1
|
|
|
<?php |
2
|
|
|
|
3
|
|
|
/* |
4
|
|
|
* This file is part of the kaloa/xmp package. |
5
|
|
|
* |
6
|
|
|
* For full copyright and license information, please view the LICENSE file |
7
|
|
|
* that was distributed with this source code. |
8
|
|
|
*/ |
9
|
|
|
|
10
|
|
|
namespace Kaloa\Xmp; |
11
|
|
|
|
12
|
|
|
use DOMDocument; |
13
|
|
|
use ErrorException; |
14
|
|
|
use Exception; |
15
|
|
|
use Kaloa\Xmp\Document as XmpDocument; |
16
|
|
|
use Kaloa\Xmp\ReaderException; |
17
|
|
|
|
18
|
|
|
/** |
19
|
|
|
* Extracts an XMP document from a data stream. |
20
|
|
|
* |
21
|
|
|
* The current algorithm ignores specific features and requirements of file |
22
|
|
|
* formats. It simply looks for the first occurrences of $tokenStart and |
23
|
|
|
* $tokenEnd and returns the content in between. This is a flexible approach but |
24
|
|
|
* it is not a correct one. There are cases in which the algorithm won't |
25
|
|
|
* succeed. Both false positives and false negatives are possible. |
26
|
|
|
*/ |
27
|
|
|
class Reader |
28
|
|
|
{ |
29
|
|
|
/** |
30
|
|
|
* Start token of XMP data. |
31
|
|
|
* |
32
|
|
|
* @var string |
33
|
|
|
*/ |
34
|
|
|
private $tokenStart = '<x:xmpmeta'; |
35
|
|
|
|
36
|
|
|
/** |
37
|
|
|
* End token of XMP data. |
38
|
|
|
* |
39
|
|
|
* @var string |
40
|
|
|
*/ |
41
|
|
|
private $tokenEnd = '</x:xmpmeta>'; |
42
|
|
|
|
43
|
|
|
/** |
44
|
|
|
* Size (in bytes) of data chunks read from the stream. |
45
|
|
|
* |
46
|
|
|
* @var int |
47
|
|
|
*/ |
48
|
|
|
private $chunkSize = 1024; |
49
|
|
|
|
50
|
|
|
/** |
51
|
|
|
* Buffer to construct XMP data in. |
52
|
|
|
* |
53
|
|
|
* @var string |
54
|
|
|
*/ |
55
|
|
|
private $buffer; |
56
|
|
|
|
57
|
|
|
/** |
58
|
|
|
* True if $tokenStart has been found. |
59
|
|
|
* |
60
|
|
|
* @var bool |
61
|
|
|
*/ |
62
|
|
|
private $started; |
63
|
|
|
|
64
|
|
|
/** |
65
|
|
|
* True if $started and $tokenEnd has been found. |
66
|
|
|
* |
67
|
|
|
* @var bool |
68
|
|
|
*/ |
69
|
|
|
private $ended; |
70
|
|
|
|
71
|
|
|
/** |
72
|
|
|
* Counts how many characters of the token that is currently searched for |
73
|
|
|
* have been found. |
74
|
|
|
* |
75
|
|
|
* This is reset whenever a character that doesn't equal the next one from |
76
|
|
|
* the token is found. If $delimPos reaches token length, the token has been |
77
|
|
|
* found. |
78
|
|
|
* |
79
|
|
|
* This variable is needed because a token might be split over two chunks of |
80
|
|
|
* input data so that functions such as strpos aren't sufficient. |
81
|
|
|
* |
82
|
|
|
* @var int |
83
|
|
|
*/ |
84
|
|
|
private $delimPos; |
85
|
|
|
|
86
|
|
|
/** |
87
|
|
|
* Length (in byte) of $tokenStart. |
88
|
|
|
* |
89
|
|
|
* @var int |
90
|
|
|
*/ |
91
|
|
|
private $tokenStartLen; |
92
|
|
|
|
93
|
|
|
/** |
94
|
|
|
* Length (in byte) of $tokenEnd. |
95
|
|
|
* |
96
|
|
|
* @var int |
97
|
|
|
*/ |
98
|
|
|
private $tokenEndLen; |
99
|
|
|
|
100
|
|
|
/** |
101
|
|
|
* Initializes the instance. |
102
|
|
|
*/ |
103
|
4 |
|
public function __construct() |
104
|
|
|
{ |
105
|
4 |
|
$this->reset(); |
106
|
4 |
|
} |
107
|
|
|
|
108
|
|
|
/** |
109
|
|
|
* Resets instance data to clean starting state. |
110
|
|
|
*/ |
111
|
4 |
|
private function reset() |
112
|
|
|
{ |
113
|
4 |
|
$this->buffer = ''; |
114
|
4 |
|
$this->started = false; |
115
|
4 |
|
$this->ended = false; |
116
|
4 |
|
$this->tokenStartLen = strlen($this->tokenStart); |
117
|
4 |
|
$this->tokenEndLen = strlen($this->tokenEnd); |
118
|
4 |
|
$this->delimPos = 0; |
119
|
4 |
|
} |
120
|
|
|
|
121
|
|
|
/** |
122
|
|
|
* Searches incoming data for $tokenStart adapting internal state if found. |
123
|
|
|
* |
124
|
|
|
* @param string $char A single byte |
125
|
|
|
*/ |
126
|
3 |
|
private function searchForTokenStart($char) |
127
|
|
|
{ |
128
|
3 |
|
if ($char === $this->tokenStart[$this->delimPos]) { |
129
|
3 |
|
$this->delimPos++; |
130
|
3 |
|
if ($this->delimPos === $this->tokenStartLen) { |
131
|
3 |
|
$this->delimPos = 0; |
132
|
3 |
|
$this->started = true; |
133
|
3 |
|
} |
134
|
3 |
|
} elseif ($char === $this->tokenStart[0]) { |
135
|
|
|
$this->delimPos = 1; |
136
|
|
|
} else { |
137
|
3 |
|
$this->delimPos = 0; |
138
|
|
|
} |
139
|
3 |
|
} |
140
|
|
|
|
141
|
|
|
/** |
142
|
|
|
* Searches incoming data for $tokenEnd adapting internal state if found. |
143
|
|
|
* |
144
|
|
|
* @param string $char A single byte |
145
|
|
|
*/ |
146
|
3 |
|
private function searchForTokenEnd($char) |
147
|
|
|
{ |
148
|
3 |
|
$this->buffer .= $char; |
149
|
3 |
|
if ($char === $this->tokenEnd[$this->delimPos]) { |
150
|
3 |
|
$this->delimPos++; |
151
|
3 |
|
if ($this->delimPos === $this->tokenEndLen) { |
152
|
2 |
|
$this->ended = true; |
153
|
2 |
|
} |
154
|
3 |
|
} elseif ($char === $this->tokenEnd[0]) { |
155
|
|
|
$this->delimPos = 1; |
156
|
|
|
} else { |
157
|
3 |
|
$this->delimPos = 0; |
158
|
|
|
} |
159
|
3 |
|
} |
160
|
|
|
|
161
|
|
|
/** |
162
|
|
|
* Extracts the first found XMP document from the stream. |
163
|
|
|
* |
164
|
|
|
* The stream is read in chunks and processed byte by byte in an |
165
|
|
|
* automaton-like fashion. |
166
|
|
|
* |
167
|
|
|
* After the execution of this method, instance variables $buffer, $started, |
168
|
|
|
* and $ended will contain meaningful values. |
169
|
|
|
* |
170
|
|
|
* @param resource $stream A stream resource |
171
|
|
|
*/ |
172
|
3 |
|
private function getXmpData($stream) |
173
|
|
|
{ |
174
|
3 |
|
while (!feof($stream)) { |
175
|
3 |
|
$chunk = fread($stream, $this->chunkSize); |
176
|
|
|
|
177
|
3 |
|
foreach (str_split($chunk) as $char) { |
178
|
3 |
|
if (!$this->started) { |
179
|
3 |
|
$this->searchForTokenStart($char); |
180
|
3 |
|
} else { |
181
|
3 |
|
$this->searchForTokenEnd($char); |
182
|
|
|
} |
183
|
|
|
|
184
|
3 |
|
if ($this->ended) { |
185
|
2 |
|
break 2; |
186
|
|
|
} |
187
|
3 |
|
} |
188
|
3 |
|
} |
189
|
|
|
|
190
|
3 |
|
if ($this->started && $this->ended) { |
191
|
2 |
|
$this->buffer = $this->tokenStart . $this->buffer; |
192
|
2 |
|
} else { |
193
|
1 |
|
$this->buffer = ''; |
194
|
|
|
} |
195
|
3 |
|
} |
196
|
|
|
|
197
|
|
|
/** |
198
|
|
|
* Returns a Kaloa\Xmp\Document of the first occurrence of XMP data in the |
199
|
|
|
* stream. |
200
|
|
|
* |
201
|
|
|
* @todo The method of error handling (set_error_handler) is just insane. |
202
|
|
|
* |
203
|
|
|
* @param resource $stream A stream resource |
204
|
|
|
* @return XmpDocument |
205
|
|
|
* @throws ReaderException |
206
|
|
|
*/ |
207
|
4 |
|
public function getXmpDocument($stream) |
208
|
|
|
{ |
209
|
4 |
|
if (!is_resource($stream) || get_resource_type($stream) !== 'stream') { |
210
|
1 |
|
throw new ReaderException('$stream is not a valid stream resource'); |
211
|
|
|
} |
212
|
|
|
|
213
|
3 |
|
$this->getXmpData($stream); |
214
|
|
|
|
215
|
3 |
|
if ($this->buffer === '') { |
216
|
1 |
|
$this->reset(); |
217
|
1 |
|
throw new ReaderException('No XMP document found in stream'); |
218
|
|
|
} |
219
|
|
|
|
220
|
|
|
|
221
|
2 |
|
set_error_handler(function($errno, $errstr, $errfile, $errline) { |
222
|
1 |
|
throw new ErrorException($errstr, $errno, 0, $errfile, $errline); |
223
|
2 |
|
}); |
224
|
|
|
|
225
|
|
|
try { |
226
|
2 |
|
$dom = new DOMDocument(); |
227
|
2 |
|
$ret = $dom->loadXML($this->buffer); |
228
|
|
|
|
229
|
|
|
// Added to make testErroneousXmpDataThrowsException work with hhvm |
230
|
1 |
|
if (false === $ret) { |
231
|
|
|
throw new Exception('loadXML returned false.'); |
232
|
|
|
} |
233
|
2 |
|
} catch (Exception $e) { |
234
|
|
|
// Finally |
235
|
1 |
|
restore_error_handler(); |
236
|
1 |
|
$this->reset(); |
237
|
|
|
|
238
|
1 |
|
throw new ReaderException($e->getMessage()); |
239
|
|
|
} |
240
|
|
|
|
241
|
|
|
// Finally |
242
|
1 |
|
restore_error_handler(); |
243
|
1 |
|
$this->reset(); |
244
|
|
|
|
245
|
|
|
|
246
|
1 |
|
$xmpDoc = new XmpDocument($dom); |
247
|
|
|
|
248
|
1 |
|
return $xmpDoc; |
249
|
|
|
} |
250
|
|
|
} |
251
|
|
|
|