tests: adds positive test case with nothrow to enhance test coverage
[luajson.git] / lua / json / decode / strings.lua
blobcb3b5cc069680ecb437523d8cd1c931bf5ee7a4f
1 --[[
2 Licensed according to the included 'LICENSE' document
3 Author: Thomas Harning Jr <harningt@gmail.com>
4 ]]
5 local lpeg = require("lpeg")
6 local jsonutil = require("json.util")
7 local util = require("json.decode.util")
8 local merge = jsonutil.merge
10 local tonumber = tonumber
11 local string_char = require("string").char
12 local floor = require("math").floor
13 local table_concat = require("table").concat
15 local error = error
17 local _ENV = nil
19 local function get_error(item)
20 local fmt_string = item .. " in string [%q] @ %i:%i"
21 return lpeg.P(function(data, index)
22 local line, line_index, bad_char, last_line = util.get_invalid_character_info(data, index)
23 local err = fmt_string:format(bad_char, line, line_index)
24 error(err)
25 end) * 1
26 end
28 local bad_unicode = get_error("Illegal unicode escape")
29 local bad_hex = get_error("Illegal hex escape")
30 local bad_character = get_error("Illegal character")
31 local bad_escape = get_error("Illegal escape")
33 local knownReplacements = {
34 ["'"] = "'",
35 ['"'] = '"',
36 ['\\'] = '\\',
37 ['/'] = '/',
38 b = '\b',
39 f = '\f',
40 n = '\n',
41 r = '\r',
42 t = '\t',
43 v = '\v',
44 z = '\z'
47 -- according to the table at http://da.wikipedia.org/wiki/UTF-8
48 local function utf8DecodeUnicode(code1, code2)
49 code1, code2 = tonumber(code1, 16), tonumber(code2, 16)
50 if code1 == 0 and code2 < 0x80 then
51 return string_char(code2)
52 end
53 if code1 < 0x08 then
54 return string_char(
55 0xC0 + code1 * 4 + floor(code2 / 64),
56 0x80 + code2 % 64)
57 end
58 return string_char(
59 0xE0 + floor(code1 / 16),
60 0x80 + (code1 % 16) * 4 + floor(code2 / 64),
61 0x80 + code2 % 64)
62 end
64 local function decodeX(code)
65 code = tonumber(code, 16)
66 return string_char(code)
67 end
69 local doSimpleSub = lpeg.C(lpeg.S("'\"\\/bfnrtvz")) / knownReplacements
70 local doUniSub = lpeg.P('u') * (lpeg.C(util.hexpair) * lpeg.C(util.hexpair) + bad_unicode)
71 local doXSub = lpeg.P('x') * (lpeg.C(util.hexpair) + bad_hex)
73 local defaultOptions = {
74 badChars = '',
75 additionalEscapes = false, -- disallow untranslated escapes
76 escapeCheck = #lpeg.S('bfnrtv/\\"xu\'z'), -- no check on valid characters
77 decodeUnicode = utf8DecodeUnicode,
78 strict_quotes = false
81 local modeOptions = {}
83 modeOptions.strict = {
84 badChars = '\b\f\n\r\t\v',
85 additionalEscapes = false, -- no additional escapes
86 escapeCheck = #lpeg.S('bfnrtv/\\"u'), --only these chars are allowed to be escaped
87 strict_quotes = true
90 local function mergeOptions(options, mode)
91 jsonutil.doOptionMerge(options, false, 'strings', defaultOptions, mode and modeOptions[mode])
92 end
94 local function buildCaptureString(quote, badChars, escapeMatch)
95 local captureChar = (1 - lpeg.S("\\" .. badChars .. quote)) + (lpeg.P("\\") / "" * escapeMatch)
96 -- During error, force end
97 local captureString = captureChar^0 + (-#lpeg.P(quote) * bad_character + -1)
98 return lpeg.P(quote) * lpeg.Cs(captureString) * lpeg.P(quote)
99 end
101 local function generateLexer(options)
102 options = options.strings
103 local quotes = { '"' }
104 if not options.strict_quotes then
105 quotes[#quotes + 1] = "'"
107 local escapeMatch = doSimpleSub
108 escapeMatch = escapeMatch + doXSub / decodeX
109 escapeMatch = escapeMatch + doUniSub / options.decodeUnicode
110 if options.additionalEscapes then
111 escapeMatch = escapeMatch + options.additionalEscapes
113 if options.escapeCheck then
114 escapeMatch = options.escapeCheck * escapeMatch + bad_escape
116 local captureString
117 for i = 1, #quotes do
118 local cap = buildCaptureString(quotes[i], options.badChars, escapeMatch)
119 if captureString == nil then
120 captureString = cap
121 else
122 captureString = captureString + cap
125 return captureString
128 local strings = {
129 mergeOptions = mergeOptions,
130 generateLexer = generateLexer
133 return strings