1package yaml
2
3import (
4	"io"
5)
6
7// Set the reader error and return 0.
8func yaml_parser_set_reader_error(parser *yaml_parser_t, problem string, offset int, value int) bool {
9	parser.error = yaml_READER_ERROR
10	parser.problem = problem
11	parser.problem_offset = offset
12	parser.problem_value = value
13	return false
14}
15
16// Byte order marks.
17const (
18	bom_UTF8    = "\xef\xbb\xbf"
19	bom_UTF16LE = "\xff\xfe"
20	bom_UTF16BE = "\xfe\xff"
21)
22
23// Determine the input stream encoding by checking the BOM symbol. If no BOM is
24// found, the UTF-8 encoding is assumed. Return 1 on success, 0 on failure.
25func yaml_parser_determine_encoding(parser *yaml_parser_t) bool {
26	// Ensure that we had enough bytes in the raw buffer.
27	for !parser.eof && len(parser.raw_buffer)-parser.raw_buffer_pos < 3 {
28		if !yaml_parser_update_raw_buffer(parser) {
29			return false
30		}
31	}
32
33	// Determine the encoding.
34	buf := parser.raw_buffer
35	pos := parser.raw_buffer_pos
36	avail := len(buf) - pos
37	if avail >= 2 && buf[pos] == bom_UTF16LE[0] && buf[pos+1] == bom_UTF16LE[1] {
38		parser.encoding = yaml_UTF16LE_ENCODING
39		parser.raw_buffer_pos += 2
40		parser.offset += 2
41	} else if avail >= 2 && buf[pos] == bom_UTF16BE[0] && buf[pos+1] == bom_UTF16BE[1] {
42		parser.encoding = yaml_UTF16BE_ENCODING
43		parser.raw_buffer_pos += 2
44		parser.offset += 2
45	} else if avail >= 3 && buf[pos] == bom_UTF8[0] && buf[pos+1] == bom_UTF8[1] && buf[pos+2] == bom_UTF8[2] {
46		parser.encoding = yaml_UTF8_ENCODING
47		parser.raw_buffer_pos += 3
48		parser.offset += 3
49	} else {
50		parser.encoding = yaml_UTF8_ENCODING
51	}
52	return true
53}
54
55// Update the raw buffer.
56func yaml_parser_update_raw_buffer(parser *yaml_parser_t) bool {
57	size_read := 0
58
59	// Return if the raw buffer is full.
60	if parser.raw_buffer_pos == 0 && len(parser.raw_buffer) == cap(parser.raw_buffer) {
61		return true
62	}
63
64	// Return on EOF.
65	if parser.eof {
66		return true
67	}
68
69	// Move the remaining bytes in the raw buffer to the beginning.
70	if parser.raw_buffer_pos > 0 && parser.raw_buffer_pos < len(parser.raw_buffer) {
71		copy(parser.raw_buffer, parser.raw_buffer[parser.raw_buffer_pos:])
72	}
73	parser.raw_buffer = parser.raw_buffer[:len(parser.raw_buffer)-parser.raw_buffer_pos]
74	parser.raw_buffer_pos = 0
75
76	// Call the read handler to fill the buffer.
77	size_read, err := parser.read_handler(parser, parser.raw_buffer[len(parser.raw_buffer):cap(parser.raw_buffer)])
78	parser.raw_buffer = parser.raw_buffer[:len(parser.raw_buffer)+size_read]
79	if err == io.EOF {
80		parser.eof = true
81	} else if err != nil {
82		return yaml_parser_set_reader_error(parser, "input error: "+err.Error(), parser.offset, -1)
83	}
84	return true
85}
86
87// Ensure that the buffer contains at least `length` characters.
88// Return true on success, false on failure.
89//
90// The length is supposed to be significantly less that the buffer size.
91func yaml_parser_update_buffer(parser *yaml_parser_t, length int) bool {
92	if parser.read_handler == nil {
93		panic("read handler must be set")
94	}
95
96	// If the EOF flag is set and the raw buffer is empty, do nothing.
97	if parser.eof && parser.raw_buffer_pos == len(parser.raw_buffer) {
98		return true
99	}
100
101	// Return if the buffer contains enough characters.
102	if parser.unread >= length {
103		return true
104	}
105
106	// Determine the input encoding if it is not known yet.
107	if parser.encoding == yaml_ANY_ENCODING {
108		if !yaml_parser_determine_encoding(parser) {
109			return false
110		}
111	}
112
113	// Move the unread characters to the beginning of the buffer.
114	buffer_len := len(parser.buffer)
115	if parser.buffer_pos > 0 && parser.buffer_pos < buffer_len {
116		copy(parser.buffer, parser.buffer[parser.buffer_pos:])
117		buffer_len -= parser.buffer_pos
118		parser.buffer_pos = 0
119	} else if parser.buffer_pos == buffer_len {
120		buffer_len = 0
121		parser.buffer_pos = 0
122	}
123
124	// Open the whole buffer for writing, and cut it before returning.
125	parser.buffer = parser.buffer[:cap(parser.buffer)]
126
127	// Fill the buffer until it has enough characters.
128	first := true
129	for parser.unread < length {
130
131		// Fill the raw buffer if necessary.
132		if !first || parser.raw_buffer_pos == len(parser.raw_buffer) {
133			if !yaml_parser_update_raw_buffer(parser) {
134				parser.buffer = parser.buffer[:buffer_len]
135				return false
136			}
137		}
138		first = false
139
140		// Decode the raw buffer.
141	inner:
142		for parser.raw_buffer_pos != len(parser.raw_buffer) {
143			var value rune
144			var width int
145
146			raw_unread := len(parser.raw_buffer) - parser.raw_buffer_pos
147
148			// Decode the next character.
149			switch parser.encoding {
150			case yaml_UTF8_ENCODING:
151				// Decode a UTF-8 character.  Check RFC 3629
152				// (http://www.ietf.org/rfc/rfc3629.txt) for more details.
153				//
154				// The following table (taken from the RFC) is used for
155				// decoding.
156				//
157				//    Char. number range |        UTF-8 octet sequence
158				//      (hexadecimal)    |              (binary)
159				//   --------------------+------------------------------------
160				//   0000 0000-0000 007F | 0xxxxxxx
161				//   0000 0080-0000 07FF | 110xxxxx 10xxxxxx
162				//   0000 0800-0000 FFFF | 1110xxxx 10xxxxxx 10xxxxxx
163				//   0001 0000-0010 FFFF | 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
164				//
165				// Additionally, the characters in the range 0xD800-0xDFFF
166				// are prohibited as they are reserved for use with UTF-16
167				// surrogate pairs.
168
169				// Determine the length of the UTF-8 sequence.
170				octet := parser.raw_buffer[parser.raw_buffer_pos]
171				switch {
172				case octet&0x80 == 0x00:
173					width = 1
174				case octet&0xE0 == 0xC0:
175					width = 2
176				case octet&0xF0 == 0xE0:
177					width = 3
178				case octet&0xF8 == 0xF0:
179					width = 4
180				default:
181					// The leading octet is invalid.
182					return yaml_parser_set_reader_error(parser,
183						"invalid leading UTF-8 octet",
184						parser.offset, int(octet))
185				}
186
187				// Check if the raw buffer contains an incomplete character.
188				if width > raw_unread {
189					if parser.eof {
190						return yaml_parser_set_reader_error(parser,
191							"incomplete UTF-8 octet sequence",
192							parser.offset, -1)
193					}
194					break inner
195				}
196
197				// Decode the leading octet.
198				switch {
199				case octet&0x80 == 0x00:
200					value = rune(octet & 0x7F)
201				case octet&0xE0 == 0xC0:
202					value = rune(octet & 0x1F)
203				case octet&0xF0 == 0xE0:
204					value = rune(octet & 0x0F)
205				case octet&0xF8 == 0xF0:
206					value = rune(octet & 0x07)
207				default:
208					value = 0
209				}
210
211				// Check and decode the trailing octets.
212				for k := 1; k < width; k++ {
213					octet = parser.raw_buffer[parser.raw_buffer_pos+k]
214
215					// Check if the octet is valid.
216					if (octet & 0xC0) != 0x80 {
217						return yaml_parser_set_reader_error(parser,
218							"invalid trailing UTF-8 octet",
219							parser.offset+k, int(octet))
220					}
221
222					// Decode the octet.
223					value = (value << 6) + rune(octet&0x3F)
224				}
225
226				// Check the length of the sequence against the value.
227				switch {
228				case width == 1:
229				case width == 2 && value >= 0x80:
230				case width == 3 && value >= 0x800:
231				case width == 4 && value >= 0x10000:
232				default:
233					return yaml_parser_set_reader_error(parser,
234						"invalid length of a UTF-8 sequence",
235						parser.offset, -1)
236				}
237
238				// Check the range of the value.
239				if value >= 0xD800 && value <= 0xDFFF || value > 0x10FFFF {
240					return yaml_parser_set_reader_error(parser,
241						"invalid Unicode character",
242						parser.offset, int(value))
243				}
244
245			case yaml_UTF16LE_ENCODING, yaml_UTF16BE_ENCODING:
246				var low, high int
247				if parser.encoding == yaml_UTF16LE_ENCODING {
248					low, high = 0, 1
249				} else {
250					low, high = 1, 0
251				}
252
253				// The UTF-16 encoding is not as simple as one might
254				// naively think.  Check RFC 2781
255				// (http://www.ietf.org/rfc/rfc2781.txt).
256				//
257				// Normally, two subsequent bytes describe a Unicode
258				// character.  However a special technique (called a
259				// surrogate pair) is used for specifying character
260				// values larger than 0xFFFF.
261				//
262				// A surrogate pair consists of two pseudo-characters:
263				//      high surrogate area (0xD800-0xDBFF)
264				//      low surrogate area (0xDC00-0xDFFF)
265				//
266				// The following formulas are used for decoding
267				// and encoding characters using surrogate pairs:
268				//
269				//  U  = U' + 0x10000   (0x01 00 00 <= U <= 0x10 FF FF)
270				//  U' = yyyyyyyyyyxxxxxxxxxx   (0 <= U' <= 0x0F FF FF)
271				//  W1 = 110110yyyyyyyyyy
272				//  W2 = 110111xxxxxxxxxx
273				//
274				// where U is the character value, W1 is the high surrogate
275				// area, W2 is the low surrogate area.
276
277				// Check for incomplete UTF-16 character.
278				if raw_unread < 2 {
279					if parser.eof {
280						return yaml_parser_set_reader_error(parser,
281							"incomplete UTF-16 character",
282							parser.offset, -1)
283					}
284					break inner
285				}
286
287				// Get the character.
288				value = rune(parser.raw_buffer[parser.raw_buffer_pos+low]) +
289					(rune(parser.raw_buffer[parser.raw_buffer_pos+high]) << 8)
290
291				// Check for unexpected low surrogate area.
292				if value&0xFC00 == 0xDC00 {
293					return yaml_parser_set_reader_error(parser,
294						"unexpected low surrogate area",
295						parser.offset, int(value))
296				}
297
298				// Check for a high surrogate area.
299				if value&0xFC00 == 0xD800 {
300					width = 4
301
302					// Check for incomplete surrogate pair.
303					if raw_unread < 4 {
304						if parser.eof {
305							return yaml_parser_set_reader_error(parser,
306								"incomplete UTF-16 surrogate pair",
307								parser.offset, -1)
308						}
309						break inner
310					}
311
312					// Get the next character.
313					value2 := rune(parser.raw_buffer[parser.raw_buffer_pos+low+2]) +
314						(rune(parser.raw_buffer[parser.raw_buffer_pos+high+2]) << 8)
315
316					// Check for a low surrogate area.
317					if value2&0xFC00 != 0xDC00 {
318						return yaml_parser_set_reader_error(parser,
319							"expected low surrogate area",
320							parser.offset+2, int(value2))
321					}
322
323					// Generate the value of the surrogate pair.
324					value = 0x10000 + ((value & 0x3FF) << 10) + (value2 & 0x3FF)
325				} else {
326					width = 2
327				}
328
329			default:
330				panic("impossible")
331			}
332
333			// Check if the character is in the allowed range:
334			//      #x9 | #xA | #xD | [#x20-#x7E]               (8 bit)
335			//      | #x85 | [#xA0-#xD7FF] | [#xE000-#xFFFD]    (16 bit)
336			//      | [#x10000-#x10FFFF]                        (32 bit)
337			switch {
338			case value == 0x09:
339			case value == 0x0A:
340			case value == 0x0D:
341			case value >= 0x20 && value <= 0x7E:
342			case value == 0x85:
343			case value >= 0xA0 && value <= 0xD7FF:
344			case value >= 0xE000 && value <= 0xFFFD:
345			case value >= 0x10000 && value <= 0x10FFFF:
346			default:
347				return yaml_parser_set_reader_error(parser,
348					"control characters are not allowed",
349					parser.offset, int(value))
350			}
351
352			// Move the raw pointers.
353			parser.raw_buffer_pos += width
354			parser.offset += width
355
356			// Finally put the character into the buffer.
357			if value <= 0x7F {
358				// 0000 0000-0000 007F . 0xxxxxxx
359				parser.buffer[buffer_len+0] = byte(value)
360				buffer_len += 1
361			} else if value <= 0x7FF {
362				// 0000 0080-0000 07FF . 110xxxxx 10xxxxxx
363				parser.buffer[buffer_len+0] = byte(0xC0 + (value >> 6))
364				parser.buffer[buffer_len+1] = byte(0x80 + (value & 0x3F))
365				buffer_len += 2
366			} else if value <= 0xFFFF {
367				// 0000 0800-0000 FFFF . 1110xxxx 10xxxxxx 10xxxxxx
368				parser.buffer[buffer_len+0] = byte(0xE0 + (value >> 12))
369				parser.buffer[buffer_len+1] = byte(0x80 + ((value >> 6) & 0x3F))
370				parser.buffer[buffer_len+2] = byte(0x80 + (value & 0x3F))
371				buffer_len += 3
372			} else {
373				// 0001 0000-0010 FFFF . 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
374				parser.buffer[buffer_len+0] = byte(0xF0 + (value >> 18))
375				parser.buffer[buffer_len+1] = byte(0x80 + ((value >> 12) & 0x3F))
376				parser.buffer[buffer_len+2] = byte(0x80 + ((value >> 6) & 0x3F))
377				parser.buffer[buffer_len+3] = byte(0x80 + (value & 0x3F))
378				buffer_len += 4
379			}
380
381			parser.unread++
382		}
383
384		// On EOF, put NUL into the buffer and return.
385		if parser.eof {
386			parser.buffer[buffer_len] = 0
387			buffer_len++
388			parser.unread++
389			break
390		}
391	}
392	parser.buffer = parser.buffer[:buffer_len]
393	return true
394}
395