/src/libyaml/src/reader.c

Source

#include "yaml_private.h"

/*
 * Declarations.
 */

static int
yaml_parser_set_reader_error(yaml_parser_t *parser, const char *problem,
        size_t offset, int value);

static int
yaml_parser_update_raw_buffer(yaml_parser_t *parser);

static int
yaml_parser_determine_encoding(yaml_parser_t *parser);

YAML_DECLARE(int)
yaml_parser_update_buffer(yaml_parser_t *parser, size_t length);

/*
 * Set the reader error and return 0.
 */

static int
yaml_parser_set_reader_error(yaml_parser_t *parser, const char *problem,
        size_t offset, int value)
{
    parser->error = YAML_READER_ERROR;
    parser->problem = problem;
    parser->problem_offset = offset;
    parser->problem_value = value;

    return 0;
}

/*
 * Byte order marks.
 */

#define BOM_UTF8    "\xef\xbb\xbf"
#define BOM_UTF16LE "\xff\xfe"
#define BOM_UTF16BE "\xfe\xff"

/*
 * Determine the input stream encoding by checking the BOM symbol. If no BOM is
 * found, the UTF-8 encoding is assumed. Return 1 on success, 0 on failure.
 */

static int
yaml_parser_determine_encoding(yaml_parser_t *parser)
{
    /* Ensure that we had enough bytes in the raw buffer. */

    while (!parser->eof
            && parser->raw_buffer.last - parser->raw_buffer.pointer < 3) {
        if (!yaml_parser_update_raw_buffer(parser)) {
            return 0;
        }
    }

    /* Determine the encoding. */

    if (parser->raw_buffer.last - parser->raw_buffer.pointer >= 2
            && !memcmp(parser->raw_buffer.pointer, BOM_UTF16LE, 2)) {
        parser->encoding = YAML_UTF16LE_ENCODING;
        parser->raw_buffer.pointer += 2;
        parser->offset += 2;
    }
    else if (parser->raw_buffer.last - parser->raw_buffer.pointer >= 2
            && !memcmp(parser->raw_buffer.pointer, BOM_UTF16BE, 2)) {
        parser->encoding = YAML_UTF16BE_ENCODING;
        parser->raw_buffer.pointer += 2;
        parser->offset += 2;
    }
    else if (parser->raw_buffer.last - parser->raw_buffer.pointer >= 3
            && !memcmp(parser->raw_buffer.pointer, BOM_UTF8, 3)) {
        parser->encoding = YAML_UTF8_ENCODING;
        parser->raw_buffer.pointer += 3;
        parser->offset += 3;
    }
    else {
        parser->encoding = YAML_UTF8_ENCODING;
    }

    return 1;
}

/*
 * Update the raw buffer.
 */

static int
yaml_parser_update_raw_buffer(yaml_parser_t *parser)
{
    size_t size_read = 0;

    /* Return if the raw buffer is full. */

    if (parser->raw_buffer.start == parser->raw_buffer.pointer
            && parser->raw_buffer.last == parser->raw_buffer.end)
        return 1;

    /* Return on EOF. */

    if (parser->eof) return 1;

    /* Move the remaining bytes in the raw buffer to the beginning. */

    if (parser->raw_buffer.start < parser->raw_buffer.pointer
            && parser->raw_buffer.pointer < parser->raw_buffer.last) {
        memmove(parser->raw_buffer.start, parser->raw_buffer.pointer,
                parser->raw_buffer.last - parser->raw_buffer.pointer);
    }
    parser->raw_buffer.last -=
        parser->raw_buffer.pointer - parser->raw_buffer.start;
    parser->raw_buffer.pointer = parser->raw_buffer.start;

    /* Call the read handler to fill the buffer. */

    if (!parser->read_handler(parser->read_handler_data, parser->raw_buffer.last,
                parser->raw_buffer.end - parser->raw_buffer.last, &size_read)) {
        return yaml_parser_set_reader_error(parser, "input error",
                parser->offset, -1);
    }
    parser->raw_buffer.last += size_read;
    if (!size_read) {
        parser->eof = 1;
    }

    return 1;
}

/*
 * Ensure that the buffer contains at least `length` characters.
 * Return 1 on success, 0 on failure.
 *
 * The length is supposed to be significantly less that the buffer size.
 */

YAML_DECLARE(int)
yaml_parser_update_buffer(yaml_parser_t *parser, size_t length)
{
    int first = 1;

    assert(parser->read_handler);   /* Read handler must be set. */

    /* If the EOF flag is set and the raw buffer is empty, do nothing. */

    if (parser->eof && parser->raw_buffer.pointer == parser->raw_buffer.last)
        return 1;

    /* Return if the buffer contains enough characters. */

    if (parser->unread >= length)
        return 1;

    /* Determine the input encoding if it is not known yet. */

    if (!parser->encoding) {
        if (!yaml_parser_determine_encoding(parser))
            return 0;
    }

    /* Move the unread characters to the beginning of the buffer. */

    if (parser->buffer.start < parser->buffer.pointer
            && parser->buffer.pointer < parser->buffer.last) {
        size_t size = parser->buffer.last - parser->buffer.pointer;
        memmove(parser->buffer.start, parser->buffer.pointer, size);
        parser->buffer.pointer = parser->buffer.start;
        parser->buffer.last = parser->buffer.start + size;
    }
    else if (parser->buffer.pointer == parser->buffer.last) {
        parser->buffer.pointer = parser->buffer.start;
        parser->buffer.last = parser->buffer.start;
    }

    /* Fill the buffer until it has enough characters. */

    while (parser->unread < length)
    {
        /* Fill the raw buffer if necessary. */

        if (!first || parser->raw_buffer.pointer == parser->raw_buffer.last) {
            if (!yaml_parser_update_raw_buffer(parser)) return 0;
        }
        first = 0;

        /* Decode the raw buffer. */

        while (parser->raw_buffer.pointer != parser->raw_buffer.last)
        {
            unsigned int value = 0, value2 = 0;
            int incomplete = 0;
            unsigned char octet;
            unsigned int width = 0;
            int low, high;
            size_t k;
            size_t raw_unread = parser->raw_buffer.last - parser->raw_buffer.pointer;

            /* Decode the next character. */

            switch (parser->encoding)
            {
                case YAML_UTF8_ENCODING:

                    /*
                     * Decode a UTF-8 character.  Check RFC 3629
                     * (http://www.ietf.org/rfc/rfc3629.txt) for more details.
                     *
                     * The following table (taken from the RFC) is used for
                     * decoding.
                     *
                     *    Char. number range |        UTF-8 octet sequence
                     *      (hexadecimal)    |              (binary)
                     *   --------------------+------------------------------------
                     *   0000 0000-0000 007F | 0xxxxxxx
                     *   0000 0080-0000 07FF | 110xxxxx 10xxxxxx
                     *   0000 0800-0000 FFFF | 1110xxxx 10xxxxxx 10xxxxxx
                     *   0001 0000-0010 FFFF | 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
                     *
                     * Additionally, the characters in the range 0xD800-0xDFFF
                     * are prohibited as they are reserved for use with UTF-16
                     * surrogate pairs.
                     */

                    /* Determine the length of the UTF-8 sequence. */

                    octet = parser->raw_buffer.pointer[0];
                    width = (octet & 0x80) == 0x00 ? 1 :
                            (octet & 0xE0) == 0xC0 ? 2 :
                            (octet & 0xF0) == 0xE0 ? 3 :
                            (octet & 0xF8) == 0xF0 ? 4 : 0;

                    /* Check if the leading octet is valid. */

                    if (!width)
                        return yaml_parser_set_reader_error(parser,
                                "invalid leading UTF-8 octet",
                                parser->offset, octet);

                    /* Check if the raw buffer contains an incomplete character. */

                    if (width > raw_unread) {
                        if (parser->eof) {
                            return yaml_parser_set_reader_error(parser,
                                    "incomplete UTF-8 octet sequence",
                                    parser->offset, -1);
                        }
                        incomplete = 1;
                        break;
                    }

                    /* Decode the leading octet. */

                    value = (octet & 0x80) == 0x00 ? octet & 0x7F :
                            (octet & 0xE0) == 0xC0 ? octet & 0x1F :
                            (octet & 0xF0) == 0xE0 ? octet & 0x0F :
                            (octet & 0xF8) == 0xF0 ? octet & 0x07 : 0;

                    /* Check and decode the trailing octets. */

                    for (k = 1; k < width; k ++)
                    {
                        octet = parser->raw_buffer.pointer[k];

                        /* Check if the octet is valid. */

                        if ((octet & 0xC0) != 0x80)
                            return yaml_parser_set_reader_error(parser,
                                    "invalid trailing UTF-8 octet",
                                    parser->offset+k, octet);

                        /* Decode the octet. */

                        value = (value << 6) + (octet & 0x3F);
                    }

                    /* Check the length of the sequence against the value. */

                    if (!((width == 1) ||
                            (width == 2 && value >= 0x80) ||
                            (width == 3 && value >= 0x800) ||
                            (width == 4 && value >= 0x10000)))
                        return yaml_parser_set_reader_error(parser,
                                "invalid length of a UTF-8 sequence",
                                parser->offset, -1);

                    /* Check the range of the value. */

                    if ((value >= 0xD800 && value <= 0xDFFF) || value > 0x10FFFF)
                        return yaml_parser_set_reader_error(parser,
                                "invalid Unicode character",
                                parser->offset, value);

                    break;

                case YAML_UTF16LE_ENCODING:
                case YAML_UTF16BE_ENCODING:

                    low = (parser->encoding == YAML_UTF16LE_ENCODING ? 0 : 1);
                    high = (parser->encoding == YAML_UTF16LE_ENCODING ? 1 : 0);

                    /*
                     * The UTF-16 encoding is not as simple as one might
                     * naively think.  Check RFC 2781
                     * (http://www.ietf.org/rfc/rfc2781.txt).
                     *
                     * Normally, two subsequent bytes describe a Unicode
                     * character.  However a special technique (called a
                     * surrogate pair) is used for specifying character
                     * values larger than 0xFFFF.
                     *
                     * A surrogate pair consists of two pseudo-characters:
                     *      high surrogate area (0xD800-0xDBFF)
                     *      low surrogate area (0xDC00-0xDFFF)
                     *
                     * The following formulas are used for decoding
                     * and encoding characters using surrogate pairs:
                     *
                     *  U  = U' + 0x10000   (0x01 00 00 <= U <= 0x10 FF FF)
                     *  U' = yyyyyyyyyyxxxxxxxxxx   (0 <= U' <= 0x0F FF FF)
                     *  W1 = 110110yyyyyyyyyy
                     *  W2 = 110111xxxxxxxxxx
                     *
                     * where U is the character value, W1 is the high surrogate
                     * area, W2 is the low surrogate area.
                     */

                    /* Check for incomplete UTF-16 character. */

                    if (raw_unread < 2) {
                        if (parser->eof) {
                            return yaml_parser_set_reader_error(parser,
                                    "incomplete UTF-16 character",
                                    parser->offset, -1);
                        }
                        incomplete = 1;
                        break;
                    }

                    /* Get the character. */

                    value = parser->raw_buffer.pointer[low]
                        + (parser->raw_buffer.pointer[high] << 8);

                    /* Check for unexpected low surrogate area. */

                    if ((value & 0xFC00) == 0xDC00)
                        return yaml_parser_set_reader_error(parser,
                                "unexpected low surrogate area",
                                parser->offset, value);

                    /* Check for a high surrogate area. */

                    if ((value & 0xFC00) == 0xD800) {

                        width = 4;

                        /* Check for incomplete surrogate pair. */

                        if (raw_unread < 4) {
                            if (parser->eof) {
                                return yaml_parser_set_reader_error(parser,
                                        "incomplete UTF-16 surrogate pair",
                                        parser->offset, -1);
                            }
                            incomplete = 1;
                            break;
                        }

                        /* Get the next character. */

                        value2 = parser->raw_buffer.pointer[low+2]
                            + (parser->raw_buffer.pointer[high+2] << 8);

                        /* Check for a low surrogate area. */

                        if ((value2 & 0xFC00) != 0xDC00)
                            return yaml_parser_set_reader_error(parser,
                                    "expected low surrogate area",
                                    parser->offset+2, value2);

                        /* Generate the value of the surrogate pair. */

                        value = 0x10000 + ((value & 0x3FF) << 10) + (value2 & 0x3FF);
                    }

                    else {
                        width = 2;
                    }

                    break;

                default:
                    assert(1);      /* Impossible. */
            }

            /* Check if the raw buffer contains enough bytes to form a character. */

            if (incomplete) break;

            /*
             * Check if the character is in the allowed range:
             *      #x9 | #xA | #xD | [#x20-#x7E]               (8 bit)
             *      | #x85 | [#xA0-#xD7FF] | [#xE000-#xFFFD]    (16 bit)
             *      | [#x10000-#x10FFFF]                        (32 bit)
             */

            if (! (value == 0x09 || value == 0x0A || value == 0x0D
                        || (value >= 0x20 && value <= 0x7E)
                        || (value == 0x85) || (value >= 0xA0 && value <= 0xD7FF)
                        || (value >= 0xE000 && value <= 0xFFFD)
                        || (value >= 0x10000 && value <= 0x10FFFF)))
                return yaml_parser_set_reader_error(parser,
                        "control characters are not allowed",
                        parser->offset, value);

            /* Move the raw pointers. */

            parser->raw_buffer.pointer += width;
            parser->offset += width;

            /* Finally put the character into the buffer. */

            /* 0000 0000-0000 007F -> 0xxxxxxx */
            if (value <= 0x7F) {
                *(parser->buffer.last++) = value;
            }
            /* 0000 0080-0000 07FF -> 110xxxxx 10xxxxxx */
            else if (value <= 0x7FF) {
                *(parser->buffer.last++) = 0xC0 + (value >> 6);
                *(parser->buffer.last++) = 0x80 + (value & 0x3F);
            }
            /* 0000 0800-0000 FFFF -> 1110xxxx 10xxxxxx 10xxxxxx */
            else if (value <= 0xFFFF) {
                *(parser->buffer.last++) = 0xE0 + (value >> 12);
                *(parser->buffer.last++) = 0x80 + ((value >> 6) & 0x3F);
                *(parser->buffer.last++) = 0x80 + (value & 0x3F);
            }
            /* 0001 0000-0010 FFFF -> 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx */
            else {
                *(parser->buffer.last++) = 0xF0 + (value >> 18);
                *(parser->buffer.last++) = 0x80 + ((value >> 12) & 0x3F);
                *(parser->buffer.last++) = 0x80 + ((value >> 6) & 0x3F);
                *(parser->buffer.last++) = 0x80 + (value & 0x3F);
            }

            parser->unread ++;
        }

        /* On EOF, put NUL into the buffer and return. */

        if (parser->eof) {
            *(parser->buffer.last++) = '\0';
            parser->unread ++;
            return 1;
        }

    }

    if (parser->offset >= MAX_FILE_SIZE) {
        return yaml_parser_set_reader_error(parser, "input is too long",
            parser->offset, -1);
    }

    return 1;
}

Coverage Report

Created: 2026-08-31 06:27

Line	Count	Source
1
2		#include "yaml_private.h"
3
4		/*
5		* Declarations.
6		*/
7
8		static int
9		yaml_parser_set_reader_error(yaml_parser_t parser, const char problem,
10		size_t offset, int value);
11
12		static int
13		yaml_parser_update_raw_buffer(yaml_parser_t *parser);
14
15		static int
16		yaml_parser_determine_encoding(yaml_parser_t *parser);
17
18		YAML_DECLARE(int)
19		yaml_parser_update_buffer(yaml_parser_t *parser, size_t length);
20
21		/*
22		* Set the reader error and return 0.
23		*/
24
25		static int
26		yaml_parser_set_reader_error(yaml_parser_t parser, const char problem,
27		size_t offset, int value)
28	280	{
29	280	parser->error = YAML_READER_ERROR;
30	280	parser->problem = problem;
31	280	parser->problem_offset = offset;
32	280	parser->problem_value = value;
33
34	280	return 0;
35	280	}
36
37		/*
38		* Byte order marks.
39		*/
40
41	7.23k	#define BOM_UTF8 "\xef\xbb\xbf"
42	7.82k	#define BOM_UTF16LE "\xff\xfe"
43	7.69k	#define BOM_UTF16BE "\xfe\xff"
44
45		/*
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.
48		*/
49
50		static int
51		yaml_parser_determine_encoding(yaml_parser_t *parser)
52	7.91k	{
53		/* Ensure that we had enough bytes in the raw buffer. */
54
55	16.2k	while (!parser->eof
56	15.8k	&& parser->raw_buffer.last - parser->raw_buffer.pointer < 3) {
57	8.35k	if (!yaml_parser_update_raw_buffer(parser)) {
58	0	return 0;
59	0	}
60	8.35k	}
61
62		/* Determine the encoding. */
63
64	7.91k	if (parser->raw_buffer.last - parser->raw_buffer.pointer >= 2
65	7.82k	&& !memcmp(parser->raw_buffer.pointer, BOM_UTF16LE, 2)) {
66	135	parser->encoding = YAML_UTF16LE_ENCODING;
67	135	parser->raw_buffer.pointer += 2;
68	135	parser->offset += 2;
69	135	}
70	7.77k	else if (parser->raw_buffer.last - parser->raw_buffer.pointer >= 2
71	7.69k	&& !memcmp(parser->raw_buffer.pointer, BOM_UTF16BE, 2)) {
72	105	parser->encoding = YAML_UTF16BE_ENCODING;
73	105	parser->raw_buffer.pointer += 2;
74	105	parser->offset += 2;
75	105	}
76	7.67k	else if (parser->raw_buffer.last - parser->raw_buffer.pointer >= 3
77	7.23k	&& !memcmp(parser->raw_buffer.pointer, BOM_UTF8, 3)) {
78	14	parser->encoding = YAML_UTF8_ENCODING;
79	14	parser->raw_buffer.pointer += 3;
80	14	parser->offset += 3;
81	14	}
82	7.65k	else {
83	7.65k	parser->encoding = YAML_UTF8_ENCODING;
84	7.65k	}
85
86	7.91k	return 1;
87	7.91k	}
88
89		/*
90		* Update the raw buffer.
91		*/
92
93		static int
94		yaml_parser_update_raw_buffer(yaml_parser_t *parser)
95	21.6k	{
96	21.6k	size_t size_read = 0;
97
98		/* Return if the raw buffer is full. */
99
100	21.6k	if (parser->raw_buffer.start == parser->raw_buffer.pointer
101	8.35k	&& parser->raw_buffer.last == parser->raw_buffer.end)
102	0	return 1;
103
104		/* Return on EOF. */
105
106	21.6k	if (parser->eof) return 1;
107
108		/* Move the remaining bytes in the raw buffer to the beginning. */
109
110	21.5k	if (parser->raw_buffer.start < parser->raw_buffer.pointer
111	13.2k	&& parser->raw_buffer.pointer < parser->raw_buffer.last) {
112	408	memmove(parser->raw_buffer.start, parser->raw_buffer.pointer,
113	408	parser->raw_buffer.last - parser->raw_buffer.pointer);
114	408	}
115	21.5k	parser->raw_buffer.last -=
116	21.5k	parser->raw_buffer.pointer - parser->raw_buffer.start;
117	21.5k	parser->raw_buffer.pointer = parser->raw_buffer.start;
118
119		/* Call the read handler to fill the buffer. */
120
121	21.5k	if (!parser->read_handler(parser->read_handler_data, parser->raw_buffer.last,
122	21.5k	parser->raw_buffer.end - parser->raw_buffer.last, &size_read)) {
123	0	return yaml_parser_set_reader_error(parser, "input error",
124	0	parser->offset, -1);
125	0	}
126	21.5k	parser->raw_buffer.last += size_read;
127	21.5k	if (!size_read) {
128	7.41k	parser->eof = 1;
129	7.41k	}
130
131	21.5k	return 1;
132	21.5k	}
133
134		/*
135		* Ensure that the buffer contains at least `length` characters.
136		* Return 1 on success, 0 on failure.
137		*
138		* The length is supposed to be significantly less that the buffer size.
139		*/
140
141		YAML_DECLARE(int)
142		yaml_parser_update_buffer(yaml_parser_t *parser, size_t length)
143	32.3k	{
144	32.3k	int first = 1;
145
146	32.3k	assert(parser->read_handler); /* Read handler must be set. */
147
148		/* If the EOF flag is set and the raw buffer is empty, do nothing. */
149
150	32.3k	if (parser->eof && parser->raw_buffer.pointer == parser->raw_buffer.last)
151	11.2k	return 1;
152
153		/* Return if the buffer contains enough characters. */
154
155	21.1k	if (parser->unread >= length)
156	0	return 1;
157
158		/* Determine the input encoding if it is not known yet. */
159
160	21.1k	if (!parser->encoding) {
161	7.91k	if (!yaml_parser_determine_encoding(parser))
162	0	return 0;
163	7.91k	}
164
165		/* Move the unread characters to the beginning of the buffer. */
166
167	21.1k	if (parser->buffer.start < parser->buffer.pointer
168	12.4k	&& parser->buffer.pointer < parser->buffer.last) {
169	6.05k	size_t size = parser->buffer.last - parser->buffer.pointer;
170	6.05k	memmove(parser->buffer.start, parser->buffer.pointer, size);
171	6.05k	parser->buffer.pointer = parser->buffer.start;
172	6.05k	parser->buffer.last = parser->buffer.start + size;
173	6.05k	}
174	15.0k	else if (parser->buffer.pointer == parser->buffer.last) {
175	14.2k	parser->buffer.pointer = parser->buffer.start;
176	14.2k	parser->buffer.last = parser->buffer.start;
177	14.2k	}
178
179		/* Fill the buffer until it has enough characters. */
180
181	35.1k	while (parser->unread < length)
182	21.5k	{
183		/* Fill the raw buffer if necessary. */
184
185	21.5k	if (!first \|\| parser->raw_buffer.pointer == parser->raw_buffer.last) {
186	13.2k	if (!yaml_parser_update_raw_buffer(parser)) return 0;
187	13.2k	}
188	21.5k	first = 0;
189
190		/* Decode the raw buffer. */
191
192	101M	while (parser->raw_buffer.pointer != parser->raw_buffer.last)
193	101M	{
194	101M	unsigned int value = 0, value2 = 0;
195	101M	int incomplete = 0;
196	101M	unsigned char octet;
197	101M	unsigned int width = 0;
198	101M	int low, high;
199	101M	size_t k;
200	101M	size_t raw_unread = parser->raw_buffer.last - parser->raw_buffer.pointer;
201
202		/* Decode the next character. */
203
204	101M	switch (parser->encoding)
205	101M	{
206	99.8M	case YAML_UTF8_ENCODING:
207
208		/*
209		* Decode a UTF-8 character. Check RFC 3629
210		* (http://www.ietf.org/rfc/rfc3629.txt) for more details.
211		*
212		* The following table (taken from the RFC) is used for
213		* decoding.
214		*
215		* Char. number range \| UTF-8 octet sequence
216		* (hexadecimal) \| (binary)
217		* --------------------+------------------------------------
218		* 0000 0000-0000 007F \| 0xxxxxxx
219		* 0000 0080-0000 07FF \| 110xxxxx 10xxxxxx
220		* 0000 0800-0000 FFFF \| 1110xxxx 10xxxxxx 10xxxxxx
221		* 0001 0000-0010 FFFF \| 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
222		*
223		* Additionally, the characters in the range 0xD800-0xDFFF
224		* are prohibited as they are reserved for use with UTF-16
225		* surrogate pairs.
226		*/
227
228		/* Determine the length of the UTF-8 sequence. */
229
230	99.8M	octet = parser->raw_buffer.pointer[0];
231	99.8M	width = (octet & 0x80) == 0x00 ? 1 :
232	99.8M	(octet & 0xE0) == 0xC0 ? 2 :
233	1.68M	(octet & 0xF0) == 0xE0 ? 3 :
234	1.67M	(octet & 0xF8) == 0xF0 ? 4 : 0;
235
236		/* Check if the leading octet is valid. */
237
238	99.8M	if (!width)
239	38	return yaml_parser_set_reader_error(parser,
240	38	"invalid leading UTF-8 octet",
241	38	parser->offset, octet);
242
243		/* Check if the raw buffer contains an incomplete character. */
244
245	99.8M	if (width > raw_unread) {
246	823	if (parser->eof) {
247	87	return yaml_parser_set_reader_error(parser,
248	87	"incomplete UTF-8 octet sequence",
249	87	parser->offset, -1);
250	87	}
251	736	incomplete = 1;
252	736	break;
253	823	}
254
255		/* Decode the leading octet. */
256
257	99.8M	value = (octet & 0x80) == 0x00 ? octet & 0x7F :
258	99.8M	(octet & 0xE0) == 0xC0 ? octet & 0x1F :
259	1.68M	(octet & 0xF0) == 0xE0 ? octet & 0x0F :
260	1.67M	(octet & 0xF8) == 0xF0 ? octet & 0x07 : 0;
261
262		/* Check and decode the trailing octets. */
263
264	103M	for (k = 1; k < width; k ++)
265	3.35M	{
266	3.35M	octet = parser->raw_buffer.pointer[k];
267
268		/* Check if the octet is valid. */
269
270	3.35M	if ((octet & 0xC0) != 0x80)
271	18	return yaml_parser_set_reader_error(parser,
272	18	"invalid trailing UTF-8 octet",
273	18	parser->offset+k, octet);
274
275		/* Decode the octet. */
276
277	3.35M	value = (value << 6) + (octet & 0x3F);
278	3.35M	}
279
280		/* Check the length of the sequence against the value. */
281
282	99.8M	if (!((width == 1) \|\|
283	1.68M	(width == 2 && value >= 0x80) \|\|
284	1.67M	(width == 3 && value >= 0x800) \|\|
285	2.15k	(width == 4 && value >= 0x10000)))
286	37	return yaml_parser_set_reader_error(parser,
287	37	"invalid length of a UTF-8 sequence",
288	37	parser->offset, -1);
289
290		/* Check the range of the value. */
291
292	99.8M	if ((value >= 0xD800 && value <= 0xDFFF) \|\| value > 0x10FFFF)
293	18	return yaml_parser_set_reader_error(parser,
294	18	"invalid Unicode character",
295	18	parser->offset, value);
296
297	99.8M	break;
298
299	99.8M	case YAML_UTF16LE_ENCODING:
300	2.13M	case YAML_UTF16BE_ENCODING:
301
302	2.13M	low = (parser->encoding == YAML_UTF16LE_ENCODING ? 0 : 1);
303	2.13M	high = (parser->encoding == YAML_UTF16LE_ENCODING ? 1 : 0);
304
305		/*
306		* The UTF-16 encoding is not as simple as one might
307		* naively think. Check RFC 2781
308		* (http://www.ietf.org/rfc/rfc2781.txt).
309		*
310		* Normally, two subsequent bytes describe a Unicode
311		* character. However a special technique (called a
312		* surrogate pair) is used for specifying character
313		* values larger than 0xFFFF.
314		*
315		* A surrogate pair consists of two pseudo-characters:
316		* high surrogate area (0xD800-0xDBFF)
317		* low surrogate area (0xDC00-0xDFFF)
318		*
319		* The following formulas are used for decoding
320		* and encoding characters using surrogate pairs:
321		*
322		* U = U' + 0x10000 (0x01 00 00 <= U <= 0x10 FF FF)
323		* U' = yyyyyyyyyyxxxxxxxxxx (0 <= U' <= 0x0F FF FF)
324		* W1 = 110110yyyyyyyyyy
325		* W2 = 110111xxxxxxxxxx
326		*
327		* where U is the character value, W1 is the high surrogate
328		* area, W2 is the low surrogate area.
329		*/
330
331		/* Check for incomplete UTF-16 character. */
332
333	2.13M	if (raw_unread < 2) {
334	64	if (parser->eof) {
335	22	return yaml_parser_set_reader_error(parser,
336	22	"incomplete UTF-16 character",
337	22	parser->offset, -1);
338	22	}
339	42	incomplete = 1;
340	42	break;
341	64	}
342
343		/* Get the character. */
344
345	2.13M	value = parser->raw_buffer.pointer[low]
346	2.13M	+ (parser->raw_buffer.pointer[high] << 8);
347
348		/* Check for unexpected low surrogate area. */
349
350	2.13M	if ((value & 0xFC00) == 0xDC00)
351	2	return yaml_parser_set_reader_error(parser,
352	2	"unexpected low surrogate area",
353	2	parser->offset, value);
354
355		/* Check for a high surrogate area. */
356
357	2.13M	if ((value & 0xFC00) == 0xD800) {
358
359	33.3k	width = 4;
360
361		/* Check for incomplete surrogate pair. */
362
363	33.3k	if (raw_unread < 4) {
364	46	if (parser->eof) {
365	6	return yaml_parser_set_reader_error(parser,
366	6	"incomplete UTF-16 surrogate pair",
367	6	parser->offset, -1);
368	6	}
369	40	incomplete = 1;
370	40	break;
371	46	}
372
373		/* Get the next character. */
374
375	33.2k	value2 = parser->raw_buffer.pointer[low+2]
376	33.2k	+ (parser->raw_buffer.pointer[high+2] << 8);
377
378		/* Check for a low surrogate area. */
379
380	33.2k	if ((value2 & 0xFC00) != 0xDC00)
381	12	return yaml_parser_set_reader_error(parser,
382	12	"expected low surrogate area",
383	12	parser->offset+2, value2);
384
385		/* Generate the value of the surrogate pair. */
386
387	33.2k	value = 0x10000 + ((value & 0x3FF) << 10) + (value2 & 0x3FF);
388	33.2k	}
389
390	2.10M	else {
391	2.10M	width = 2;
392	2.10M	}
393
394	2.13M	break;
395
396	2.13M	default:
397	0	assert(1); /* Impossible. */
398	101M	}
399
400		/* Check if the raw buffer contains enough bytes to form a character. */
401
402	101M	if (incomplete) break;
403
404		/*
405		* Check if the character is in the allowed range:
406		* #x9 \| #xA \| #xD \| [#x20-#x7E] (8 bit)
407		* \| #x85 \| [#xA0-#xD7FF] \| [#xE000-#xFFFD] (16 bit)
408		* \| [#x10000-#x10FFFF] (32 bit)
409		*/
410
411	101M	if (! (value == 0x09 \|\| value == 0x0A \|\| value == 0x0D
412	91.7M	\|\| (value >= 0x20 && value <= 0x7E)
413	3.81M	\|\| (value == 0x85) \|\| (value >= 0xA0 && value <= 0xD7FF)
414	311k	\|\| (value >= 0xE000 && value <= 0xFFFD)
415	35.4k	\|\| (value >= 0x10000 && value <= 0x10FFFF)))
416	40	return yaml_parser_set_reader_error(parser,
417	40	"control characters are not allowed",
418	40	parser->offset, value);
419
420		/* Move the raw pointers. */
421
422	101M	parser->raw_buffer.pointer += width;
423	101M	parser->offset += width;
424
425		/* Finally put the character into the buffer. */
426
427		/* 0000 0000-0000 007F -> 0xxxxxxx */
428	101M	if (value <= 0x7F) {
429	98.1M	*(parser->buffer.last++) = value;
430	98.1M	}
431		/* 0000 0080-0000 07FF -> 110xxxxx 10xxxxxx */
432	3.81M	else if (value <= 0x7FF) {
433	17.1k	*(parser->buffer.last++) = 0xC0 + (value >> 6);
434	17.1k	*(parser->buffer.last++) = 0x80 + (value & 0x3F);
435	17.1k	}
436		/* 0000 0800-0000 FFFF -> 1110xxxx 10xxxxxx 10xxxxxx */
437	3.79M	else if (value <= 0xFFFF) {
438	3.76M	*(parser->buffer.last++) = 0xE0 + (value >> 12);
439	3.76M	*(parser->buffer.last++) = 0x80 + ((value >> 6) & 0x3F);
440	3.76M	*(parser->buffer.last++) = 0x80 + (value & 0x3F);
441	3.76M	}
442		/* 0001 0000-0010 FFFF -> 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx */
443	35.3k	else {
444	35.3k	*(parser->buffer.last++) = 0xF0 + (value >> 18);
445	35.3k	*(parser->buffer.last++) = 0x80 + ((value >> 12) & 0x3F);
446	35.3k	*(parser->buffer.last++) = 0x80 + ((value >> 6) & 0x3F);
447	35.3k	*(parser->buffer.last++) = 0x80 + (value & 0x3F);
448	35.3k	}
449
450	101M	parser->unread ++;
451	101M	}
452
453		/* On EOF, put NUL into the buffer and return. */
454
455	21.2k	if (parser->eof) {
456	7.23k	*(parser->buffer.last++) = '\0';
457	7.23k	parser->unread ++;
458	7.23k	return 1;
459	7.23k	}
460
461	21.2k	}
462
463	13.6k	if (parser->offset >= MAX_FILE_SIZE) {
464	0	return yaml_parser_set_reader_error(parser, "input is too long",
465	0	parser->offset, -1);
466	0	}
467
468	13.6k	return 1;
469	13.6k	}