lex:add_rule('directive', lex:tag(lexer.PREPROCESSOR, lexer.starts_line(lexer.to_eol('%')))) lexer.property['scintillua.comment'] = '#' return lex