2 #include "yaml_private.h"
9 yaml_parser_set_reader_error(yaml_parser_t *parser, const char *problem,
10 size_t offset, int value);
13 yaml_parser_update_raw_buffer(yaml_parser_t *parser);
16 yaml_parser_determine_encoding(yaml_parser_t *parser);
19 yaml_parser_update_buffer(yaml_parser_t *parser, size_t length);
22 * Set the reader error and return 0.
26 yaml_parser_set_reader_error(yaml_parser_t *parser, const char *problem,
27 size_t offset, int value)
29 parser->error = YAML_READER_ERROR;
30 parser->problem = problem;
31 parser->problem_offset = offset;
32 parser->problem_value = value;
41 #define BOM_UTF8 "\xef\xbb\xbf"
42 #define BOM_UTF16LE "\xff\xfe"
43 #define BOM_UTF16BE "\xfe\xff"
46 * Determine the input stream encoding by checking the BOM symbol. If no BOM is
47 * found, the UTF-8 encoding is assumed. Return 1 on success, 0 on failure.
51 yaml_parser_determine_encoding(yaml_parser_t *parser)
53 /* Ensure that we had enough bytes in the raw buffer. */
56 && parser->raw_buffer.last - parser->raw_buffer.pointer < 3) {
57 if (!yaml_parser_update_raw_buffer(parser)) {
62 /* Determine the encoding. */
64 if (parser->raw_buffer.last - parser->raw_buffer.pointer >= 2
65 && !memcmp(parser->raw_buffer.pointer, BOM_UTF16LE, 2)) {
66 parser->encoding = YAML_UTF16LE_ENCODING;
67 parser->raw_buffer.pointer += 2;
70 else if (parser->raw_buffer.last - parser->raw_buffer.pointer >= 2
71 && !memcmp(parser->raw_buffer.pointer, BOM_UTF16BE, 2)) {
72 parser->encoding = YAML_UTF16BE_ENCODING;
73 parser->raw_buffer.pointer += 2;
76 else if (parser->raw_buffer.last - parser->raw_buffer.pointer >= 3
77 && !memcmp(parser->raw_buffer.pointer, BOM_UTF8, 3)) {
78 parser->encoding = YAML_UTF8_ENCODING;
79 parser->raw_buffer.pointer += 3;
83 parser->encoding = YAML_UTF8_ENCODING;
90 * Update the raw buffer.
94 yaml_parser_update_raw_buffer(yaml_parser_t *parser)
98 /* Return if the raw buffer is full. */
100 if (parser->raw_buffer.start == parser->raw_buffer.pointer
101 && parser->raw_buffer.last == parser->raw_buffer.end)
106 if (parser->eof) return 1;
108 /* Move the remaining bytes in the raw buffer to the beginning. */
110 if (parser->raw_buffer.start < parser->raw_buffer.pointer
111 && parser->raw_buffer.pointer < parser->raw_buffer.last) {
112 memmove(parser->raw_buffer.start, parser->raw_buffer.pointer,
113 parser->raw_buffer.last - parser->raw_buffer.pointer);
115 parser->raw_buffer.last -=
116 parser->raw_buffer.pointer - parser->raw_buffer.start;
117 parser->raw_buffer.pointer = parser->raw_buffer.start;
119 /* Call the read handler to fill the buffer. */
121 if (!parser->read_handler(parser->read_handler_data, parser->raw_buffer.last,
122 parser->raw_buffer.end - parser->raw_buffer.last, &size_read)) {
123 return yaml_parser_set_reader_error(parser, "Input error",
126 parser->raw_buffer.last += size_read;
135 * Ensure that the buffer contains at least `length` characters.
136 * Return 1 on success, 0 on failure.
138 * The length is supposed to be significantly less that the buffer size.
142 yaml_parser_update_buffer(yaml_parser_t *parser, size_t length)
144 assert(parser->read_handler); /* Read handler must be set. */
146 /* If the EOF flag is set and the raw buffer is empty, do nothing. */
148 if (parser->eof && parser->raw_buffer.pointer == parser->raw_buffer.last)
151 /* Return if the buffer contains enough characters. */
153 if (parser->unread >= length)
156 /* Determine the input encoding if it is not known yet. */
158 if (!parser->encoding) {
159 if (!yaml_parser_determine_encoding(parser))
163 /* Move the unread characters to the beginning of the buffer. */
165 if (parser->buffer.start < parser->buffer.pointer
166 && parser->buffer.pointer < parser->buffer.last) {
167 size_t size = parser->buffer.last - parser->buffer.pointer;
168 memmove(parser->buffer.start, parser->buffer.pointer, size);
169 parser->buffer.pointer = parser->buffer.start;
170 parser->buffer.last = parser->buffer.start + size;
172 else if (parser->buffer.pointer == parser->buffer.last) {
173 parser->buffer.pointer = parser->buffer.start;
174 parser->buffer.last = parser->buffer.start;
177 /* Fill the buffer until it has enough characters. */
179 while (parser->unread < length)
181 /* Fill the raw buffer. */
183 if (!yaml_parser_update_raw_buffer(parser)) return 0;
185 /* Decode the raw buffer. */
187 while (parser->raw_buffer.pointer != parser->raw_buffer.last)
189 unsigned int value, value2;
194 int raw_unread = parser->raw_buffer.last - parser->raw_buffer.pointer;
196 /* Decode the next character. */
198 switch (parser->encoding)
200 case YAML_UTF8_ENCODING:
203 * Decode a UTF-8 character. Check RFC 3629
204 * (http://www.ietf.org/rfc/rfc3629.txt) for more details.
206 * The following table (taken from the RFC) is used for
209 * Char. number range | UTF-8 octet sequence
210 * (hexadecimal) | (binary)
211 * --------------------+------------------------------------
212 * 0000 0000-0000 007F | 0xxxxxxx
213 * 0000 0080-0000 07FF | 110xxxxx 10xxxxxx
214 * 0000 0800-0000 FFFF | 1110xxxx 10xxxxxx 10xxxxxx
215 * 0001 0000-0010 FFFF | 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
217 * Additionally, the characters in the range 0xD800-0xDFFF
218 * are prohibited as they are reserved for use with UTF-16
222 /* Determine the length of the UTF-8 sequence. */
224 octet = parser->raw_buffer.pointer[0];
225 width = (octet & 0x80) == 0x00 ? 1 :
226 (octet & 0xE0) == 0xC0 ? 2 :
227 (octet & 0xF0) == 0xE0 ? 3 :
228 (octet & 0xF8) == 0xF0 ? 4 : 0;
230 /* Check if the leading octet is valid. */
233 return yaml_parser_set_reader_error(parser,
234 "Invalid leading UTF-8 octet",
235 parser->offset, octet);
237 /* Check if the raw buffer contains an incomplete character. */
239 if (width > raw_unread) {
241 return yaml_parser_set_reader_error(parser,
242 "Incomplete UTF-8 octet sequence",
249 /* Decode the leading octet. */
251 value = (octet & 0x80) == 0x00 ? octet & 0x7F :
252 (octet & 0xE0) == 0xC0 ? octet & 0x1F :
253 (octet & 0xF0) == 0xE0 ? octet & 0x0F :
254 (octet & 0xF8) == 0xF0 ? octet & 0x07 : 0;
256 /* Check and decode the trailing octets. */
258 for (k = 1; k < width; k ++)
260 octet = parser->raw_buffer.pointer[k];
262 /* Check if the octet is valid. */
264 if ((octet & 0xC0) != 0x80)
265 return yaml_parser_set_reader_error(parser,
266 "Invalid trailing UTF-8 octet",
267 parser->offset+k, octet);
269 /* Decode the octet. */
271 value = (value << 6) + (octet & 0x3F);
274 /* Check the length of the sequence against the value. */
276 if (!((width == 1) ||
277 (width == 2 && value >= 0x80) ||
278 (width == 3 && value >= 0x800) ||
279 (width == 4 && value >= 0x10000)))
280 return yaml_parser_set_reader_error(parser,
281 "Invalid length of a UTF-8 sequence",
284 /* Check the range of the value. */
286 if ((value >= 0xD800 && value <= 0xDFFF) || value > 0x10FFFF)
287 return yaml_parser_set_reader_error(parser,
288 "Invalid Unicode character",
289 parser->offset, value);
293 case YAML_UTF16LE_ENCODING:
294 case YAML_UTF16BE_ENCODING:
296 low = (parser->encoding == YAML_UTF16LE_ENCODING ? 0 : 1);
297 high = (parser->encoding == YAML_UTF16LE_ENCODING ? 1 : 0);
300 * The UTF-16 encoding is not as simple as one might
301 * naively think. Check RFC 2781
302 * (http://www.ietf.org/rfc/rfc2781.txt).
304 * Normally, two subsequent bytes describe a Unicode
305 * character. However a special technique (called a
306 * surrogate pair) is used for specifying character
307 * values larger than 0xFFFF.
309 * A surrogate pair consists of two pseudo-characters:
310 * high surrogate area (0xD800-0xDBFF)
311 * low surrogate area (0xDC00-0xDFFF)
313 * The following formulas are used for decoding
314 * and encoding characters using surrogate pairs:
316 * U = U' + 0x10000 (0x01 00 00 <= U <= 0x10 FF FF)
317 * U' = yyyyyyyyyyxxxxxxxxxx (0 <= U' <= 0x0F FF FF)
318 * W1 = 110110yyyyyyyyyy
319 * W2 = 110111xxxxxxxxxx
321 * where U is the character value, W1 is the high surrogate
322 * area, W2 is the low surrogate area.
325 /* Check for incomplete UTF-16 character. */
327 if (raw_unread < 2) {
329 return yaml_parser_set_reader_error(parser,
330 "Incomplete UTF-16 character",
337 /* Get the character. */
339 value = parser->raw_buffer.pointer[low]
340 + (parser->raw_buffer.pointer[high] << 8);
342 /* Check for unexpected low surrogate area. */
344 if ((value & 0xFC00) == 0xDC00)
345 return yaml_parser_set_reader_error(parser,
346 "Unexpected low surrogate area",
347 parser->offset, value);
349 /* Check for a high surrogate area. */
351 if ((value & 0xFC00) == 0xD800) {
355 /* Check for incomplete surrogate pair. */
357 if (raw_unread < 4) {
359 return yaml_parser_set_reader_error(parser,
360 "Incomplete UTF-16 surrogate pair",
367 /* Get the next character. */
369 unsigned int value2 = parser->raw_buffer.pointer[low+2]
370 + (parser->raw_buffer.pointer[high+2] << 8);
372 /* Check for a low surrogate area. */
374 if ((value2 & 0xFC00) != 0xDC00)
375 return yaml_parser_set_reader_error(parser,
376 "Expected low surrogate area",
377 parser->offset+2, value2);
379 /* Generate the value of the surrogate pair. */
381 value = 0x10000 + ((value & 0x3FF) << 10) + (value2 & 0x3FF);
391 /* Check if the raw buffer contains enough bytes to form a character. */
393 if (incomplete) break;
396 * Check if the character is in the allowed range:
397 * #x9 | #xA | #xD | [#x20-#x7E] (8 bit)
398 * | #x85 | [#xA0-#xD7FF] | [#xE000-#xFFFD] (16 bit)
399 * | [#x10000-#x10FFFF] (32 bit)
402 if (! (value == 0x09 || value == 0x0A || value == 0x0D
403 || (value >= 0x20 && value <= 0x7E)
404 || (value == 0x85) || (value >= 0xA0 && value <= 0xD7FF)
405 || (value >= 0xE000 && value <= 0xFFFD)
406 || (value >= 0x10000 && value <= 0x10FFFF)))
407 return yaml_parser_set_reader_error(parser,
408 "Control characters are not allowed",
409 parser->offset, value);
411 /* Move the raw pointers. */
413 parser->raw_buffer.pointer += width;
414 parser->offset += width;
416 /* Finally put the character into the buffer. */
418 /* 0000 0000-0000 007F -> 0xxxxxxx */
420 *(parser->buffer.last++) = value;
422 /* 0000 0080-0000 07FF -> 110xxxxx 10xxxxxx */
423 else if (value <= 0x7FF) {
424 *(parser->buffer.last++) = 0xC0 + (value >> 6);
425 *(parser->buffer.last++) = 0x80 + (value & 0x3F);
427 /* 0000 0800-0000 FFFF -> 1110xxxx 10xxxxxx 10xxxxxx */
428 else if (value <= 0xFFFF) {
429 *(parser->buffer.last++) = 0xE0 + (value >> 12);
430 *(parser->buffer.last++) = 0x80 + ((value >> 6) & 0x3F);
431 *(parser->buffer.last++) = 0x80 + (value & 0x3F);
433 /* 0001 0000-0010 FFFF -> 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx */
435 *(parser->buffer.last++) = 0xF0 + (value >> 18);
436 *(parser->buffer.last++) = 0x80 + ((value >> 12) & 0x3F);
437 *(parser->buffer.last++) = 0x80 + ((value >> 6) & 0x3F);
438 *(parser->buffer.last++) = 0x80 + (value & 0x3F);
444 /* On EOF, put NUL into the buffer and return. */
447 *(parser->buffer.last++) = '\0';