1 | // __ _____ _____ _____ |
2 | // __| | __| | | | JSON for Modern C++ |
3 | // | | |__ | | | | | | version 3.11.3 |
4 | // |_____|_____|_____|_|___| https://github.com/nlohmann/json |
5 | // |
6 | // SPDX-FileCopyrightText: 2013-2023 Niels Lohmann <https://nlohmann.me> |
7 | // SPDX-License-Identifier: MIT |
8 | |
9 | #pragma once |
10 | |
11 | #include <algorithm> // generate_n |
12 | #include <array> // array |
13 | #include <cmath> // ldexp |
14 | #include <cstddef> // size_t |
15 | #include <cstdint> // uint8_t, uint16_t, uint32_t, uint64_t |
16 | #include <cstdio> // snprintf |
17 | #include <cstring> // memcpy |
18 | #include <iterator> // back_inserter |
19 | #include <limits> // numeric_limits |
20 | #include <string> // char_traits, string |
21 | #include <utility> // make_pair, move |
22 | #include <vector> // vector |
23 | |
24 | #include <nlohmann/detail/exceptions.hpp> |
25 | #include <nlohmann/detail/input/input_adapters.hpp> |
26 | #include <nlohmann/detail/input/json_sax.hpp> |
27 | #include <nlohmann/detail/input/lexer.hpp> |
28 | #include <nlohmann/detail/macro_scope.hpp> |
29 | #include <nlohmann/detail/meta/is_sax.hpp> |
30 | #include <nlohmann/detail/meta/type_traits.hpp> |
31 | #include <nlohmann/detail/string_concat.hpp> |
32 | #include <nlohmann/detail/value_t.hpp> |
33 | |
34 | NLOHMANN_JSON_NAMESPACE_BEGIN |
35 | namespace detail |
36 | { |
37 | |
38 | /// how to treat CBOR tags |
39 | enum class cbor_tag_handler_t |
40 | { |
41 | error, ///< throw a parse_error exception in case of a tag |
42 | ignore, ///< ignore tags |
43 | store ///< store tags as binary type |
44 | }; |
45 | |
46 | /*! |
47 | @brief determine system byte order |
48 | |
49 | @return true if and only if system's byte order is little endian |
50 | |
51 | @note from https://stackoverflow.com/a/1001328/266378 |
52 | */ |
53 | static inline bool little_endianness(int num = 1) noexcept |
54 | { |
55 | return *reinterpret_cast<char*>(&num) == 1; |
56 | } |
57 | |
58 | /////////////////// |
59 | // binary reader // |
60 | /////////////////// |
61 | |
62 | /*! |
63 | @brief deserialization of CBOR, MessagePack, and UBJSON values |
64 | */ |
65 | template<typename BasicJsonType, typename InputAdapterType, typename SAX = json_sax_dom_parser<BasicJsonType>> |
66 | class binary_reader |
67 | { |
68 | using number_integer_t = typename BasicJsonType::number_integer_t; |
69 | using number_unsigned_t = typename BasicJsonType::number_unsigned_t; |
70 | using number_float_t = typename BasicJsonType::number_float_t; |
71 | using string_t = typename BasicJsonType::string_t; |
72 | using binary_t = typename BasicJsonType::binary_t; |
73 | using json_sax_t = SAX; |
74 | using char_type = typename InputAdapterType::char_type; |
75 | using char_int_type = typename char_traits<char_type>::int_type; |
76 | |
77 | public: |
78 | /*! |
79 | @brief create a binary reader |
80 | |
81 | @param[in] adapter input adapter to read from |
82 | */ |
83 | explicit binary_reader(InputAdapterType&& adapter, const input_format_t format = input_format_t::json) noexcept : ia(std::move(adapter)), input_format(format) |
84 | { |
85 | (void)detail::is_sax_static_asserts<SAX, BasicJsonType> {}; |
86 | } |
87 | |
88 | // make class move-only |
89 | binary_reader(const binary_reader&) = delete; |
90 | binary_reader(binary_reader&&) = default; // NOLINT(hicpp-noexcept-move,performance-noexcept-move-constructor) |
91 | binary_reader& operator=(const binary_reader&) = delete; |
92 | binary_reader& operator=(binary_reader&&) = default; // NOLINT(hicpp-noexcept-move,performance-noexcept-move-constructor) |
93 | ~binary_reader() = default; |
94 | |
95 | /*! |
96 | @param[in] format the binary format to parse |
97 | @param[in] sax_ a SAX event processor |
98 | @param[in] strict whether to expect the input to be consumed completed |
99 | @param[in] tag_handler how to treat CBOR tags |
100 | |
101 | @return whether parsing was successful |
102 | */ |
103 | JSON_HEDLEY_NON_NULL(3) |
104 | bool sax_parse(const input_format_t format, |
105 | json_sax_t* sax_, |
106 | const bool strict = true, |
107 | const cbor_tag_handler_t tag_handler = cbor_tag_handler_t::error) |
108 | { |
109 | sax = sax_; |
110 | bool result = false; |
111 | |
112 | switch (format) |
113 | { |
114 | case input_format_t::bson: |
115 | result = parse_bson_internal(); |
116 | break; |
117 | |
118 | case input_format_t::cbor: |
119 | result = parse_cbor_internal(get_char: true, tag_handler); |
120 | break; |
121 | |
122 | case input_format_t::msgpack: |
123 | result = parse_msgpack_internal(); |
124 | break; |
125 | |
126 | case input_format_t::ubjson: |
127 | case input_format_t::bjdata: |
128 | result = parse_ubjson_internal(); |
129 | break; |
130 | |
131 | case input_format_t::json: // LCOV_EXCL_LINE |
132 | default: // LCOV_EXCL_LINE |
133 | JSON_ASSERT(false); // NOLINT(cert-dcl03-c,hicpp-static-assert,misc-static-assert) LCOV_EXCL_LINE |
134 | } |
135 | |
136 | // strict mode: next byte must be EOF |
137 | if (result && strict) |
138 | { |
139 | if (input_format == input_format_t::ubjson || input_format == input_format_t::bjdata) |
140 | { |
141 | get_ignore_noop(); |
142 | } |
143 | else |
144 | { |
145 | get(); |
146 | } |
147 | |
148 | if (JSON_HEDLEY_UNLIKELY(current != char_traits<char_type>::eof())) |
149 | { |
150 | return sax->parse_error(chars_read, get_token_string(), parse_error::create(110, chars_read, |
151 | exception_message(format: input_format, detail: concat("expected end of input; last byte: 0x" , get_token_string()), context: "value" ), nullptr)); |
152 | } |
153 | } |
154 | |
155 | return result; |
156 | } |
157 | |
158 | private: |
159 | ////////// |
160 | // BSON // |
161 | ////////// |
162 | |
163 | /*! |
164 | @brief Reads in a BSON-object and passes it to the SAX-parser. |
165 | @return whether a valid BSON-value was passed to the SAX parser |
166 | */ |
167 | bool parse_bson_internal() |
168 | { |
169 | std::int32_t document_size{}; |
170 | get_number<std::int32_t, true>(input_format_t::bson, document_size); |
171 | |
172 | if (JSON_HEDLEY_UNLIKELY(!sax->start_object(static_cast<std::size_t>(-1)))) |
173 | { |
174 | return false; |
175 | } |
176 | |
177 | if (JSON_HEDLEY_UNLIKELY(!parse_bson_element_list(/*is_array*/false))) |
178 | { |
179 | return false; |
180 | } |
181 | |
182 | return sax->end_object(); |
183 | } |
184 | |
185 | /*! |
186 | @brief Parses a C-style string from the BSON input. |
187 | @param[in,out] result A reference to the string variable where the read |
188 | string is to be stored. |
189 | @return `true` if the \x00-byte indicating the end of the string was |
190 | encountered before the EOF; false` indicates an unexpected EOF. |
191 | */ |
192 | bool get_bson_cstr(string_t& result) |
193 | { |
194 | auto out = std::back_inserter(result); |
195 | while (true) |
196 | { |
197 | get(); |
198 | if (JSON_HEDLEY_UNLIKELY(!unexpect_eof(input_format_t::bson, "cstring" ))) |
199 | { |
200 | return false; |
201 | } |
202 | if (current == 0x00) |
203 | { |
204 | return true; |
205 | } |
206 | *out++ = static_cast<typename string_t::value_type>(current); |
207 | } |
208 | } |
209 | |
210 | /*! |
211 | @brief Parses a zero-terminated string of length @a len from the BSON |
212 | input. |
213 | @param[in] len The length (including the zero-byte at the end) of the |
214 | string to be read. |
215 | @param[in,out] result A reference to the string variable where the read |
216 | string is to be stored. |
217 | @tparam NumberType The type of the length @a len |
218 | @pre len >= 1 |
219 | @return `true` if the string was successfully parsed |
220 | */ |
221 | template<typename NumberType> |
222 | bool get_bson_string(const NumberType len, string_t& result) |
223 | { |
224 | if (JSON_HEDLEY_UNLIKELY(len < 1)) |
225 | { |
226 | auto last_token = get_token_string(); |
227 | return sax->parse_error(chars_read, last_token, parse_error::create(112, chars_read, |
228 | exception_message(format: input_format_t::bson, detail: concat("string length must be at least 1, is " , std::to_string(len)), context: "string" ), nullptr)); |
229 | } |
230 | |
231 | return get_string(input_format_t::bson, len - static_cast<NumberType>(1), result) && get() != char_traits<char_type>::eof(); |
232 | } |
233 | |
234 | /*! |
235 | @brief Parses a byte array input of length @a len from the BSON input. |
236 | @param[in] len The length of the byte array to be read. |
237 | @param[in,out] result A reference to the binary variable where the read |
238 | array is to be stored. |
239 | @tparam NumberType The type of the length @a len |
240 | @pre len >= 0 |
241 | @return `true` if the byte array was successfully parsed |
242 | */ |
243 | template<typename NumberType> |
244 | bool get_bson_binary(const NumberType len, binary_t& result) |
245 | { |
246 | if (JSON_HEDLEY_UNLIKELY(len < 0)) |
247 | { |
248 | auto last_token = get_token_string(); |
249 | return sax->parse_error(chars_read, last_token, parse_error::create(112, chars_read, |
250 | exception_message(format: input_format_t::bson, detail: concat("byte array length cannot be negative, is " , std::to_string(len)), context: "binary" ), nullptr)); |
251 | } |
252 | |
253 | // All BSON binary values have a subtype |
254 | std::uint8_t subtype{}; |
255 | get_number<std::uint8_t>(input_format_t::bson, subtype); |
256 | result.set_subtype(subtype); |
257 | |
258 | return get_binary(input_format_t::bson, len, result); |
259 | } |
260 | |
261 | /*! |
262 | @brief Read a BSON document element of the given @a element_type. |
263 | @param[in] element_type The BSON element type, c.f. http://bsonspec.org/spec.html |
264 | @param[in] element_type_parse_position The position in the input stream, |
265 | where the `element_type` was read. |
266 | @warning Not all BSON element types are supported yet. An unsupported |
267 | @a element_type will give rise to a parse_error.114: |
268 | Unsupported BSON record type 0x... |
269 | @return whether a valid BSON-object/array was passed to the SAX parser |
270 | */ |
271 | bool parse_bson_element_internal(const char_int_type element_type, |
272 | const std::size_t element_type_parse_position) |
273 | { |
274 | switch (element_type) |
275 | { |
276 | case 0x01: // double |
277 | { |
278 | double number{}; |
279 | return get_number<double, true>(input_format_t::bson, number) && sax->number_float(static_cast<number_float_t>(number), "" ); |
280 | } |
281 | |
282 | case 0x02: // string |
283 | { |
284 | std::int32_t len{}; |
285 | string_t value; |
286 | return get_number<std::int32_t, true>(input_format_t::bson, len) && get_bson_string(len, value) && sax->string(value); |
287 | } |
288 | |
289 | case 0x03: // object |
290 | { |
291 | return parse_bson_internal(); |
292 | } |
293 | |
294 | case 0x04: // array |
295 | { |
296 | return parse_bson_array(); |
297 | } |
298 | |
299 | case 0x05: // binary |
300 | { |
301 | std::int32_t len{}; |
302 | binary_t value; |
303 | return get_number<std::int32_t, true>(input_format_t::bson, len) && get_bson_binary(len, value) && sax->binary(value); |
304 | } |
305 | |
306 | case 0x08: // boolean |
307 | { |
308 | return sax->boolean(get() != 0); |
309 | } |
310 | |
311 | case 0x0A: // null |
312 | { |
313 | return sax->null(); |
314 | } |
315 | |
316 | case 0x10: // int32 |
317 | { |
318 | std::int32_t value{}; |
319 | return get_number<std::int32_t, true>(input_format_t::bson, value) && sax->number_integer(value); |
320 | } |
321 | |
322 | case 0x12: // int64 |
323 | { |
324 | std::int64_t value{}; |
325 | return get_number<std::int64_t, true>(input_format_t::bson, value) && sax->number_integer(value); |
326 | } |
327 | |
328 | default: // anything else not supported (yet) |
329 | { |
330 | std::array<char, 3> cr{._M_elems: {}}; |
331 | static_cast<void>((std::snprintf)(s: cr.data(), maxlen: cr.size(), format: "%.2hhX" , static_cast<unsigned char>(element_type))); // NOLINT(cppcoreguidelines-pro-type-vararg,hicpp-vararg) |
332 | const std::string cr_str{cr.data()}; |
333 | return sax->parse_error(element_type_parse_position, cr_str, |
334 | parse_error::create(id_: 114, byte_: element_type_parse_position, what_arg: concat(args: "Unsupported BSON record type 0x" , args: cr_str), context: nullptr)); |
335 | } |
336 | } |
337 | } |
338 | |
339 | /*! |
340 | @brief Read a BSON element list (as specified in the BSON-spec) |
341 | |
342 | The same binary layout is used for objects and arrays, hence it must be |
343 | indicated with the argument @a is_array which one is expected |
344 | (true --> array, false --> object). |
345 | |
346 | @param[in] is_array Determines if the element list being read is to be |
347 | treated as an object (@a is_array == false), or as an |
348 | array (@a is_array == true). |
349 | @return whether a valid BSON-object/array was passed to the SAX parser |
350 | */ |
351 | bool parse_bson_element_list(const bool is_array) |
352 | { |
353 | string_t key; |
354 | |
355 | while (auto element_type = get()) |
356 | { |
357 | if (JSON_HEDLEY_UNLIKELY(!unexpect_eof(input_format_t::bson, "element list" ))) |
358 | { |
359 | return false; |
360 | } |
361 | |
362 | const std::size_t element_type_parse_position = chars_read; |
363 | if (JSON_HEDLEY_UNLIKELY(!get_bson_cstr(key))) |
364 | { |
365 | return false; |
366 | } |
367 | |
368 | if (!is_array && !sax->key(key)) |
369 | { |
370 | return false; |
371 | } |
372 | |
373 | if (JSON_HEDLEY_UNLIKELY(!parse_bson_element_internal(element_type, element_type_parse_position))) |
374 | { |
375 | return false; |
376 | } |
377 | |
378 | // get_bson_cstr only appends |
379 | key.clear(); |
380 | } |
381 | |
382 | return true; |
383 | } |
384 | |
385 | /*! |
386 | @brief Reads an array from the BSON input and passes it to the SAX-parser. |
387 | @return whether a valid BSON-array was passed to the SAX parser |
388 | */ |
389 | bool parse_bson_array() |
390 | { |
391 | std::int32_t document_size{}; |
392 | get_number<std::int32_t, true>(input_format_t::bson, document_size); |
393 | |
394 | if (JSON_HEDLEY_UNLIKELY(!sax->start_array(static_cast<std::size_t>(-1)))) |
395 | { |
396 | return false; |
397 | } |
398 | |
399 | if (JSON_HEDLEY_UNLIKELY(!parse_bson_element_list(/*is_array*/true))) |
400 | { |
401 | return false; |
402 | } |
403 | |
404 | return sax->end_array(); |
405 | } |
406 | |
407 | ////////// |
408 | // CBOR // |
409 | ////////// |
410 | |
411 | /*! |
412 | @param[in] get_char whether a new character should be retrieved from the |
413 | input (true) or whether the last read character should |
414 | be considered instead (false) |
415 | @param[in] tag_handler how CBOR tags should be treated |
416 | |
417 | @return whether a valid CBOR value was passed to the SAX parser |
418 | */ |
419 | bool parse_cbor_internal(const bool get_char, |
420 | const cbor_tag_handler_t tag_handler) |
421 | { |
422 | switch (get_char ? get() : current) |
423 | { |
424 | // EOF |
425 | case char_traits<char_type>::eof(): |
426 | return unexpect_eof(format: input_format_t::cbor, context: "value" ); |
427 | |
428 | // Integer 0x00..0x17 (0..23) |
429 | case 0x00: |
430 | case 0x01: |
431 | case 0x02: |
432 | case 0x03: |
433 | case 0x04: |
434 | case 0x05: |
435 | case 0x06: |
436 | case 0x07: |
437 | case 0x08: |
438 | case 0x09: |
439 | case 0x0A: |
440 | case 0x0B: |
441 | case 0x0C: |
442 | case 0x0D: |
443 | case 0x0E: |
444 | case 0x0F: |
445 | case 0x10: |
446 | case 0x11: |
447 | case 0x12: |
448 | case 0x13: |
449 | case 0x14: |
450 | case 0x15: |
451 | case 0x16: |
452 | case 0x17: |
453 | return sax->number_unsigned(static_cast<number_unsigned_t>(current)); |
454 | |
455 | case 0x18: // Unsigned integer (one-byte uint8_t follows) |
456 | { |
457 | std::uint8_t number{}; |
458 | return get_number(input_format_t::cbor, number) && sax->number_unsigned(number); |
459 | } |
460 | |
461 | case 0x19: // Unsigned integer (two-byte uint16_t follows) |
462 | { |
463 | std::uint16_t number{}; |
464 | return get_number(input_format_t::cbor, number) && sax->number_unsigned(number); |
465 | } |
466 | |
467 | case 0x1A: // Unsigned integer (four-byte uint32_t follows) |
468 | { |
469 | std::uint32_t number{}; |
470 | return get_number(input_format_t::cbor, number) && sax->number_unsigned(number); |
471 | } |
472 | |
473 | case 0x1B: // Unsigned integer (eight-byte uint64_t follows) |
474 | { |
475 | std::uint64_t number{}; |
476 | return get_number(input_format_t::cbor, number) && sax->number_unsigned(number); |
477 | } |
478 | |
479 | // Negative integer -1-0x00..-1-0x17 (-1..-24) |
480 | case 0x20: |
481 | case 0x21: |
482 | case 0x22: |
483 | case 0x23: |
484 | case 0x24: |
485 | case 0x25: |
486 | case 0x26: |
487 | case 0x27: |
488 | case 0x28: |
489 | case 0x29: |
490 | case 0x2A: |
491 | case 0x2B: |
492 | case 0x2C: |
493 | case 0x2D: |
494 | case 0x2E: |
495 | case 0x2F: |
496 | case 0x30: |
497 | case 0x31: |
498 | case 0x32: |
499 | case 0x33: |
500 | case 0x34: |
501 | case 0x35: |
502 | case 0x36: |
503 | case 0x37: |
504 | return sax->number_integer(static_cast<std::int8_t>(0x20 - 1 - current)); |
505 | |
506 | case 0x38: // Negative integer (one-byte uint8_t follows) |
507 | { |
508 | std::uint8_t number{}; |
509 | return get_number(input_format_t::cbor, number) && sax->number_integer(static_cast<number_integer_t>(-1) - number); |
510 | } |
511 | |
512 | case 0x39: // Negative integer -1-n (two-byte uint16_t follows) |
513 | { |
514 | std::uint16_t number{}; |
515 | return get_number(input_format_t::cbor, number) && sax->number_integer(static_cast<number_integer_t>(-1) - number); |
516 | } |
517 | |
518 | case 0x3A: // Negative integer -1-n (four-byte uint32_t follows) |
519 | { |
520 | std::uint32_t number{}; |
521 | return get_number(input_format_t::cbor, number) && sax->number_integer(static_cast<number_integer_t>(-1) - number); |
522 | } |
523 | |
524 | case 0x3B: // Negative integer -1-n (eight-byte uint64_t follows) |
525 | { |
526 | std::uint64_t number{}; |
527 | return get_number(input_format_t::cbor, number) && sax->number_integer(static_cast<number_integer_t>(-1) |
528 | - static_cast<number_integer_t>(number)); |
529 | } |
530 | |
531 | // Binary data (0x00..0x17 bytes follow) |
532 | case 0x40: |
533 | case 0x41: |
534 | case 0x42: |
535 | case 0x43: |
536 | case 0x44: |
537 | case 0x45: |
538 | case 0x46: |
539 | case 0x47: |
540 | case 0x48: |
541 | case 0x49: |
542 | case 0x4A: |
543 | case 0x4B: |
544 | case 0x4C: |
545 | case 0x4D: |
546 | case 0x4E: |
547 | case 0x4F: |
548 | case 0x50: |
549 | case 0x51: |
550 | case 0x52: |
551 | case 0x53: |
552 | case 0x54: |
553 | case 0x55: |
554 | case 0x56: |
555 | case 0x57: |
556 | case 0x58: // Binary data (one-byte uint8_t for n follows) |
557 | case 0x59: // Binary data (two-byte uint16_t for n follow) |
558 | case 0x5A: // Binary data (four-byte uint32_t for n follow) |
559 | case 0x5B: // Binary data (eight-byte uint64_t for n follow) |
560 | case 0x5F: // Binary data (indefinite length) |
561 | { |
562 | binary_t b; |
563 | return get_cbor_binary(result&: b) && sax->binary(b); |
564 | } |
565 | |
566 | // UTF-8 string (0x00..0x17 bytes follow) |
567 | case 0x60: |
568 | case 0x61: |
569 | case 0x62: |
570 | case 0x63: |
571 | case 0x64: |
572 | case 0x65: |
573 | case 0x66: |
574 | case 0x67: |
575 | case 0x68: |
576 | case 0x69: |
577 | case 0x6A: |
578 | case 0x6B: |
579 | case 0x6C: |
580 | case 0x6D: |
581 | case 0x6E: |
582 | case 0x6F: |
583 | case 0x70: |
584 | case 0x71: |
585 | case 0x72: |
586 | case 0x73: |
587 | case 0x74: |
588 | case 0x75: |
589 | case 0x76: |
590 | case 0x77: |
591 | case 0x78: // UTF-8 string (one-byte uint8_t for n follows) |
592 | case 0x79: // UTF-8 string (two-byte uint16_t for n follow) |
593 | case 0x7A: // UTF-8 string (four-byte uint32_t for n follow) |
594 | case 0x7B: // UTF-8 string (eight-byte uint64_t for n follow) |
595 | case 0x7F: // UTF-8 string (indefinite length) |
596 | { |
597 | string_t s; |
598 | return get_cbor_string(result&: s) && sax->string(s); |
599 | } |
600 | |
601 | // array (0x00..0x17 data items follow) |
602 | case 0x80: |
603 | case 0x81: |
604 | case 0x82: |
605 | case 0x83: |
606 | case 0x84: |
607 | case 0x85: |
608 | case 0x86: |
609 | case 0x87: |
610 | case 0x88: |
611 | case 0x89: |
612 | case 0x8A: |
613 | case 0x8B: |
614 | case 0x8C: |
615 | case 0x8D: |
616 | case 0x8E: |
617 | case 0x8F: |
618 | case 0x90: |
619 | case 0x91: |
620 | case 0x92: |
621 | case 0x93: |
622 | case 0x94: |
623 | case 0x95: |
624 | case 0x96: |
625 | case 0x97: |
626 | return get_cbor_array( |
627 | len: conditional_static_cast<std::size_t>(value: static_cast<unsigned int>(current) & 0x1Fu), tag_handler); |
628 | |
629 | case 0x98: // array (one-byte uint8_t for n follows) |
630 | { |
631 | std::uint8_t len{}; |
632 | return get_number(input_format_t::cbor, len) && get_cbor_array(len: static_cast<std::size_t>(len), tag_handler); |
633 | } |
634 | |
635 | case 0x99: // array (two-byte uint16_t for n follow) |
636 | { |
637 | std::uint16_t len{}; |
638 | return get_number(input_format_t::cbor, len) && get_cbor_array(len: static_cast<std::size_t>(len), tag_handler); |
639 | } |
640 | |
641 | case 0x9A: // array (four-byte uint32_t for n follow) |
642 | { |
643 | std::uint32_t len{}; |
644 | return get_number(input_format_t::cbor, len) && get_cbor_array(len: conditional_static_cast<std::size_t>(value: len), tag_handler); |
645 | } |
646 | |
647 | case 0x9B: // array (eight-byte uint64_t for n follow) |
648 | { |
649 | std::uint64_t len{}; |
650 | return get_number(input_format_t::cbor, len) && get_cbor_array(len: conditional_static_cast<std::size_t>(value: len), tag_handler); |
651 | } |
652 | |
653 | case 0x9F: // array (indefinite length) |
654 | return get_cbor_array(len: static_cast<std::size_t>(-1), tag_handler); |
655 | |
656 | // map (0x00..0x17 pairs of data items follow) |
657 | case 0xA0: |
658 | case 0xA1: |
659 | case 0xA2: |
660 | case 0xA3: |
661 | case 0xA4: |
662 | case 0xA5: |
663 | case 0xA6: |
664 | case 0xA7: |
665 | case 0xA8: |
666 | case 0xA9: |
667 | case 0xAA: |
668 | case 0xAB: |
669 | case 0xAC: |
670 | case 0xAD: |
671 | case 0xAE: |
672 | case 0xAF: |
673 | case 0xB0: |
674 | case 0xB1: |
675 | case 0xB2: |
676 | case 0xB3: |
677 | case 0xB4: |
678 | case 0xB5: |
679 | case 0xB6: |
680 | case 0xB7: |
681 | return get_cbor_object(len: conditional_static_cast<std::size_t>(value: static_cast<unsigned int>(current) & 0x1Fu), tag_handler); |
682 | |
683 | case 0xB8: // map (one-byte uint8_t for n follows) |
684 | { |
685 | std::uint8_t len{}; |
686 | return get_number(input_format_t::cbor, len) && get_cbor_object(len: static_cast<std::size_t>(len), tag_handler); |
687 | } |
688 | |
689 | case 0xB9: // map (two-byte uint16_t for n follow) |
690 | { |
691 | std::uint16_t len{}; |
692 | return get_number(input_format_t::cbor, len) && get_cbor_object(len: static_cast<std::size_t>(len), tag_handler); |
693 | } |
694 | |
695 | case 0xBA: // map (four-byte uint32_t for n follow) |
696 | { |
697 | std::uint32_t len{}; |
698 | return get_number(input_format_t::cbor, len) && get_cbor_object(len: conditional_static_cast<std::size_t>(value: len), tag_handler); |
699 | } |
700 | |
701 | case 0xBB: // map (eight-byte uint64_t for n follow) |
702 | { |
703 | std::uint64_t len{}; |
704 | return get_number(input_format_t::cbor, len) && get_cbor_object(len: conditional_static_cast<std::size_t>(value: len), tag_handler); |
705 | } |
706 | |
707 | case 0xBF: // map (indefinite length) |
708 | return get_cbor_object(len: static_cast<std::size_t>(-1), tag_handler); |
709 | |
710 | case 0xC6: // tagged item |
711 | case 0xC7: |
712 | case 0xC8: |
713 | case 0xC9: |
714 | case 0xCA: |
715 | case 0xCB: |
716 | case 0xCC: |
717 | case 0xCD: |
718 | case 0xCE: |
719 | case 0xCF: |
720 | case 0xD0: |
721 | case 0xD1: |
722 | case 0xD2: |
723 | case 0xD3: |
724 | case 0xD4: |
725 | case 0xD8: // tagged item (1 bytes follow) |
726 | case 0xD9: // tagged item (2 bytes follow) |
727 | case 0xDA: // tagged item (4 bytes follow) |
728 | case 0xDB: // tagged item (8 bytes follow) |
729 | { |
730 | switch (tag_handler) |
731 | { |
732 | case cbor_tag_handler_t::error: |
733 | { |
734 | auto last_token = get_token_string(); |
735 | return sax->parse_error(chars_read, last_token, parse_error::create(112, chars_read, |
736 | exception_message(format: input_format_t::cbor, detail: concat("invalid byte: 0x" , last_token), context: "value" ), nullptr)); |
737 | } |
738 | |
739 | case cbor_tag_handler_t::ignore: |
740 | { |
741 | // ignore binary subtype |
742 | switch (current) |
743 | { |
744 | case 0xD8: |
745 | { |
746 | std::uint8_t subtype_to_ignore{}; |
747 | get_number(input_format_t::cbor, subtype_to_ignore); |
748 | break; |
749 | } |
750 | case 0xD9: |
751 | { |
752 | std::uint16_t subtype_to_ignore{}; |
753 | get_number(input_format_t::cbor, subtype_to_ignore); |
754 | break; |
755 | } |
756 | case 0xDA: |
757 | { |
758 | std::uint32_t subtype_to_ignore{}; |
759 | get_number(input_format_t::cbor, subtype_to_ignore); |
760 | break; |
761 | } |
762 | case 0xDB: |
763 | { |
764 | std::uint64_t subtype_to_ignore{}; |
765 | get_number(input_format_t::cbor, subtype_to_ignore); |
766 | break; |
767 | } |
768 | default: |
769 | break; |
770 | } |
771 | return parse_cbor_internal(get_char: true, tag_handler); |
772 | } |
773 | |
774 | case cbor_tag_handler_t::store: |
775 | { |
776 | binary_t b; |
777 | // use binary subtype and store in binary container |
778 | switch (current) |
779 | { |
780 | case 0xD8: |
781 | { |
782 | std::uint8_t subtype{}; |
783 | get_number(input_format_t::cbor, subtype); |
784 | b.set_subtype(detail::conditional_static_cast<typename binary_t::subtype_type>(subtype)); |
785 | break; |
786 | } |
787 | case 0xD9: |
788 | { |
789 | std::uint16_t subtype{}; |
790 | get_number(input_format_t::cbor, subtype); |
791 | b.set_subtype(detail::conditional_static_cast<typename binary_t::subtype_type>(subtype)); |
792 | break; |
793 | } |
794 | case 0xDA: |
795 | { |
796 | std::uint32_t subtype{}; |
797 | get_number(input_format_t::cbor, subtype); |
798 | b.set_subtype(detail::conditional_static_cast<typename binary_t::subtype_type>(subtype)); |
799 | break; |
800 | } |
801 | case 0xDB: |
802 | { |
803 | std::uint64_t subtype{}; |
804 | get_number(input_format_t::cbor, subtype); |
805 | b.set_subtype(detail::conditional_static_cast<typename binary_t::subtype_type>(subtype)); |
806 | break; |
807 | } |
808 | default: |
809 | return parse_cbor_internal(get_char: true, tag_handler); |
810 | } |
811 | get(); |
812 | return get_cbor_binary(result&: b) && sax->binary(b); |
813 | } |
814 | |
815 | default: // LCOV_EXCL_LINE |
816 | JSON_ASSERT(false); // NOLINT(cert-dcl03-c,hicpp-static-assert,misc-static-assert) LCOV_EXCL_LINE |
817 | return false; // LCOV_EXCL_LINE |
818 | } |
819 | } |
820 | |
821 | case 0xF4: // false |
822 | return sax->boolean(false); |
823 | |
824 | case 0xF5: // true |
825 | return sax->boolean(true); |
826 | |
827 | case 0xF6: // null |
828 | return sax->null(); |
829 | |
830 | case 0xF9: // Half-Precision Float (two-byte IEEE 754) |
831 | { |
832 | const auto byte1_raw = get(); |
833 | if (JSON_HEDLEY_UNLIKELY(!unexpect_eof(input_format_t::cbor, "number" ))) |
834 | { |
835 | return false; |
836 | } |
837 | const auto byte2_raw = get(); |
838 | if (JSON_HEDLEY_UNLIKELY(!unexpect_eof(input_format_t::cbor, "number" ))) |
839 | { |
840 | return false; |
841 | } |
842 | |
843 | const auto byte1 = static_cast<unsigned char>(byte1_raw); |
844 | const auto byte2 = static_cast<unsigned char>(byte2_raw); |
845 | |
846 | // code from RFC 7049, Appendix D, Figure 3: |
847 | // As half-precision floating-point numbers were only added |
848 | // to IEEE 754 in 2008, today's programming platforms often |
849 | // still only have limited support for them. It is very |
850 | // easy to include at least decoding support for them even |
851 | // without such support. An example of a small decoder for |
852 | // half-precision floating-point numbers in the C language |
853 | // is shown in Fig. 3. |
854 | const auto half = static_cast<unsigned int>((byte1 << 8u) + byte2); |
855 | const double val = [&half] |
856 | { |
857 | const int exp = (half >> 10u) & 0x1Fu; |
858 | const unsigned int mant = half & 0x3FFu; |
859 | JSON_ASSERT(0 <= exp&& exp <= 32); |
860 | JSON_ASSERT(mant <= 1024); |
861 | switch (exp) |
862 | { |
863 | case 0: |
864 | return std::ldexp(x: mant, exp: -24); |
865 | case 31: |
866 | return (mant == 0) |
867 | ? std::numeric_limits<double>::infinity() |
868 | : std::numeric_limits<double>::quiet_NaN(); |
869 | default: |
870 | return std::ldexp(x: mant + 1024, exp: exp - 25); |
871 | } |
872 | }(); |
873 | return sax->number_float((half & 0x8000u) != 0 |
874 | ? static_cast<number_float_t>(-val) |
875 | : static_cast<number_float_t>(val), "" ); |
876 | } |
877 | |
878 | case 0xFA: // Single-Precision Float (four-byte IEEE 754) |
879 | { |
880 | float number{}; |
881 | return get_number(input_format_t::cbor, number) && sax->number_float(static_cast<number_float_t>(number), "" ); |
882 | } |
883 | |
884 | case 0xFB: // Double-Precision Float (eight-byte IEEE 754) |
885 | { |
886 | double number{}; |
887 | return get_number(input_format_t::cbor, number) && sax->number_float(static_cast<number_float_t>(number), "" ); |
888 | } |
889 | |
890 | default: // anything else (0xFF is handled inside the other types) |
891 | { |
892 | auto last_token = get_token_string(); |
893 | return sax->parse_error(chars_read, last_token, parse_error::create(112, chars_read, |
894 | exception_message(format: input_format_t::cbor, detail: concat("invalid byte: 0x" , last_token), context: "value" ), nullptr)); |
895 | } |
896 | } |
897 | } |
898 | |
899 | /*! |
900 | @brief reads a CBOR string |
901 | |
902 | This function first reads starting bytes to determine the expected |
903 | string length and then copies this number of bytes into a string. |
904 | Additionally, CBOR's strings with indefinite lengths are supported. |
905 | |
906 | @param[out] result created string |
907 | |
908 | @return whether string creation completed |
909 | */ |
910 | bool get_cbor_string(string_t& result) |
911 | { |
912 | if (JSON_HEDLEY_UNLIKELY(!unexpect_eof(input_format_t::cbor, "string" ))) |
913 | { |
914 | return false; |
915 | } |
916 | |
917 | switch (current) |
918 | { |
919 | // UTF-8 string (0x00..0x17 bytes follow) |
920 | case 0x60: |
921 | case 0x61: |
922 | case 0x62: |
923 | case 0x63: |
924 | case 0x64: |
925 | case 0x65: |
926 | case 0x66: |
927 | case 0x67: |
928 | case 0x68: |
929 | case 0x69: |
930 | case 0x6A: |
931 | case 0x6B: |
932 | case 0x6C: |
933 | case 0x6D: |
934 | case 0x6E: |
935 | case 0x6F: |
936 | case 0x70: |
937 | case 0x71: |
938 | case 0x72: |
939 | case 0x73: |
940 | case 0x74: |
941 | case 0x75: |
942 | case 0x76: |
943 | case 0x77: |
944 | { |
945 | return get_string(input_format_t::cbor, static_cast<unsigned int>(current) & 0x1Fu, result); |
946 | } |
947 | |
948 | case 0x78: // UTF-8 string (one-byte uint8_t for n follows) |
949 | { |
950 | std::uint8_t len{}; |
951 | return get_number(input_format_t::cbor, len) && get_string(input_format_t::cbor, len, result); |
952 | } |
953 | |
954 | case 0x79: // UTF-8 string (two-byte uint16_t for n follow) |
955 | { |
956 | std::uint16_t len{}; |
957 | return get_number(input_format_t::cbor, len) && get_string(input_format_t::cbor, len, result); |
958 | } |
959 | |
960 | case 0x7A: // UTF-8 string (four-byte uint32_t for n follow) |
961 | { |
962 | std::uint32_t len{}; |
963 | return get_number(input_format_t::cbor, len) && get_string(input_format_t::cbor, len, result); |
964 | } |
965 | |
966 | case 0x7B: // UTF-8 string (eight-byte uint64_t for n follow) |
967 | { |
968 | std::uint64_t len{}; |
969 | return get_number(input_format_t::cbor, len) && get_string(input_format_t::cbor, len, result); |
970 | } |
971 | |
972 | case 0x7F: // UTF-8 string (indefinite length) |
973 | { |
974 | while (get() != 0xFF) |
975 | { |
976 | string_t chunk; |
977 | if (!get_cbor_string(result&: chunk)) |
978 | { |
979 | return false; |
980 | } |
981 | result.append(chunk); |
982 | } |
983 | return true; |
984 | } |
985 | |
986 | default: |
987 | { |
988 | auto last_token = get_token_string(); |
989 | return sax->parse_error(chars_read, last_token, parse_error::create(113, chars_read, |
990 | exception_message(format: input_format_t::cbor, detail: concat("expected length specification (0x60-0x7B) or indefinite string type (0x7F); last byte: 0x" , last_token), context: "string" ), nullptr)); |
991 | } |
992 | } |
993 | } |
994 | |
995 | /*! |
996 | @brief reads a CBOR byte array |
997 | |
998 | This function first reads starting bytes to determine the expected |
999 | byte array length and then copies this number of bytes into the byte array. |
1000 | Additionally, CBOR's byte arrays with indefinite lengths are supported. |
1001 | |
1002 | @param[out] result created byte array |
1003 | |
1004 | @return whether byte array creation completed |
1005 | */ |
1006 | bool get_cbor_binary(binary_t& result) |
1007 | { |
1008 | if (JSON_HEDLEY_UNLIKELY(!unexpect_eof(input_format_t::cbor, "binary" ))) |
1009 | { |
1010 | return false; |
1011 | } |
1012 | |
1013 | switch (current) |
1014 | { |
1015 | // Binary data (0x00..0x17 bytes follow) |
1016 | case 0x40: |
1017 | case 0x41: |
1018 | case 0x42: |
1019 | case 0x43: |
1020 | case 0x44: |
1021 | case 0x45: |
1022 | case 0x46: |
1023 | case 0x47: |
1024 | case 0x48: |
1025 | case 0x49: |
1026 | case 0x4A: |
1027 | case 0x4B: |
1028 | case 0x4C: |
1029 | case 0x4D: |
1030 | case 0x4E: |
1031 | case 0x4F: |
1032 | case 0x50: |
1033 | case 0x51: |
1034 | case 0x52: |
1035 | case 0x53: |
1036 | case 0x54: |
1037 | case 0x55: |
1038 | case 0x56: |
1039 | case 0x57: |
1040 | { |
1041 | return get_binary(input_format_t::cbor, static_cast<unsigned int>(current) & 0x1Fu, result); |
1042 | } |
1043 | |
1044 | case 0x58: // Binary data (one-byte uint8_t for n follows) |
1045 | { |
1046 | std::uint8_t len{}; |
1047 | return get_number(input_format_t::cbor, len) && |
1048 | get_binary(input_format_t::cbor, len, result); |
1049 | } |
1050 | |
1051 | case 0x59: // Binary data (two-byte uint16_t for n follow) |
1052 | { |
1053 | std::uint16_t len{}; |
1054 | return get_number(input_format_t::cbor, len) && |
1055 | get_binary(input_format_t::cbor, len, result); |
1056 | } |
1057 | |
1058 | case 0x5A: // Binary data (four-byte uint32_t for n follow) |
1059 | { |
1060 | std::uint32_t len{}; |
1061 | return get_number(input_format_t::cbor, len) && |
1062 | get_binary(input_format_t::cbor, len, result); |
1063 | } |
1064 | |
1065 | case 0x5B: // Binary data (eight-byte uint64_t for n follow) |
1066 | { |
1067 | std::uint64_t len{}; |
1068 | return get_number(input_format_t::cbor, len) && |
1069 | get_binary(input_format_t::cbor, len, result); |
1070 | } |
1071 | |
1072 | case 0x5F: // Binary data (indefinite length) |
1073 | { |
1074 | while (get() != 0xFF) |
1075 | { |
1076 | binary_t chunk; |
1077 | if (!get_cbor_binary(result&: chunk)) |
1078 | { |
1079 | return false; |
1080 | } |
1081 | result.insert(result.end(), chunk.begin(), chunk.end()); |
1082 | } |
1083 | return true; |
1084 | } |
1085 | |
1086 | default: |
1087 | { |
1088 | auto last_token = get_token_string(); |
1089 | return sax->parse_error(chars_read, last_token, parse_error::create(113, chars_read, |
1090 | exception_message(format: input_format_t::cbor, detail: concat("expected length specification (0x40-0x5B) or indefinite binary array type (0x5F); last byte: 0x" , last_token), context: "binary" ), nullptr)); |
1091 | } |
1092 | } |
1093 | } |
1094 | |
1095 | /*! |
1096 | @param[in] len the length of the array or static_cast<std::size_t>(-1) for an |
1097 | array of indefinite size |
1098 | @param[in] tag_handler how CBOR tags should be treated |
1099 | @return whether array creation completed |
1100 | */ |
1101 | bool get_cbor_array(const std::size_t len, |
1102 | const cbor_tag_handler_t tag_handler) |
1103 | { |
1104 | if (JSON_HEDLEY_UNLIKELY(!sax->start_array(len))) |
1105 | { |
1106 | return false; |
1107 | } |
1108 | |
1109 | if (len != static_cast<std::size_t>(-1)) |
1110 | { |
1111 | for (std::size_t i = 0; i < len; ++i) |
1112 | { |
1113 | if (JSON_HEDLEY_UNLIKELY(!parse_cbor_internal(true, tag_handler))) |
1114 | { |
1115 | return false; |
1116 | } |
1117 | } |
1118 | } |
1119 | else |
1120 | { |
1121 | while (get() != 0xFF) |
1122 | { |
1123 | if (JSON_HEDLEY_UNLIKELY(!parse_cbor_internal(false, tag_handler))) |
1124 | { |
1125 | return false; |
1126 | } |
1127 | } |
1128 | } |
1129 | |
1130 | return sax->end_array(); |
1131 | } |
1132 | |
1133 | /*! |
1134 | @param[in] len the length of the object or static_cast<std::size_t>(-1) for an |
1135 | object of indefinite size |
1136 | @param[in] tag_handler how CBOR tags should be treated |
1137 | @return whether object creation completed |
1138 | */ |
1139 | bool get_cbor_object(const std::size_t len, |
1140 | const cbor_tag_handler_t tag_handler) |
1141 | { |
1142 | if (JSON_HEDLEY_UNLIKELY(!sax->start_object(len))) |
1143 | { |
1144 | return false; |
1145 | } |
1146 | |
1147 | if (len != 0) |
1148 | { |
1149 | string_t key; |
1150 | if (len != static_cast<std::size_t>(-1)) |
1151 | { |
1152 | for (std::size_t i = 0; i < len; ++i) |
1153 | { |
1154 | get(); |
1155 | if (JSON_HEDLEY_UNLIKELY(!get_cbor_string(key) || !sax->key(key))) |
1156 | { |
1157 | return false; |
1158 | } |
1159 | |
1160 | if (JSON_HEDLEY_UNLIKELY(!parse_cbor_internal(true, tag_handler))) |
1161 | { |
1162 | return false; |
1163 | } |
1164 | key.clear(); |
1165 | } |
1166 | } |
1167 | else |
1168 | { |
1169 | while (get() != 0xFF) |
1170 | { |
1171 | if (JSON_HEDLEY_UNLIKELY(!get_cbor_string(key) || !sax->key(key))) |
1172 | { |
1173 | return false; |
1174 | } |
1175 | |
1176 | if (JSON_HEDLEY_UNLIKELY(!parse_cbor_internal(true, tag_handler))) |
1177 | { |
1178 | return false; |
1179 | } |
1180 | key.clear(); |
1181 | } |
1182 | } |
1183 | } |
1184 | |
1185 | return sax->end_object(); |
1186 | } |
1187 | |
1188 | ///////////// |
1189 | // MsgPack // |
1190 | ///////////// |
1191 | |
1192 | /*! |
1193 | @return whether a valid MessagePack value was passed to the SAX parser |
1194 | */ |
1195 | bool parse_msgpack_internal() |
1196 | { |
1197 | switch (get()) |
1198 | { |
1199 | // EOF |
1200 | case char_traits<char_type>::eof(): |
1201 | return unexpect_eof(format: input_format_t::msgpack, context: "value" ); |
1202 | |
1203 | // positive fixint |
1204 | case 0x00: |
1205 | case 0x01: |
1206 | case 0x02: |
1207 | case 0x03: |
1208 | case 0x04: |
1209 | case 0x05: |
1210 | case 0x06: |
1211 | case 0x07: |
1212 | case 0x08: |
1213 | case 0x09: |
1214 | case 0x0A: |
1215 | case 0x0B: |
1216 | case 0x0C: |
1217 | case 0x0D: |
1218 | case 0x0E: |
1219 | case 0x0F: |
1220 | case 0x10: |
1221 | case 0x11: |
1222 | case 0x12: |
1223 | case 0x13: |
1224 | case 0x14: |
1225 | case 0x15: |
1226 | case 0x16: |
1227 | case 0x17: |
1228 | case 0x18: |
1229 | case 0x19: |
1230 | case 0x1A: |
1231 | case 0x1B: |
1232 | case 0x1C: |
1233 | case 0x1D: |
1234 | case 0x1E: |
1235 | case 0x1F: |
1236 | case 0x20: |
1237 | case 0x21: |
1238 | case 0x22: |
1239 | case 0x23: |
1240 | case 0x24: |
1241 | case 0x25: |
1242 | case 0x26: |
1243 | case 0x27: |
1244 | case 0x28: |
1245 | case 0x29: |
1246 | case 0x2A: |
1247 | case 0x2B: |
1248 | case 0x2C: |
1249 | case 0x2D: |
1250 | case 0x2E: |
1251 | case 0x2F: |
1252 | case 0x30: |
1253 | case 0x31: |
1254 | case 0x32: |
1255 | case 0x33: |
1256 | case 0x34: |
1257 | case 0x35: |
1258 | case 0x36: |
1259 | case 0x37: |
1260 | case 0x38: |
1261 | case 0x39: |
1262 | case 0x3A: |
1263 | case 0x3B: |
1264 | case 0x3C: |
1265 | case 0x3D: |
1266 | case 0x3E: |
1267 | case 0x3F: |
1268 | case 0x40: |
1269 | case 0x41: |
1270 | case 0x42: |
1271 | case 0x43: |
1272 | case 0x44: |
1273 | case 0x45: |
1274 | case 0x46: |
1275 | case 0x47: |
1276 | case 0x48: |
1277 | case 0x49: |
1278 | case 0x4A: |
1279 | case 0x4B: |
1280 | case 0x4C: |
1281 | case 0x4D: |
1282 | case 0x4E: |
1283 | case 0x4F: |
1284 | case 0x50: |
1285 | case 0x51: |
1286 | case 0x52: |
1287 | case 0x53: |
1288 | case 0x54: |
1289 | case 0x55: |
1290 | case 0x56: |
1291 | case 0x57: |
1292 | case 0x58: |
1293 | case 0x59: |
1294 | case 0x5A: |
1295 | case 0x5B: |
1296 | case 0x5C: |
1297 | case 0x5D: |
1298 | case 0x5E: |
1299 | case 0x5F: |
1300 | case 0x60: |
1301 | case 0x61: |
1302 | case 0x62: |
1303 | case 0x63: |
1304 | case 0x64: |
1305 | case 0x65: |
1306 | case 0x66: |
1307 | case 0x67: |
1308 | case 0x68: |
1309 | case 0x69: |
1310 | case 0x6A: |
1311 | case 0x6B: |
1312 | case 0x6C: |
1313 | case 0x6D: |
1314 | case 0x6E: |
1315 | case 0x6F: |
1316 | case 0x70: |
1317 | case 0x71: |
1318 | case 0x72: |
1319 | case 0x73: |
1320 | case 0x74: |
1321 | case 0x75: |
1322 | case 0x76: |
1323 | case 0x77: |
1324 | case 0x78: |
1325 | case 0x79: |
1326 | case 0x7A: |
1327 | case 0x7B: |
1328 | case 0x7C: |
1329 | case 0x7D: |
1330 | case 0x7E: |
1331 | case 0x7F: |
1332 | return sax->number_unsigned(static_cast<number_unsigned_t>(current)); |
1333 | |
1334 | // fixmap |
1335 | case 0x80: |
1336 | case 0x81: |
1337 | case 0x82: |
1338 | case 0x83: |
1339 | case 0x84: |
1340 | case 0x85: |
1341 | case 0x86: |
1342 | case 0x87: |
1343 | case 0x88: |
1344 | case 0x89: |
1345 | case 0x8A: |
1346 | case 0x8B: |
1347 | case 0x8C: |
1348 | case 0x8D: |
1349 | case 0x8E: |
1350 | case 0x8F: |
1351 | return get_msgpack_object(len: conditional_static_cast<std::size_t>(value: static_cast<unsigned int>(current) & 0x0Fu)); |
1352 | |
1353 | // fixarray |
1354 | case 0x90: |
1355 | case 0x91: |
1356 | case 0x92: |
1357 | case 0x93: |
1358 | case 0x94: |
1359 | case 0x95: |
1360 | case 0x96: |
1361 | case 0x97: |
1362 | case 0x98: |
1363 | case 0x99: |
1364 | case 0x9A: |
1365 | case 0x9B: |
1366 | case 0x9C: |
1367 | case 0x9D: |
1368 | case 0x9E: |
1369 | case 0x9F: |
1370 | return get_msgpack_array(len: conditional_static_cast<std::size_t>(value: static_cast<unsigned int>(current) & 0x0Fu)); |
1371 | |
1372 | // fixstr |
1373 | case 0xA0: |
1374 | case 0xA1: |
1375 | case 0xA2: |
1376 | case 0xA3: |
1377 | case 0xA4: |
1378 | case 0xA5: |
1379 | case 0xA6: |
1380 | case 0xA7: |
1381 | case 0xA8: |
1382 | case 0xA9: |
1383 | case 0xAA: |
1384 | case 0xAB: |
1385 | case 0xAC: |
1386 | case 0xAD: |
1387 | case 0xAE: |
1388 | case 0xAF: |
1389 | case 0xB0: |
1390 | case 0xB1: |
1391 | case 0xB2: |
1392 | case 0xB3: |
1393 | case 0xB4: |
1394 | case 0xB5: |
1395 | case 0xB6: |
1396 | case 0xB7: |
1397 | case 0xB8: |
1398 | case 0xB9: |
1399 | case 0xBA: |
1400 | case 0xBB: |
1401 | case 0xBC: |
1402 | case 0xBD: |
1403 | case 0xBE: |
1404 | case 0xBF: |
1405 | case 0xD9: // str 8 |
1406 | case 0xDA: // str 16 |
1407 | case 0xDB: // str 32 |
1408 | { |
1409 | string_t s; |
1410 | return get_msgpack_string(result&: s) && sax->string(s); |
1411 | } |
1412 | |
1413 | case 0xC0: // nil |
1414 | return sax->null(); |
1415 | |
1416 | case 0xC2: // false |
1417 | return sax->boolean(false); |
1418 | |
1419 | case 0xC3: // true |
1420 | return sax->boolean(true); |
1421 | |
1422 | case 0xC4: // bin 8 |
1423 | case 0xC5: // bin 16 |
1424 | case 0xC6: // bin 32 |
1425 | case 0xC7: // ext 8 |
1426 | case 0xC8: // ext 16 |
1427 | case 0xC9: // ext 32 |
1428 | case 0xD4: // fixext 1 |
1429 | case 0xD5: // fixext 2 |
1430 | case 0xD6: // fixext 4 |
1431 | case 0xD7: // fixext 8 |
1432 | case 0xD8: // fixext 16 |
1433 | { |
1434 | binary_t b; |
1435 | return get_msgpack_binary(result&: b) && sax->binary(b); |
1436 | } |
1437 | |
1438 | case 0xCA: // float 32 |
1439 | { |
1440 | float number{}; |
1441 | return get_number(input_format_t::msgpack, number) && sax->number_float(static_cast<number_float_t>(number), "" ); |
1442 | } |
1443 | |
1444 | case 0xCB: // float 64 |
1445 | { |
1446 | double number{}; |
1447 | return get_number(input_format_t::msgpack, number) && sax->number_float(static_cast<number_float_t>(number), "" ); |
1448 | } |
1449 | |
1450 | case 0xCC: // uint 8 |
1451 | { |
1452 | std::uint8_t number{}; |
1453 | return get_number(input_format_t::msgpack, number) && sax->number_unsigned(number); |
1454 | } |
1455 | |
1456 | case 0xCD: // uint 16 |
1457 | { |
1458 | std::uint16_t number{}; |
1459 | return get_number(input_format_t::msgpack, number) && sax->number_unsigned(number); |
1460 | } |
1461 | |
1462 | case 0xCE: // uint 32 |
1463 | { |
1464 | std::uint32_t number{}; |
1465 | return get_number(input_format_t::msgpack, number) && sax->number_unsigned(number); |
1466 | } |
1467 | |
1468 | case 0xCF: // uint 64 |
1469 | { |
1470 | std::uint64_t number{}; |
1471 | return get_number(input_format_t::msgpack, number) && sax->number_unsigned(number); |
1472 | } |
1473 | |
1474 | case 0xD0: // int 8 |
1475 | { |
1476 | std::int8_t number{}; |
1477 | return get_number(input_format_t::msgpack, number) && sax->number_integer(number); |
1478 | } |
1479 | |
1480 | case 0xD1: // int 16 |
1481 | { |
1482 | std::int16_t number{}; |
1483 | return get_number(input_format_t::msgpack, number) && sax->number_integer(number); |
1484 | } |
1485 | |
1486 | case 0xD2: // int 32 |
1487 | { |
1488 | std::int32_t number{}; |
1489 | return get_number(input_format_t::msgpack, number) && sax->number_integer(number); |
1490 | } |
1491 | |
1492 | case 0xD3: // int 64 |
1493 | { |
1494 | std::int64_t number{}; |
1495 | return get_number(input_format_t::msgpack, number) && sax->number_integer(number); |
1496 | } |
1497 | |
1498 | case 0xDC: // array 16 |
1499 | { |
1500 | std::uint16_t len{}; |
1501 | return get_number(input_format_t::msgpack, len) && get_msgpack_array(len: static_cast<std::size_t>(len)); |
1502 | } |
1503 | |
1504 | case 0xDD: // array 32 |
1505 | { |
1506 | std::uint32_t len{}; |
1507 | return get_number(input_format_t::msgpack, len) && get_msgpack_array(len: conditional_static_cast<std::size_t>(value: len)); |
1508 | } |
1509 | |
1510 | case 0xDE: // map 16 |
1511 | { |
1512 | std::uint16_t len{}; |
1513 | return get_number(input_format_t::msgpack, len) && get_msgpack_object(len: static_cast<std::size_t>(len)); |
1514 | } |
1515 | |
1516 | case 0xDF: // map 32 |
1517 | { |
1518 | std::uint32_t len{}; |
1519 | return get_number(input_format_t::msgpack, len) && get_msgpack_object(len: conditional_static_cast<std::size_t>(value: len)); |
1520 | } |
1521 | |
1522 | // negative fixint |
1523 | case 0xE0: |
1524 | case 0xE1: |
1525 | case 0xE2: |
1526 | case 0xE3: |
1527 | case 0xE4: |
1528 | case 0xE5: |
1529 | case 0xE6: |
1530 | case 0xE7: |
1531 | case 0xE8: |
1532 | case 0xE9: |
1533 | case 0xEA: |
1534 | case 0xEB: |
1535 | case 0xEC: |
1536 | case 0xED: |
1537 | case 0xEE: |
1538 | case 0xEF: |
1539 | case 0xF0: |
1540 | case 0xF1: |
1541 | case 0xF2: |
1542 | case 0xF3: |
1543 | case 0xF4: |
1544 | case 0xF5: |
1545 | case 0xF6: |
1546 | case 0xF7: |
1547 | case 0xF8: |
1548 | case 0xF9: |
1549 | case 0xFA: |
1550 | case 0xFB: |
1551 | case 0xFC: |
1552 | case 0xFD: |
1553 | case 0xFE: |
1554 | case 0xFF: |
1555 | return sax->number_integer(static_cast<std::int8_t>(current)); |
1556 | |
1557 | default: // anything else |
1558 | { |
1559 | auto last_token = get_token_string(); |
1560 | return sax->parse_error(chars_read, last_token, parse_error::create(112, chars_read, |
1561 | exception_message(format: input_format_t::msgpack, detail: concat("invalid byte: 0x" , last_token), context: "value" ), nullptr)); |
1562 | } |
1563 | } |
1564 | } |
1565 | |
1566 | /*! |
1567 | @brief reads a MessagePack string |
1568 | |
1569 | This function first reads starting bytes to determine the expected |
1570 | string length and then copies this number of bytes into a string. |
1571 | |
1572 | @param[out] result created string |
1573 | |
1574 | @return whether string creation completed |
1575 | */ |
1576 | bool get_msgpack_string(string_t& result) |
1577 | { |
1578 | if (JSON_HEDLEY_UNLIKELY(!unexpect_eof(input_format_t::msgpack, "string" ))) |
1579 | { |
1580 | return false; |
1581 | } |
1582 | |
1583 | switch (current) |
1584 | { |
1585 | // fixstr |
1586 | case 0xA0: |
1587 | case 0xA1: |
1588 | case 0xA2: |
1589 | case 0xA3: |
1590 | case 0xA4: |
1591 | case 0xA5: |
1592 | case 0xA6: |
1593 | case 0xA7: |
1594 | case 0xA8: |
1595 | case 0xA9: |
1596 | case 0xAA: |
1597 | case 0xAB: |
1598 | case 0xAC: |
1599 | case 0xAD: |
1600 | case 0xAE: |
1601 | case 0xAF: |
1602 | case 0xB0: |
1603 | case 0xB1: |
1604 | case 0xB2: |
1605 | case 0xB3: |
1606 | case 0xB4: |
1607 | case 0xB5: |
1608 | case 0xB6: |
1609 | case 0xB7: |
1610 | case 0xB8: |
1611 | case 0xB9: |
1612 | case 0xBA: |
1613 | case 0xBB: |
1614 | case 0xBC: |
1615 | case 0xBD: |
1616 | case 0xBE: |
1617 | case 0xBF: |
1618 | { |
1619 | return get_string(input_format_t::msgpack, static_cast<unsigned int>(current) & 0x1Fu, result); |
1620 | } |
1621 | |
1622 | case 0xD9: // str 8 |
1623 | { |
1624 | std::uint8_t len{}; |
1625 | return get_number(input_format_t::msgpack, len) && get_string(input_format_t::msgpack, len, result); |
1626 | } |
1627 | |
1628 | case 0xDA: // str 16 |
1629 | { |
1630 | std::uint16_t len{}; |
1631 | return get_number(input_format_t::msgpack, len) && get_string(input_format_t::msgpack, len, result); |
1632 | } |
1633 | |
1634 | case 0xDB: // str 32 |
1635 | { |
1636 | std::uint32_t len{}; |
1637 | return get_number(input_format_t::msgpack, len) && get_string(input_format_t::msgpack, len, result); |
1638 | } |
1639 | |
1640 | default: |
1641 | { |
1642 | auto last_token = get_token_string(); |
1643 | return sax->parse_error(chars_read, last_token, parse_error::create(113, chars_read, |
1644 | exception_message(format: input_format_t::msgpack, detail: concat("expected length specification (0xA0-0xBF, 0xD9-0xDB); last byte: 0x" , last_token), context: "string" ), nullptr)); |
1645 | } |
1646 | } |
1647 | } |
1648 | |
1649 | /*! |
1650 | @brief reads a MessagePack byte array |
1651 | |
1652 | This function first reads starting bytes to determine the expected |
1653 | byte array length and then copies this number of bytes into a byte array. |
1654 | |
1655 | @param[out] result created byte array |
1656 | |
1657 | @return whether byte array creation completed |
1658 | */ |
1659 | bool get_msgpack_binary(binary_t& result) |
1660 | { |
1661 | // helper function to set the subtype |
1662 | auto assign_and_return_true = [&result](std::int8_t subtype) |
1663 | { |
1664 | result.set_subtype(static_cast<std::uint8_t>(subtype)); |
1665 | return true; |
1666 | }; |
1667 | |
1668 | switch (current) |
1669 | { |
1670 | case 0xC4: // bin 8 |
1671 | { |
1672 | std::uint8_t len{}; |
1673 | return get_number(input_format_t::msgpack, len) && |
1674 | get_binary(input_format_t::msgpack, len, result); |
1675 | } |
1676 | |
1677 | case 0xC5: // bin 16 |
1678 | { |
1679 | std::uint16_t len{}; |
1680 | return get_number(input_format_t::msgpack, len) && |
1681 | get_binary(input_format_t::msgpack, len, result); |
1682 | } |
1683 | |
1684 | case 0xC6: // bin 32 |
1685 | { |
1686 | std::uint32_t len{}; |
1687 | return get_number(input_format_t::msgpack, len) && |
1688 | get_binary(input_format_t::msgpack, len, result); |
1689 | } |
1690 | |
1691 | case 0xC7: // ext 8 |
1692 | { |
1693 | std::uint8_t len{}; |
1694 | std::int8_t subtype{}; |
1695 | return get_number(input_format_t::msgpack, len) && |
1696 | get_number(input_format_t::msgpack, subtype) && |
1697 | get_binary(input_format_t::msgpack, len, result) && |
1698 | assign_and_return_true(subtype); |
1699 | } |
1700 | |
1701 | case 0xC8: // ext 16 |
1702 | { |
1703 | std::uint16_t len{}; |
1704 | std::int8_t subtype{}; |
1705 | return get_number(input_format_t::msgpack, len) && |
1706 | get_number(input_format_t::msgpack, subtype) && |
1707 | get_binary(input_format_t::msgpack, len, result) && |
1708 | assign_and_return_true(subtype); |
1709 | } |
1710 | |
1711 | case 0xC9: // ext 32 |
1712 | { |
1713 | std::uint32_t len{}; |
1714 | std::int8_t subtype{}; |
1715 | return get_number(input_format_t::msgpack, len) && |
1716 | get_number(input_format_t::msgpack, subtype) && |
1717 | get_binary(input_format_t::msgpack, len, result) && |
1718 | assign_and_return_true(subtype); |
1719 | } |
1720 | |
1721 | case 0xD4: // fixext 1 |
1722 | { |
1723 | std::int8_t subtype{}; |
1724 | return get_number(input_format_t::msgpack, subtype) && |
1725 | get_binary(input_format_t::msgpack, 1, result) && |
1726 | assign_and_return_true(subtype); |
1727 | } |
1728 | |
1729 | case 0xD5: // fixext 2 |
1730 | { |
1731 | std::int8_t subtype{}; |
1732 | return get_number(input_format_t::msgpack, subtype) && |
1733 | get_binary(input_format_t::msgpack, 2, result) && |
1734 | assign_and_return_true(subtype); |
1735 | } |
1736 | |
1737 | case 0xD6: // fixext 4 |
1738 | { |
1739 | std::int8_t subtype{}; |
1740 | return get_number(input_format_t::msgpack, subtype) && |
1741 | get_binary(input_format_t::msgpack, 4, result) && |
1742 | assign_and_return_true(subtype); |
1743 | } |
1744 | |
1745 | case 0xD7: // fixext 8 |
1746 | { |
1747 | std::int8_t subtype{}; |
1748 | return get_number(input_format_t::msgpack, subtype) && |
1749 | get_binary(input_format_t::msgpack, 8, result) && |
1750 | assign_and_return_true(subtype); |
1751 | } |
1752 | |
1753 | case 0xD8: // fixext 16 |
1754 | { |
1755 | std::int8_t subtype{}; |
1756 | return get_number(input_format_t::msgpack, subtype) && |
1757 | get_binary(input_format_t::msgpack, 16, result) && |
1758 | assign_and_return_true(subtype); |
1759 | } |
1760 | |
1761 | default: // LCOV_EXCL_LINE |
1762 | return false; // LCOV_EXCL_LINE |
1763 | } |
1764 | } |
1765 | |
1766 | /*! |
1767 | @param[in] len the length of the array |
1768 | @return whether array creation completed |
1769 | */ |
1770 | bool get_msgpack_array(const std::size_t len) |
1771 | { |
1772 | if (JSON_HEDLEY_UNLIKELY(!sax->start_array(len))) |
1773 | { |
1774 | return false; |
1775 | } |
1776 | |
1777 | for (std::size_t i = 0; i < len; ++i) |
1778 | { |
1779 | if (JSON_HEDLEY_UNLIKELY(!parse_msgpack_internal())) |
1780 | { |
1781 | return false; |
1782 | } |
1783 | } |
1784 | |
1785 | return sax->end_array(); |
1786 | } |
1787 | |
1788 | /*! |
1789 | @param[in] len the length of the object |
1790 | @return whether object creation completed |
1791 | */ |
1792 | bool get_msgpack_object(const std::size_t len) |
1793 | { |
1794 | if (JSON_HEDLEY_UNLIKELY(!sax->start_object(len))) |
1795 | { |
1796 | return false; |
1797 | } |
1798 | |
1799 | string_t key; |
1800 | for (std::size_t i = 0; i < len; ++i) |
1801 | { |
1802 | get(); |
1803 | if (JSON_HEDLEY_UNLIKELY(!get_msgpack_string(key) || !sax->key(key))) |
1804 | { |
1805 | return false; |
1806 | } |
1807 | |
1808 | if (JSON_HEDLEY_UNLIKELY(!parse_msgpack_internal())) |
1809 | { |
1810 | return false; |
1811 | } |
1812 | key.clear(); |
1813 | } |
1814 | |
1815 | return sax->end_object(); |
1816 | } |
1817 | |
1818 | //////////// |
1819 | // UBJSON // |
1820 | //////////// |
1821 | |
1822 | /*! |
1823 | @param[in] get_char whether a new character should be retrieved from the |
1824 | input (true, default) or whether the last read |
1825 | character should be considered instead |
1826 | |
1827 | @return whether a valid UBJSON value was passed to the SAX parser |
1828 | */ |
1829 | bool parse_ubjson_internal(const bool get_char = true) |
1830 | { |
1831 | return get_ubjson_value(prefix: get_char ? get_ignore_noop() : current); |
1832 | } |
1833 | |
1834 | /*! |
1835 | @brief reads a UBJSON string |
1836 | |
1837 | This function is either called after reading the 'S' byte explicitly |
1838 | indicating a string, or in case of an object key where the 'S' byte can be |
1839 | left out. |
1840 | |
1841 | @param[out] result created string |
1842 | @param[in] get_char whether a new character should be retrieved from the |
1843 | input (true, default) or whether the last read |
1844 | character should be considered instead |
1845 | |
1846 | @return whether string creation completed |
1847 | */ |
1848 | bool get_ubjson_string(string_t& result, const bool get_char = true) |
1849 | { |
1850 | if (get_char) |
1851 | { |
1852 | get(); // TODO(niels): may we ignore N here? |
1853 | } |
1854 | |
1855 | if (JSON_HEDLEY_UNLIKELY(!unexpect_eof(input_format, "value" ))) |
1856 | { |
1857 | return false; |
1858 | } |
1859 | |
1860 | switch (current) |
1861 | { |
1862 | case 'U': |
1863 | { |
1864 | std::uint8_t len{}; |
1865 | return get_number(input_format, len) && get_string(input_format, len, result); |
1866 | } |
1867 | |
1868 | case 'i': |
1869 | { |
1870 | std::int8_t len{}; |
1871 | return get_number(input_format, len) && get_string(input_format, len, result); |
1872 | } |
1873 | |
1874 | case 'I': |
1875 | { |
1876 | std::int16_t len{}; |
1877 | return get_number(input_format, len) && get_string(input_format, len, result); |
1878 | } |
1879 | |
1880 | case 'l': |
1881 | { |
1882 | std::int32_t len{}; |
1883 | return get_number(input_format, len) && get_string(input_format, len, result); |
1884 | } |
1885 | |
1886 | case 'L': |
1887 | { |
1888 | std::int64_t len{}; |
1889 | return get_number(input_format, len) && get_string(input_format, len, result); |
1890 | } |
1891 | |
1892 | case 'u': |
1893 | { |
1894 | if (input_format != input_format_t::bjdata) |
1895 | { |
1896 | break; |
1897 | } |
1898 | std::uint16_t len{}; |
1899 | return get_number(input_format, len) && get_string(input_format, len, result); |
1900 | } |
1901 | |
1902 | case 'm': |
1903 | { |
1904 | if (input_format != input_format_t::bjdata) |
1905 | { |
1906 | break; |
1907 | } |
1908 | std::uint32_t len{}; |
1909 | return get_number(input_format, len) && get_string(input_format, len, result); |
1910 | } |
1911 | |
1912 | case 'M': |
1913 | { |
1914 | if (input_format != input_format_t::bjdata) |
1915 | { |
1916 | break; |
1917 | } |
1918 | std::uint64_t len{}; |
1919 | return get_number(input_format, len) && get_string(input_format, len, result); |
1920 | } |
1921 | |
1922 | default: |
1923 | break; |
1924 | } |
1925 | auto last_token = get_token_string(); |
1926 | std::string message; |
1927 | |
1928 | if (input_format != input_format_t::bjdata) |
1929 | { |
1930 | message = "expected length type specification (U, i, I, l, L); last byte: 0x" + last_token; |
1931 | } |
1932 | else |
1933 | { |
1934 | message = "expected length type specification (U, i, u, I, m, l, M, L); last byte: 0x" + last_token; |
1935 | } |
1936 | return sax->parse_error(chars_read, last_token, parse_error::create(113, chars_read, exception_message(format: input_format, detail: message, context: "string" ), nullptr)); |
1937 | } |
1938 | |
1939 | /*! |
1940 | @param[out] dim an integer vector storing the ND array dimensions |
1941 | @return whether reading ND array size vector is successful |
1942 | */ |
1943 | bool get_ubjson_ndarray_size(std::vector<size_t>& dim) |
1944 | { |
1945 | std::pair<std::size_t, char_int_type> size_and_type; |
1946 | size_t dimlen = 0; |
1947 | bool no_ndarray = true; |
1948 | |
1949 | if (JSON_HEDLEY_UNLIKELY(!get_ubjson_size_type(size_and_type, no_ndarray))) |
1950 | { |
1951 | return false; |
1952 | } |
1953 | |
1954 | if (size_and_type.first != npos) |
1955 | { |
1956 | if (size_and_type.second != 0) |
1957 | { |
1958 | if (size_and_type.second != 'N') |
1959 | { |
1960 | for (std::size_t i = 0; i < size_and_type.first; ++i) |
1961 | { |
1962 | if (JSON_HEDLEY_UNLIKELY(!get_ubjson_size_value(dimlen, no_ndarray, size_and_type.second))) |
1963 | { |
1964 | return false; |
1965 | } |
1966 | dim.push_back(x: dimlen); |
1967 | } |
1968 | } |
1969 | } |
1970 | else |
1971 | { |
1972 | for (std::size_t i = 0; i < size_and_type.first; ++i) |
1973 | { |
1974 | if (JSON_HEDLEY_UNLIKELY(!get_ubjson_size_value(dimlen, no_ndarray))) |
1975 | { |
1976 | return false; |
1977 | } |
1978 | dim.push_back(x: dimlen); |
1979 | } |
1980 | } |
1981 | } |
1982 | else |
1983 | { |
1984 | while (current != ']') |
1985 | { |
1986 | if (JSON_HEDLEY_UNLIKELY(!get_ubjson_size_value(dimlen, no_ndarray, current))) |
1987 | { |
1988 | return false; |
1989 | } |
1990 | dim.push_back(x: dimlen); |
1991 | get_ignore_noop(); |
1992 | } |
1993 | } |
1994 | return true; |
1995 | } |
1996 | |
1997 | /*! |
1998 | @param[out] result determined size |
1999 | @param[in,out] is_ndarray for input, `true` means already inside an ndarray vector |
2000 | or ndarray dimension is not allowed; `false` means ndarray |
2001 | is allowed; for output, `true` means an ndarray is found; |
2002 | is_ndarray can only return `true` when its initial value |
2003 | is `false` |
2004 | @param[in] prefix type marker if already read, otherwise set to 0 |
2005 | |
2006 | @return whether size determination completed |
2007 | */ |
2008 | bool get_ubjson_size_value(std::size_t& result, bool& is_ndarray, char_int_type prefix = 0) |
2009 | { |
2010 | if (prefix == 0) |
2011 | { |
2012 | prefix = get_ignore_noop(); |
2013 | } |
2014 | |
2015 | switch (prefix) |
2016 | { |
2017 | case 'U': |
2018 | { |
2019 | std::uint8_t number{}; |
2020 | if (JSON_HEDLEY_UNLIKELY(!get_number(input_format, number))) |
2021 | { |
2022 | return false; |
2023 | } |
2024 | result = static_cast<std::size_t>(number); |
2025 | return true; |
2026 | } |
2027 | |
2028 | case 'i': |
2029 | { |
2030 | std::int8_t number{}; |
2031 | if (JSON_HEDLEY_UNLIKELY(!get_number(input_format, number))) |
2032 | { |
2033 | return false; |
2034 | } |
2035 | if (number < 0) |
2036 | { |
2037 | return sax->parse_error(chars_read, get_token_string(), parse_error::create(113, chars_read, |
2038 | exception_message(format: input_format, detail: "count in an optimized container must be positive" , context: "size" ), nullptr)); |
2039 | } |
2040 | result = static_cast<std::size_t>(number); // NOLINT(bugprone-signed-char-misuse,cert-str34-c): number is not a char |
2041 | return true; |
2042 | } |
2043 | |
2044 | case 'I': |
2045 | { |
2046 | std::int16_t number{}; |
2047 | if (JSON_HEDLEY_UNLIKELY(!get_number(input_format, number))) |
2048 | { |
2049 | return false; |
2050 | } |
2051 | if (number < 0) |
2052 | { |
2053 | return sax->parse_error(chars_read, get_token_string(), parse_error::create(113, chars_read, |
2054 | exception_message(format: input_format, detail: "count in an optimized container must be positive" , context: "size" ), nullptr)); |
2055 | } |
2056 | result = static_cast<std::size_t>(number); |
2057 | return true; |
2058 | } |
2059 | |
2060 | case 'l': |
2061 | { |
2062 | std::int32_t number{}; |
2063 | if (JSON_HEDLEY_UNLIKELY(!get_number(input_format, number))) |
2064 | { |
2065 | return false; |
2066 | } |
2067 | if (number < 0) |
2068 | { |
2069 | return sax->parse_error(chars_read, get_token_string(), parse_error::create(113, chars_read, |
2070 | exception_message(format: input_format, detail: "count in an optimized container must be positive" , context: "size" ), nullptr)); |
2071 | } |
2072 | result = static_cast<std::size_t>(number); |
2073 | return true; |
2074 | } |
2075 | |
2076 | case 'L': |
2077 | { |
2078 | std::int64_t number{}; |
2079 | if (JSON_HEDLEY_UNLIKELY(!get_number(input_format, number))) |
2080 | { |
2081 | return false; |
2082 | } |
2083 | if (number < 0) |
2084 | { |
2085 | return sax->parse_error(chars_read, get_token_string(), parse_error::create(113, chars_read, |
2086 | exception_message(format: input_format, detail: "count in an optimized container must be positive" , context: "size" ), nullptr)); |
2087 | } |
2088 | if (!value_in_range_of<std::size_t>(val: number)) |
2089 | { |
2090 | return sax->parse_error(chars_read, get_token_string(), out_of_range::create(408, |
2091 | exception_message(format: input_format, detail: "integer value overflow" , context: "size" ), nullptr)); |
2092 | } |
2093 | result = static_cast<std::size_t>(number); |
2094 | return true; |
2095 | } |
2096 | |
2097 | case 'u': |
2098 | { |
2099 | if (input_format != input_format_t::bjdata) |
2100 | { |
2101 | break; |
2102 | } |
2103 | std::uint16_t number{}; |
2104 | if (JSON_HEDLEY_UNLIKELY(!get_number(input_format, number))) |
2105 | { |
2106 | return false; |
2107 | } |
2108 | result = static_cast<std::size_t>(number); |
2109 | return true; |
2110 | } |
2111 | |
2112 | case 'm': |
2113 | { |
2114 | if (input_format != input_format_t::bjdata) |
2115 | { |
2116 | break; |
2117 | } |
2118 | std::uint32_t number{}; |
2119 | if (JSON_HEDLEY_UNLIKELY(!get_number(input_format, number))) |
2120 | { |
2121 | return false; |
2122 | } |
2123 | result = conditional_static_cast<std::size_t>(value: number); |
2124 | return true; |
2125 | } |
2126 | |
2127 | case 'M': |
2128 | { |
2129 | if (input_format != input_format_t::bjdata) |
2130 | { |
2131 | break; |
2132 | } |
2133 | std::uint64_t number{}; |
2134 | if (JSON_HEDLEY_UNLIKELY(!get_number(input_format, number))) |
2135 | { |
2136 | return false; |
2137 | } |
2138 | if (!value_in_range_of<std::size_t>(val: number)) |
2139 | { |
2140 | return sax->parse_error(chars_read, get_token_string(), out_of_range::create(408, |
2141 | exception_message(format: input_format, detail: "integer value overflow" , context: "size" ), nullptr)); |
2142 | } |
2143 | result = detail::conditional_static_cast<std::size_t>(value: number); |
2144 | return true; |
2145 | } |
2146 | |
2147 | case '[': |
2148 | { |
2149 | if (input_format != input_format_t::bjdata) |
2150 | { |
2151 | break; |
2152 | } |
2153 | if (is_ndarray) // ndarray dimensional vector can only contain integers, and can not embed another array |
2154 | { |
2155 | return sax->parse_error(chars_read, get_token_string(), parse_error::create(113, chars_read, exception_message(format: input_format, detail: "ndarray dimensional vector is not allowed" , context: "size" ), nullptr)); |
2156 | } |
2157 | std::vector<size_t> dim; |
2158 | if (JSON_HEDLEY_UNLIKELY(!get_ubjson_ndarray_size(dim))) |
2159 | { |
2160 | return false; |
2161 | } |
2162 | if (dim.size() == 1 || (dim.size() == 2 && dim.at(n: 0) == 1)) // return normal array size if 1D row vector |
2163 | { |
2164 | result = dim.at(n: dim.size() - 1); |
2165 | return true; |
2166 | } |
2167 | if (!dim.empty()) // if ndarray, convert to an object in JData annotated array format |
2168 | { |
2169 | for (auto i : dim) // test if any dimension in an ndarray is 0, if so, return a 1D empty container |
2170 | { |
2171 | if ( i == 0 ) |
2172 | { |
2173 | result = 0; |
2174 | return true; |
2175 | } |
2176 | } |
2177 | |
2178 | string_t key = "_ArraySize_" ; |
2179 | if (JSON_HEDLEY_UNLIKELY(!sax->start_object(3) || !sax->key(key) || !sax->start_array(dim.size()))) |
2180 | { |
2181 | return false; |
2182 | } |
2183 | result = 1; |
2184 | for (auto i : dim) |
2185 | { |
2186 | result *= i; |
2187 | if (result == 0 || result == npos) // because dim elements shall not have zeros, result = 0 means overflow happened; it also can't be npos as it is used to initialize size in get_ubjson_size_type() |
2188 | { |
2189 | return sax->parse_error(chars_read, get_token_string(), out_of_range::create(408, exception_message(format: input_format, detail: "excessive ndarray size caused overflow" , context: "size" ), nullptr)); |
2190 | } |
2191 | if (JSON_HEDLEY_UNLIKELY(!sax->number_unsigned(static_cast<number_unsigned_t>(i)))) |
2192 | { |
2193 | return false; |
2194 | } |
2195 | } |
2196 | is_ndarray = true; |
2197 | return sax->end_array(); |
2198 | } |
2199 | result = 0; |
2200 | return true; |
2201 | } |
2202 | |
2203 | default: |
2204 | break; |
2205 | } |
2206 | auto last_token = get_token_string(); |
2207 | std::string message; |
2208 | |
2209 | if (input_format != input_format_t::bjdata) |
2210 | { |
2211 | message = "expected length type specification (U, i, I, l, L) after '#'; last byte: 0x" + last_token; |
2212 | } |
2213 | else |
2214 | { |
2215 | message = "expected length type specification (U, i, u, I, m, l, M, L) after '#'; last byte: 0x" + last_token; |
2216 | } |
2217 | return sax->parse_error(chars_read, last_token, parse_error::create(113, chars_read, exception_message(format: input_format, detail: message, context: "size" ), nullptr)); |
2218 | } |
2219 | |
2220 | /*! |
2221 | @brief determine the type and size for a container |
2222 | |
2223 | In the optimized UBJSON format, a type and a size can be provided to allow |
2224 | for a more compact representation. |
2225 | |
2226 | @param[out] result pair of the size and the type |
2227 | @param[in] inside_ndarray whether the parser is parsing an ND array dimensional vector |
2228 | |
2229 | @return whether pair creation completed |
2230 | */ |
2231 | bool get_ubjson_size_type(std::pair<std::size_t, char_int_type>& result, bool inside_ndarray = false) |
2232 | { |
2233 | result.first = npos; // size |
2234 | result.second = 0; // type |
2235 | bool is_ndarray = false; |
2236 | |
2237 | get_ignore_noop(); |
2238 | |
2239 | if (current == '$') |
2240 | { |
2241 | result.second = get(); // must not ignore 'N', because 'N' maybe the type |
2242 | if (input_format == input_format_t::bjdata |
2243 | && JSON_HEDLEY_UNLIKELY(std::binary_search(bjd_optimized_type_markers.begin(), bjd_optimized_type_markers.end(), result.second))) |
2244 | { |
2245 | auto last_token = get_token_string(); |
2246 | return sax->parse_error(chars_read, last_token, parse_error::create(112, chars_read, |
2247 | exception_message(format: input_format, detail: concat("marker 0x" , last_token, " is not a permitted optimized array type" ), context: "type" ), nullptr)); |
2248 | } |
2249 | |
2250 | if (JSON_HEDLEY_UNLIKELY(!unexpect_eof(input_format, "type" ))) |
2251 | { |
2252 | return false; |
2253 | } |
2254 | |
2255 | get_ignore_noop(); |
2256 | if (JSON_HEDLEY_UNLIKELY(current != '#')) |
2257 | { |
2258 | if (JSON_HEDLEY_UNLIKELY(!unexpect_eof(input_format, "value" ))) |
2259 | { |
2260 | return false; |
2261 | } |
2262 | auto last_token = get_token_string(); |
2263 | return sax->parse_error(chars_read, last_token, parse_error::create(112, chars_read, |
2264 | exception_message(format: input_format, detail: concat("expected '#' after type information; last byte: 0x" , last_token), context: "size" ), nullptr)); |
2265 | } |
2266 | |
2267 | const bool is_error = get_ubjson_size_value(result&: result.first, is_ndarray); |
2268 | if (input_format == input_format_t::bjdata && is_ndarray) |
2269 | { |
2270 | if (inside_ndarray) |
2271 | { |
2272 | return sax->parse_error(chars_read, get_token_string(), parse_error::create(112, chars_read, |
2273 | exception_message(format: input_format, detail: "ndarray can not be recursive" , context: "size" ), nullptr)); |
2274 | } |
2275 | result.second |= (1 << 8); // use bit 8 to indicate ndarray, all UBJSON and BJData markers should be ASCII letters |
2276 | } |
2277 | return is_error; |
2278 | } |
2279 | |
2280 | if (current == '#') |
2281 | { |
2282 | const bool is_error = get_ubjson_size_value(result&: result.first, is_ndarray); |
2283 | if (input_format == input_format_t::bjdata && is_ndarray) |
2284 | { |
2285 | return sax->parse_error(chars_read, get_token_string(), parse_error::create(112, chars_read, |
2286 | exception_message(format: input_format, detail: "ndarray requires both type and size" , context: "size" ), nullptr)); |
2287 | } |
2288 | return is_error; |
2289 | } |
2290 | |
2291 | return true; |
2292 | } |
2293 | |
2294 | /*! |
2295 | @param prefix the previously read or set type prefix |
2296 | @return whether value creation completed |
2297 | */ |
2298 | bool get_ubjson_value(const char_int_type prefix) |
2299 | { |
2300 | switch (prefix) |
2301 | { |
2302 | case char_traits<char_type>::eof(): // EOF |
2303 | return unexpect_eof(format: input_format, context: "value" ); |
2304 | |
2305 | case 'T': // true |
2306 | return sax->boolean(true); |
2307 | case 'F': // false |
2308 | return sax->boolean(false); |
2309 | |
2310 | case 'Z': // null |
2311 | return sax->null(); |
2312 | |
2313 | case 'U': |
2314 | { |
2315 | std::uint8_t number{}; |
2316 | return get_number(input_format, number) && sax->number_unsigned(number); |
2317 | } |
2318 | |
2319 | case 'i': |
2320 | { |
2321 | std::int8_t number{}; |
2322 | return get_number(input_format, number) && sax->number_integer(number); |
2323 | } |
2324 | |
2325 | case 'I': |
2326 | { |
2327 | std::int16_t number{}; |
2328 | return get_number(input_format, number) && sax->number_integer(number); |
2329 | } |
2330 | |
2331 | case 'l': |
2332 | { |
2333 | std::int32_t number{}; |
2334 | return get_number(input_format, number) && sax->number_integer(number); |
2335 | } |
2336 | |
2337 | case 'L': |
2338 | { |
2339 | std::int64_t number{}; |
2340 | return get_number(input_format, number) && sax->number_integer(number); |
2341 | } |
2342 | |
2343 | case 'u': |
2344 | { |
2345 | if (input_format != input_format_t::bjdata) |
2346 | { |
2347 | break; |
2348 | } |
2349 | std::uint16_t number{}; |
2350 | return get_number(input_format, number) && sax->number_unsigned(number); |
2351 | } |
2352 | |
2353 | case 'm': |
2354 | { |
2355 | if (input_format != input_format_t::bjdata) |
2356 | { |
2357 | break; |
2358 | } |
2359 | std::uint32_t number{}; |
2360 | return get_number(input_format, number) && sax->number_unsigned(number); |
2361 | } |
2362 | |
2363 | case 'M': |
2364 | { |
2365 | if (input_format != input_format_t::bjdata) |
2366 | { |
2367 | break; |
2368 | } |
2369 | std::uint64_t number{}; |
2370 | return get_number(input_format, number) && sax->number_unsigned(number); |
2371 | } |
2372 | |
2373 | case 'h': |
2374 | { |
2375 | if (input_format != input_format_t::bjdata) |
2376 | { |
2377 | break; |
2378 | } |
2379 | const auto byte1_raw = get(); |
2380 | if (JSON_HEDLEY_UNLIKELY(!unexpect_eof(input_format, "number" ))) |
2381 | { |
2382 | return false; |
2383 | } |
2384 | const auto byte2_raw = get(); |
2385 | if (JSON_HEDLEY_UNLIKELY(!unexpect_eof(input_format, "number" ))) |
2386 | { |
2387 | return false; |
2388 | } |
2389 | |
2390 | const auto byte1 = static_cast<unsigned char>(byte1_raw); |
2391 | const auto byte2 = static_cast<unsigned char>(byte2_raw); |
2392 | |
2393 | // code from RFC 7049, Appendix D, Figure 3: |
2394 | // As half-precision floating-point numbers were only added |
2395 | // to IEEE 754 in 2008, today's programming platforms often |
2396 | // still only have limited support for them. It is very |
2397 | // easy to include at least decoding support for them even |
2398 | // without such support. An example of a small decoder for |
2399 | // half-precision floating-point numbers in the C language |
2400 | // is shown in Fig. 3. |
2401 | const auto half = static_cast<unsigned int>((byte2 << 8u) + byte1); |
2402 | const double val = [&half] |
2403 | { |
2404 | const int exp = (half >> 10u) & 0x1Fu; |
2405 | const unsigned int mant = half & 0x3FFu; |
2406 | JSON_ASSERT(0 <= exp&& exp <= 32); |
2407 | JSON_ASSERT(mant <= 1024); |
2408 | switch (exp) |
2409 | { |
2410 | case 0: |
2411 | return std::ldexp(x: mant, exp: -24); |
2412 | case 31: |
2413 | return (mant == 0) |
2414 | ? std::numeric_limits<double>::infinity() |
2415 | : std::numeric_limits<double>::quiet_NaN(); |
2416 | default: |
2417 | return std::ldexp(x: mant + 1024, exp: exp - 25); |
2418 | } |
2419 | }(); |
2420 | return sax->number_float((half & 0x8000u) != 0 |
2421 | ? static_cast<number_float_t>(-val) |
2422 | : static_cast<number_float_t>(val), "" ); |
2423 | } |
2424 | |
2425 | case 'd': |
2426 | { |
2427 | float number{}; |
2428 | return get_number(input_format, number) && sax->number_float(static_cast<number_float_t>(number), "" ); |
2429 | } |
2430 | |
2431 | case 'D': |
2432 | { |
2433 | double number{}; |
2434 | return get_number(input_format, number) && sax->number_float(static_cast<number_float_t>(number), "" ); |
2435 | } |
2436 | |
2437 | case 'H': |
2438 | { |
2439 | return get_ubjson_high_precision_number(); |
2440 | } |
2441 | |
2442 | case 'C': // char |
2443 | { |
2444 | get(); |
2445 | if (JSON_HEDLEY_UNLIKELY(!unexpect_eof(input_format, "char" ))) |
2446 | { |
2447 | return false; |
2448 | } |
2449 | if (JSON_HEDLEY_UNLIKELY(current > 127)) |
2450 | { |
2451 | auto last_token = get_token_string(); |
2452 | return sax->parse_error(chars_read, last_token, parse_error::create(113, chars_read, |
2453 | exception_message(format: input_format, detail: concat("byte after 'C' must be in range 0x00..0x7F; last byte: 0x" , last_token), context: "char" ), nullptr)); |
2454 | } |
2455 | string_t s(1, static_cast<typename string_t::value_type>(current)); |
2456 | return sax->string(s); |
2457 | } |
2458 | |
2459 | case 'S': // string |
2460 | { |
2461 | string_t s; |
2462 | return get_ubjson_string(result&: s) && sax->string(s); |
2463 | } |
2464 | |
2465 | case '[': // array |
2466 | return get_ubjson_array(); |
2467 | |
2468 | case '{': // object |
2469 | return get_ubjson_object(); |
2470 | |
2471 | default: // anything else |
2472 | break; |
2473 | } |
2474 | auto last_token = get_token_string(); |
2475 | return sax->parse_error(chars_read, last_token, parse_error::create(112, chars_read, exception_message(format: input_format, detail: "invalid byte: 0x" + last_token, context: "value" ), nullptr)); |
2476 | } |
2477 | |
2478 | /*! |
2479 | @return whether array creation completed |
2480 | */ |
2481 | bool get_ubjson_array() |
2482 | { |
2483 | std::pair<std::size_t, char_int_type> size_and_type; |
2484 | if (JSON_HEDLEY_UNLIKELY(!get_ubjson_size_type(size_and_type))) |
2485 | { |
2486 | return false; |
2487 | } |
2488 | |
2489 | // if bit-8 of size_and_type.second is set to 1, encode bjdata ndarray as an object in JData annotated array format (https://github.com/NeuroJSON/jdata): |
2490 | // {"_ArrayType_" : "typeid", "_ArraySize_" : [n1, n2, ...], "_ArrayData_" : [v1, v2, ...]} |
2491 | |
2492 | if (input_format == input_format_t::bjdata && size_and_type.first != npos && (size_and_type.second & (1 << 8)) != 0) |
2493 | { |
2494 | size_and_type.second &= ~(static_cast<char_int_type>(1) << 8); // use bit 8 to indicate ndarray, here we remove the bit to restore the type marker |
2495 | auto it = std::lower_bound(bjd_types_map.begin(), bjd_types_map.end(), size_and_type.second, [](const bjd_type & p, char_int_type t) |
2496 | { |
2497 | return p.first < t; |
2498 | }); |
2499 | string_t key = "_ArrayType_" ; |
2500 | if (JSON_HEDLEY_UNLIKELY(it == bjd_types_map.end() || it->first != size_and_type.second)) |
2501 | { |
2502 | auto last_token = get_token_string(); |
2503 | return sax->parse_error(chars_read, last_token, parse_error::create(112, chars_read, |
2504 | exception_message(format: input_format, detail: "invalid byte: 0x" + last_token, context: "type" ), nullptr)); |
2505 | } |
2506 | |
2507 | string_t type = it->second; // sax->string() takes a reference |
2508 | if (JSON_HEDLEY_UNLIKELY(!sax->key(key) || !sax->string(type))) |
2509 | { |
2510 | return false; |
2511 | } |
2512 | |
2513 | if (size_and_type.second == 'C') |
2514 | { |
2515 | size_and_type.second = 'U'; |
2516 | } |
2517 | |
2518 | key = "_ArrayData_" ; |
2519 | if (JSON_HEDLEY_UNLIKELY(!sax->key(key) || !sax->start_array(size_and_type.first) )) |
2520 | { |
2521 | return false; |
2522 | } |
2523 | |
2524 | for (std::size_t i = 0; i < size_and_type.first; ++i) |
2525 | { |
2526 | if (JSON_HEDLEY_UNLIKELY(!get_ubjson_value(size_and_type.second))) |
2527 | { |
2528 | return false; |
2529 | } |
2530 | } |
2531 | |
2532 | return (sax->end_array() && sax->end_object()); |
2533 | } |
2534 | |
2535 | if (size_and_type.first != npos) |
2536 | { |
2537 | if (JSON_HEDLEY_UNLIKELY(!sax->start_array(size_and_type.first))) |
2538 | { |
2539 | return false; |
2540 | } |
2541 | |
2542 | if (size_and_type.second != 0) |
2543 | { |
2544 | if (size_and_type.second != 'N') |
2545 | { |
2546 | for (std::size_t i = 0; i < size_and_type.first; ++i) |
2547 | { |
2548 | if (JSON_HEDLEY_UNLIKELY(!get_ubjson_value(size_and_type.second))) |
2549 | { |
2550 | return false; |
2551 | } |
2552 | } |
2553 | } |
2554 | } |
2555 | else |
2556 | { |
2557 | for (std::size_t i = 0; i < size_and_type.first; ++i) |
2558 | { |
2559 | if (JSON_HEDLEY_UNLIKELY(!parse_ubjson_internal())) |
2560 | { |
2561 | return false; |
2562 | } |
2563 | } |
2564 | } |
2565 | } |
2566 | else |
2567 | { |
2568 | if (JSON_HEDLEY_UNLIKELY(!sax->start_array(static_cast<std::size_t>(-1)))) |
2569 | { |
2570 | return false; |
2571 | } |
2572 | |
2573 | while (current != ']') |
2574 | { |
2575 | if (JSON_HEDLEY_UNLIKELY(!parse_ubjson_internal(false))) |
2576 | { |
2577 | return false; |
2578 | } |
2579 | get_ignore_noop(); |
2580 | } |
2581 | } |
2582 | |
2583 | return sax->end_array(); |
2584 | } |
2585 | |
2586 | /*! |
2587 | @return whether object creation completed |
2588 | */ |
2589 | bool get_ubjson_object() |
2590 | { |
2591 | std::pair<std::size_t, char_int_type> size_and_type; |
2592 | if (JSON_HEDLEY_UNLIKELY(!get_ubjson_size_type(size_and_type))) |
2593 | { |
2594 | return false; |
2595 | } |
2596 | |
2597 | // do not accept ND-array size in objects in BJData |
2598 | if (input_format == input_format_t::bjdata && size_and_type.first != npos && (size_and_type.second & (1 << 8)) != 0) |
2599 | { |
2600 | auto last_token = get_token_string(); |
2601 | return sax->parse_error(chars_read, last_token, parse_error::create(112, chars_read, |
2602 | exception_message(format: input_format, detail: "BJData object does not support ND-array size in optimized format" , context: "object" ), nullptr)); |
2603 | } |
2604 | |
2605 | string_t key; |
2606 | if (size_and_type.first != npos) |
2607 | { |
2608 | if (JSON_HEDLEY_UNLIKELY(!sax->start_object(size_and_type.first))) |
2609 | { |
2610 | return false; |
2611 | } |
2612 | |
2613 | if (size_and_type.second != 0) |
2614 | { |
2615 | for (std::size_t i = 0; i < size_and_type.first; ++i) |
2616 | { |
2617 | if (JSON_HEDLEY_UNLIKELY(!get_ubjson_string(key) || !sax->key(key))) |
2618 | { |
2619 | return false; |
2620 | } |
2621 | if (JSON_HEDLEY_UNLIKELY(!get_ubjson_value(size_and_type.second))) |
2622 | { |
2623 | return false; |
2624 | } |
2625 | key.clear(); |
2626 | } |
2627 | } |
2628 | else |
2629 | { |
2630 | for (std::size_t i = 0; i < size_and_type.first; ++i) |
2631 | { |
2632 | if (JSON_HEDLEY_UNLIKELY(!get_ubjson_string(key) || !sax->key(key))) |
2633 | { |
2634 | return false; |
2635 | } |
2636 | if (JSON_HEDLEY_UNLIKELY(!parse_ubjson_internal())) |
2637 | { |
2638 | return false; |
2639 | } |
2640 | key.clear(); |
2641 | } |
2642 | } |
2643 | } |
2644 | else |
2645 | { |
2646 | if (JSON_HEDLEY_UNLIKELY(!sax->start_object(static_cast<std::size_t>(-1)))) |
2647 | { |
2648 | return false; |
2649 | } |
2650 | |
2651 | while (current != '}') |
2652 | { |
2653 | if (JSON_HEDLEY_UNLIKELY(!get_ubjson_string(key, false) || !sax->key(key))) |
2654 | { |
2655 | return false; |
2656 | } |
2657 | if (JSON_HEDLEY_UNLIKELY(!parse_ubjson_internal())) |
2658 | { |
2659 | return false; |
2660 | } |
2661 | get_ignore_noop(); |
2662 | key.clear(); |
2663 | } |
2664 | } |
2665 | |
2666 | return sax->end_object(); |
2667 | } |
2668 | |
2669 | // Note, no reader for UBJSON binary types is implemented because they do |
2670 | // not exist |
2671 | |
2672 | bool get_ubjson_high_precision_number() |
2673 | { |
2674 | // get size of following number string |
2675 | std::size_t size{}; |
2676 | bool no_ndarray = true; |
2677 | auto res = get_ubjson_size_value(result&: size, is_ndarray&: no_ndarray); |
2678 | if (JSON_HEDLEY_UNLIKELY(!res)) |
2679 | { |
2680 | return res; |
2681 | } |
2682 | |
2683 | // get number string |
2684 | std::vector<char> number_vector; |
2685 | for (std::size_t i = 0; i < size; ++i) |
2686 | { |
2687 | get(); |
2688 | if (JSON_HEDLEY_UNLIKELY(!unexpect_eof(input_format, "number" ))) |
2689 | { |
2690 | return false; |
2691 | } |
2692 | number_vector.push_back(x: static_cast<char>(current)); |
2693 | } |
2694 | |
2695 | // parse number string |
2696 | using ia_type = decltype(detail::input_adapter(container: number_vector)); |
2697 | auto number_lexer = detail::lexer<BasicJsonType, ia_type>(detail::input_adapter(container: number_vector), false); |
2698 | const auto result_number = number_lexer.scan(); |
2699 | const auto number_string = number_lexer.get_token_string(); |
2700 | const auto result_remainder = number_lexer.scan(); |
2701 | |
2702 | using token_type = typename detail::lexer_base<BasicJsonType>::token_type; |
2703 | |
2704 | if (JSON_HEDLEY_UNLIKELY(result_remainder != token_type::end_of_input)) |
2705 | { |
2706 | return sax->parse_error(chars_read, number_string, parse_error::create(115, chars_read, |
2707 | exception_message(format: input_format, detail: concat("invalid number text: " , number_lexer.get_token_string()), context: "high-precision number" ), nullptr)); |
2708 | } |
2709 | |
2710 | switch (result_number) |
2711 | { |
2712 | case token_type::value_integer: |
2713 | return sax->number_integer(number_lexer.get_number_integer()); |
2714 | case token_type::value_unsigned: |
2715 | return sax->number_unsigned(number_lexer.get_number_unsigned()); |
2716 | case token_type::value_float: |
2717 | return sax->number_float(number_lexer.get_number_float(), std::move(number_string)); |
2718 | case token_type::uninitialized: |
2719 | case token_type::literal_true: |
2720 | case token_type::literal_false: |
2721 | case token_type::literal_null: |
2722 | case token_type::value_string: |
2723 | case token_type::begin_array: |
2724 | case token_type::begin_object: |
2725 | case token_type::end_array: |
2726 | case token_type::end_object: |
2727 | case token_type::name_separator: |
2728 | case token_type::value_separator: |
2729 | case token_type::parse_error: |
2730 | case token_type::end_of_input: |
2731 | case token_type::literal_or_value: |
2732 | default: |
2733 | return sax->parse_error(chars_read, number_string, parse_error::create(115, chars_read, |
2734 | exception_message(format: input_format, detail: concat("invalid number text: " , number_lexer.get_token_string()), context: "high-precision number" ), nullptr)); |
2735 | } |
2736 | } |
2737 | |
2738 | /////////////////////// |
2739 | // Utility functions // |
2740 | /////////////////////// |
2741 | |
2742 | /*! |
2743 | @brief get next character from the input |
2744 | |
2745 | This function provides the interface to the used input adapter. It does |
2746 | not throw in case the input reached EOF, but returns a -'ve valued |
2747 | `char_traits<char_type>::eof()` in that case. |
2748 | |
2749 | @return character read from the input |
2750 | */ |
2751 | char_int_type get() |
2752 | { |
2753 | ++chars_read; |
2754 | return current = ia.get_character(); |
2755 | } |
2756 | |
2757 | /*! |
2758 | @return character read from the input after ignoring all 'N' entries |
2759 | */ |
2760 | char_int_type get_ignore_noop() |
2761 | { |
2762 | do |
2763 | { |
2764 | get(); |
2765 | } |
2766 | while (current == 'N'); |
2767 | |
2768 | return current; |
2769 | } |
2770 | |
2771 | /* |
2772 | @brief read a number from the input |
2773 | |
2774 | @tparam NumberType the type of the number |
2775 | @param[in] format the current format (for diagnostics) |
2776 | @param[out] result number of type @a NumberType |
2777 | |
2778 | @return whether conversion completed |
2779 | |
2780 | @note This function needs to respect the system's endianness, because |
2781 | bytes in CBOR, MessagePack, and UBJSON are stored in network order |
2782 | (big endian) and therefore need reordering on little endian systems. |
2783 | On the other hand, BSON and BJData use little endian and should reorder |
2784 | on big endian systems. |
2785 | */ |
2786 | template<typename NumberType, bool InputIsLittleEndian = false> |
2787 | bool get_number(const input_format_t format, NumberType& result) |
2788 | { |
2789 | // step 1: read input into array with system's byte order |
2790 | std::array<std::uint8_t, sizeof(NumberType)> vec{}; |
2791 | for (std::size_t i = 0; i < sizeof(NumberType); ++i) |
2792 | { |
2793 | get(); |
2794 | if (JSON_HEDLEY_UNLIKELY(!unexpect_eof(format, "number" ))) |
2795 | { |
2796 | return false; |
2797 | } |
2798 | |
2799 | // reverse byte order prior to conversion if necessary |
2800 | if (is_little_endian != (InputIsLittleEndian || format == input_format_t::bjdata)) |
2801 | { |
2802 | vec[sizeof(NumberType) - i - 1] = static_cast<std::uint8_t>(current); |
2803 | } |
2804 | else |
2805 | { |
2806 | vec[i] = static_cast<std::uint8_t>(current); // LCOV_EXCL_LINE |
2807 | } |
2808 | } |
2809 | |
2810 | // step 2: convert array into number of type T and return |
2811 | std::memcpy(dest: &result, src: vec.data(), n: sizeof(NumberType)); |
2812 | return true; |
2813 | } |
2814 | |
2815 | /*! |
2816 | @brief create a string by reading characters from the input |
2817 | |
2818 | @tparam NumberType the type of the number |
2819 | @param[in] format the current format (for diagnostics) |
2820 | @param[in] len number of characters to read |
2821 | @param[out] result string created by reading @a len bytes |
2822 | |
2823 | @return whether string creation completed |
2824 | |
2825 | @note We can not reserve @a len bytes for the result, because @a len |
2826 | may be too large. Usually, @ref unexpect_eof() detects the end of |
2827 | the input before we run out of string memory. |
2828 | */ |
2829 | template<typename NumberType> |
2830 | bool get_string(const input_format_t format, |
2831 | const NumberType len, |
2832 | string_t& result) |
2833 | { |
2834 | bool success = true; |
2835 | for (NumberType i = 0; i < len; i++) |
2836 | { |
2837 | get(); |
2838 | if (JSON_HEDLEY_UNLIKELY(!unexpect_eof(format, "string" ))) |
2839 | { |
2840 | success = false; |
2841 | break; |
2842 | } |
2843 | result.push_back(static_cast<typename string_t::value_type>(current)); |
2844 | } |
2845 | return success; |
2846 | } |
2847 | |
2848 | /*! |
2849 | @brief create a byte array by reading bytes from the input |
2850 | |
2851 | @tparam NumberType the type of the number |
2852 | @param[in] format the current format (for diagnostics) |
2853 | @param[in] len number of bytes to read |
2854 | @param[out] result byte array created by reading @a len bytes |
2855 | |
2856 | @return whether byte array creation completed |
2857 | |
2858 | @note We can not reserve @a len bytes for the result, because @a len |
2859 | may be too large. Usually, @ref unexpect_eof() detects the end of |
2860 | the input before we run out of memory. |
2861 | */ |
2862 | template<typename NumberType> |
2863 | bool get_binary(const input_format_t format, |
2864 | const NumberType len, |
2865 | binary_t& result) |
2866 | { |
2867 | bool success = true; |
2868 | for (NumberType i = 0; i < len; i++) |
2869 | { |
2870 | get(); |
2871 | if (JSON_HEDLEY_UNLIKELY(!unexpect_eof(format, "binary" ))) |
2872 | { |
2873 | success = false; |
2874 | break; |
2875 | } |
2876 | result.push_back(static_cast<std::uint8_t>(current)); |
2877 | } |
2878 | return success; |
2879 | } |
2880 | |
2881 | /*! |
2882 | @param[in] format the current format (for diagnostics) |
2883 | @param[in] context further context information (for diagnostics) |
2884 | @return whether the last read character is not EOF |
2885 | */ |
2886 | JSON_HEDLEY_NON_NULL(3) |
2887 | bool unexpect_eof(const input_format_t format, const char* context) const |
2888 | { |
2889 | if (JSON_HEDLEY_UNLIKELY(current == char_traits<char_type>::eof())) |
2890 | { |
2891 | return sax->parse_error(chars_read, "<end of file>" , |
2892 | parse_error::create(110, chars_read, exception_message(format, detail: "unexpected end of input" , context), nullptr)); |
2893 | } |
2894 | return true; |
2895 | } |
2896 | |
2897 | /*! |
2898 | @return a string representation of the last read byte |
2899 | */ |
2900 | std::string get_token_string() const |
2901 | { |
2902 | std::array<char, 3> cr{._M_elems: {}}; |
2903 | static_cast<void>((std::snprintf)(s: cr.data(), maxlen: cr.size(), format: "%.2hhX" , static_cast<unsigned char>(current))); // NOLINT(cppcoreguidelines-pro-type-vararg,hicpp-vararg) |
2904 | return std::string{cr.data()}; |
2905 | } |
2906 | |
2907 | /*! |
2908 | @param[in] format the current format |
2909 | @param[in] detail a detailed error message |
2910 | @param[in] context further context information |
2911 | @return a message string to use in the parse_error exceptions |
2912 | */ |
2913 | std::string exception_message(const input_format_t format, |
2914 | const std::string& detail, |
2915 | const std::string& context) const |
2916 | { |
2917 | std::string error_msg = "syntax error while parsing " ; |
2918 | |
2919 | switch (format) |
2920 | { |
2921 | case input_format_t::cbor: |
2922 | error_msg += "CBOR" ; |
2923 | break; |
2924 | |
2925 | case input_format_t::msgpack: |
2926 | error_msg += "MessagePack" ; |
2927 | break; |
2928 | |
2929 | case input_format_t::ubjson: |
2930 | error_msg += "UBJSON" ; |
2931 | break; |
2932 | |
2933 | case input_format_t::bson: |
2934 | error_msg += "BSON" ; |
2935 | break; |
2936 | |
2937 | case input_format_t::bjdata: |
2938 | error_msg += "BJData" ; |
2939 | break; |
2940 | |
2941 | case input_format_t::json: // LCOV_EXCL_LINE |
2942 | default: // LCOV_EXCL_LINE |
2943 | JSON_ASSERT(false); // NOLINT(cert-dcl03-c,hicpp-static-assert,misc-static-assert) LCOV_EXCL_LINE |
2944 | } |
2945 | |
2946 | return concat(args&: error_msg, args: ' ', args: context, args: ": " , args: detail); |
2947 | } |
2948 | |
2949 | private: |
2950 | static JSON_INLINE_VARIABLE constexpr std::size_t npos = static_cast<std::size_t>(-1); |
2951 | |
2952 | /// input adapter |
2953 | InputAdapterType ia; |
2954 | |
2955 | /// the current character |
2956 | char_int_type current = char_traits<char_type>::eof(); |
2957 | |
2958 | /// the number of characters read |
2959 | std::size_t chars_read = 0; |
2960 | |
2961 | /// whether we can assume little endianness |
2962 | const bool is_little_endian = little_endianness(); |
2963 | |
2964 | /// input format |
2965 | const input_format_t input_format = input_format_t::json; |
2966 | |
2967 | /// the SAX parser |
2968 | json_sax_t* sax = nullptr; |
2969 | |
2970 | // excluded markers in bjdata optimized type |
2971 | #define JSON_BINARY_READER_MAKE_BJD_OPTIMIZED_TYPE_MARKERS_ \ |
2972 | make_array<char_int_type>('F', 'H', 'N', 'S', 'T', 'Z', '[', '{') |
2973 | |
2974 | #define JSON_BINARY_READER_MAKE_BJD_TYPES_MAP_ \ |
2975 | make_array<bjd_type>( \ |
2976 | bjd_type{'C', "char"}, \ |
2977 | bjd_type{'D', "double"}, \ |
2978 | bjd_type{'I', "int16"}, \ |
2979 | bjd_type{'L', "int64"}, \ |
2980 | bjd_type{'M', "uint64"}, \ |
2981 | bjd_type{'U', "uint8"}, \ |
2982 | bjd_type{'d', "single"}, \ |
2983 | bjd_type{'i', "int8"}, \ |
2984 | bjd_type{'l', "int32"}, \ |
2985 | bjd_type{'m', "uint32"}, \ |
2986 | bjd_type{'u', "uint16"}) |
2987 | |
2988 | JSON_PRIVATE_UNLESS_TESTED: |
2989 | // lookup tables |
2990 | // NOLINTNEXTLINE(cppcoreguidelines-non-private-member-variables-in-classes) |
2991 | const decltype(JSON_BINARY_READER_MAKE_BJD_OPTIMIZED_TYPE_MARKERS_) bjd_optimized_type_markers = |
2992 | JSON_BINARY_READER_MAKE_BJD_OPTIMIZED_TYPE_MARKERS_; |
2993 | |
2994 | using bjd_type = std::pair<char_int_type, string_t>; |
2995 | // NOLINTNEXTLINE(cppcoreguidelines-non-private-member-variables-in-classes) |
2996 | const decltype(JSON_BINARY_READER_MAKE_BJD_TYPES_MAP_) bjd_types_map = |
2997 | JSON_BINARY_READER_MAKE_BJD_TYPES_MAP_; |
2998 | |
2999 | #undef JSON_BINARY_READER_MAKE_BJD_OPTIMIZED_TYPE_MARKERS_ |
3000 | #undef JSON_BINARY_READER_MAKE_BJD_TYPES_MAP_ |
3001 | }; |
3002 | |
3003 | #ifndef JSON_HAS_CPP_17 |
3004 | template<typename BasicJsonType, typename InputAdapterType, typename SAX> |
3005 | constexpr std::size_t binary_reader<BasicJsonType, InputAdapterType, SAX>::npos; |
3006 | #endif |
3007 | |
3008 | } // namespace detail |
3009 | NLOHMANN_JSON_NAMESPACE_END |
3010 | |