|
1
|
|
|
<?php |
|
2
|
|
|
/** |
|
3
|
|
|
* @file |
|
4
|
|
|
* |
|
5
|
|
|
* PseudoClass class. |
|
6
|
|
|
* |
|
7
|
|
|
* This is the first pass in an experiment to break PseudoClass handling |
|
8
|
|
|
* out of the normal traversal. Eventually, this should become a |
|
9
|
|
|
* top-level pluggable registry that will allow custom pseudoclasses. |
|
10
|
|
|
* For now, though, we just handle the core pseudoclasses. |
|
11
|
|
|
*/ |
|
12
|
|
|
|
|
13
|
|
|
namespace QueryPath\CSS\DOMTraverser; |
|
14
|
|
|
|
|
15
|
|
|
use \QueryPath\CSS\NotImplementedException; |
|
16
|
|
|
use \QueryPath\CSS\EventHandler; |
|
17
|
|
|
use QueryPath\CSS\ParseException; |
|
18
|
|
|
|
|
19
|
|
|
/** |
|
20
|
|
|
* The PseudoClass handler. |
|
21
|
|
|
* |
|
22
|
|
|
*/ |
|
23
|
|
|
class PseudoClass |
|
24
|
|
|
{ |
|
25
|
|
|
|
|
26
|
|
|
/** |
|
27
|
|
|
* Tests whether the given element matches the given pseudoclass. |
|
28
|
|
|
* |
|
29
|
|
|
* @param string $pseudoclass |
|
30
|
|
|
* The string name of the pseudoclass |
|
31
|
|
|
* @param resource $node |
|
32
|
|
|
* The DOMNode to be tested. |
|
33
|
|
|
* @param resource $scope |
|
34
|
|
|
* The DOMElement that is the active root for this node. |
|
35
|
|
|
* @param mixed $value |
|
36
|
|
|
* The optional value string provided with this class. This is |
|
37
|
|
|
* used, for example, in an+b psuedoclasses. |
|
38
|
|
|
* @return bool |
|
39
|
|
|
* @throws NotImplementedException |
|
40
|
|
|
* @throws \QueryPath\CSS\ParseException |
|
41
|
|
|
* @retval boolean |
|
42
|
|
|
* TRUE if the node matches, FALSE otherwise. |
|
43
|
|
|
*/ |
|
44
|
|
|
public function elementMatches($pseudoclass, $node, $scope, $value = NULL) |
|
45
|
|
|
{ |
|
46
|
|
|
$name = strtolower($pseudoclass); |
|
47
|
|
|
// Need to handle known pseudoclasses. |
|
48
|
|
|
switch ($name) { |
|
49
|
|
|
case 'current': |
|
50
|
|
|
case 'past': |
|
51
|
|
|
case 'future': |
|
52
|
|
|
case 'visited': |
|
53
|
|
|
case 'hover': |
|
54
|
|
|
case 'active': |
|
55
|
|
|
case 'focus': |
|
56
|
|
|
case 'animated': // Last 3 are from jQuery |
|
57
|
|
|
case 'visible': |
|
58
|
|
|
case 'hidden': |
|
59
|
|
|
// These require a UA, which we don't have. |
|
60
|
|
|
case 'valid': |
|
61
|
|
|
case 'invalid': |
|
62
|
|
|
case 'required': |
|
63
|
|
|
case 'optional': |
|
64
|
|
|
case 'read-only': |
|
65
|
|
|
case 'read-write': |
|
66
|
|
|
// Since we don't know how to validate elements, |
|
67
|
|
|
// we can't supply these. |
|
68
|
|
|
case 'dir': |
|
69
|
|
|
// FIXME: I don't know how to get directionality info. |
|
70
|
|
|
case 'nth-column': |
|
71
|
|
|
case 'nth-last-column': |
|
72
|
|
|
// We don't know what a column is in most documents. |
|
73
|
|
|
// FIXME: Can we do this for HTML? |
|
74
|
|
|
case 'target': |
|
75
|
|
|
// This requires a location URL, which we don't have. |
|
76
|
|
|
return false; |
|
77
|
|
|
case 'indeterminate': |
|
78
|
|
|
// Because sometimes screwing with people is fun. |
|
79
|
|
|
return (boolean)mt_rand(0, 1); |
|
80
|
|
|
case 'lang': |
|
81
|
|
|
// No value = exception. |
|
82
|
|
|
if (!isset($value)) { |
|
83
|
|
|
throw new NotImplementedException(':lang() requires a value.'); |
|
84
|
|
|
} |
|
85
|
|
|
|
|
86
|
|
|
return $this->lang($node, $value); |
|
87
|
|
|
case 'any-link': |
|
88
|
|
|
return Util::matchesAttribute($node, 'href') |
|
89
|
|
|
|| Util::matchesAttribute($node, 'src') |
|
90
|
|
|
|| Util::matchesAttribute($node, 'link'); |
|
91
|
|
|
case 'link': |
|
92
|
|
|
return Util::matchesAttribute($node, 'href'); |
|
93
|
|
|
case 'local-link': |
|
94
|
|
|
return $this->isLocalLink($node); |
|
95
|
|
|
case 'root': |
|
96
|
|
|
return $node->isSameNode($node->ownerDocument->documentElement); |
|
97
|
|
|
|
|
98
|
|
|
// CSS 4 declares the :scope pseudo-class, which describes what was |
|
99
|
|
|
// the :x-root QueryPath extension. |
|
100
|
|
|
case 'x-root': |
|
101
|
|
|
case 'x-reset': |
|
102
|
|
|
case 'scope': |
|
103
|
|
|
return $node->isSameNode($scope); |
|
104
|
|
|
// NON-STANDARD extensions for simple support of even and odd. These |
|
105
|
|
|
// are supported by jQuery, FF, and other user agents. |
|
106
|
|
|
case 'even': |
|
107
|
|
|
return $this->isNthChild($node, 'even'); |
|
108
|
|
|
case 'odd': |
|
109
|
|
|
return $this->isNthChild($node, 'odd'); |
|
110
|
|
|
case 'nth-child': |
|
111
|
|
|
return $this->isNthChild($node, $value); |
|
112
|
|
|
case 'nth-last-child': |
|
113
|
|
|
return $this->isNthChild($node, $value, true); |
|
114
|
|
|
case 'nth-of-type': |
|
115
|
|
|
return $this->isNthChild($node, $value, false, true); |
|
116
|
|
|
case 'nth-last-of-type': |
|
117
|
|
|
return $this->isNthChild($node, $value, true, true); |
|
118
|
|
|
case 'first-of-type': |
|
119
|
|
|
return $this->isFirstOfType($node); |
|
120
|
|
|
case 'last-of-type': |
|
121
|
|
|
return $this->isLastOfType($node); |
|
122
|
|
|
case 'only-of-type': |
|
123
|
|
|
return $this->isFirstOfType($node) && $this->isLastOfType($node); |
|
124
|
|
|
|
|
125
|
|
|
// Additional pseudo-classes defined in jQuery: |
|
126
|
|
|
case 'lt': |
|
127
|
|
|
// I'm treating this as "less than or equal to". |
|
128
|
|
|
$rule = sprintf('-n + %d', (int)$value); |
|
129
|
|
|
|
|
130
|
|
|
// $rule = '-n+15'; |
|
131
|
|
|
return $this->isNthChild($node, $rule); |
|
132
|
|
|
case 'gt': |
|
133
|
|
|
// I'm treating this as "greater than" |
|
134
|
|
|
// return $this->nodePositionFromEnd($node) > (int) $value; |
|
135
|
|
|
return $this->nodePositionFromStart($node) > (int)$value; |
|
136
|
|
|
case 'nth': |
|
137
|
|
|
case 'eq': |
|
138
|
|
|
$rule = (int)$value; |
|
139
|
|
|
|
|
140
|
|
|
return $this->isNthChild($node, $rule); |
|
141
|
|
|
case 'first': |
|
142
|
|
|
return $this->isNthChild($node, 1); |
|
143
|
|
|
case 'first-child': |
|
144
|
|
|
return $this->isFirst($node); |
|
145
|
|
|
case 'last': |
|
146
|
|
|
case 'last-child': |
|
147
|
|
|
return $this->isLast($node); |
|
148
|
|
|
case 'only-child': |
|
149
|
|
|
return $this->isFirst($node) && $this->isLast($node); |
|
150
|
|
|
case 'empty': |
|
151
|
|
|
return $this->isEmpty($node); |
|
152
|
|
|
case 'parent': |
|
153
|
|
|
return !$this->isEmpty($node); |
|
154
|
|
|
|
|
155
|
|
|
case 'enabled': |
|
156
|
|
|
case 'disabled': |
|
157
|
|
|
case 'checked': |
|
158
|
|
|
return Util::matchesAttribute($node, $name); |
|
159
|
|
|
case 'text': |
|
160
|
|
|
case 'radio': |
|
161
|
|
|
case 'checkbox': |
|
162
|
|
|
case 'file': |
|
163
|
|
|
case 'password': |
|
164
|
|
|
case 'submit': |
|
165
|
|
|
case 'image': |
|
166
|
|
|
case 'reset': |
|
167
|
|
|
case 'button': |
|
168
|
|
|
return Util::matchesAttribute($node, 'type', $name); |
|
169
|
|
|
|
|
170
|
|
|
case 'header': |
|
171
|
|
|
return $this->header($node); |
|
172
|
|
|
case 'has': |
|
173
|
|
|
case 'matches': |
|
174
|
|
|
return $this->has($node, $value); |
|
175
|
|
|
break; |
|
|
|
|
|
|
176
|
|
|
case 'not': |
|
177
|
|
|
if (empty($value)) { |
|
178
|
|
|
throw new ParseException(':not() requires a value.'); |
|
179
|
|
|
} |
|
180
|
|
|
|
|
181
|
|
|
return $this->isNot($node, $value); |
|
182
|
|
|
// Contains == text matches. |
|
183
|
|
|
// In QP 2.1, this was changed. |
|
184
|
|
|
case 'contains': |
|
185
|
|
|
return $this->contains($node, $value); |
|
186
|
|
|
// Since QP 2.1 |
|
187
|
|
|
case 'contains-exactly': |
|
188
|
|
|
return $this->containsExactly($node, $value); |
|
189
|
|
|
default: |
|
190
|
|
|
throw new ParseException('Unknown Pseudo-Class: ' . $name); |
|
191
|
|
|
} |
|
192
|
|
|
} |
|
193
|
|
|
|
|
194
|
|
|
/** |
|
195
|
|
|
* Pseudo-class handler for :lang |
|
196
|
|
|
* |
|
197
|
|
|
* Note that this does not implement the spec in its entirety because we do |
|
198
|
|
|
* not presume to "know the language" of the document. If anyone is interested |
|
199
|
|
|
* in making this more intelligent, please do so. |
|
200
|
|
|
*/ |
|
201
|
|
|
protected function lang($node, $value) |
|
202
|
|
|
{ |
|
203
|
|
|
// TODO: This checks for cases where an explicit language is |
|
204
|
|
|
// set. The spec seems to indicate that an element should inherit |
|
205
|
|
|
// language from the parent... but this is unclear. |
|
206
|
|
|
$operator = (strpos($value, '-') !== false) ? EventHandler::IS_EXACTLY : EventHandler::CONTAINS_WITH_HYPHEN; |
|
207
|
|
|
|
|
208
|
|
|
$match = true; |
|
|
|
|
|
|
209
|
|
|
foreach ($node->attributes as $attrNode) { |
|
210
|
|
|
if ($attrNode->localName === 'lang') { |
|
211
|
|
|
|
|
212
|
|
|
if ($attrNode->nodeName === $attrNode->localName) { |
|
213
|
|
|
// fprintf(STDOUT, "%s in NS %s\n", $attrNode->name, $attrNode->nodeName); |
|
214
|
|
|
return Util::matchesAttribute($node, 'lang', $value, $operator); |
|
215
|
|
|
} |
|
216
|
|
|
|
|
217
|
|
|
$nsuri = $attrNode->namespaceURI; |
|
218
|
|
|
// fprintf(STDOUT, "%s in NS %s\n", $attrNode->name, $nsuri); |
|
219
|
|
|
return Util::matchesAttributeNS($node, 'lang', $nsuri, $value, $operator); |
|
220
|
|
|
} |
|
221
|
|
|
} |
|
222
|
|
|
|
|
223
|
|
|
return false; |
|
224
|
|
|
} |
|
225
|
|
|
|
|
226
|
|
|
/** |
|
227
|
|
|
* Provides jQuery pseudoclass ':header'. |
|
228
|
|
|
* |
|
229
|
|
|
* @param $node |
|
230
|
|
|
* @return bool |
|
231
|
|
|
*/ |
|
232
|
|
|
protected function header($node): bool |
|
233
|
|
|
{ |
|
234
|
|
|
return preg_match('/^h[1-9]$/i', $node->tagName) === 1; |
|
235
|
|
|
} |
|
236
|
|
|
|
|
237
|
|
|
/** |
|
238
|
|
|
* Provides pseudoclass :empty. |
|
239
|
|
|
*/ |
|
240
|
|
|
protected function isEmpty($node): bool |
|
241
|
|
|
{ |
|
242
|
|
|
foreach ($node->childNodes as $kid) { |
|
243
|
|
|
// We don't want to count PIs and comments. From the spec, it |
|
244
|
|
|
// appears that CDATA is also not counted. |
|
245
|
|
|
if ($kid->nodeType === XML_ELEMENT_NODE || $kid->nodeType === XML_TEXT_NODE) { |
|
246
|
|
|
// As soon as we hit a FALSE, return. |
|
247
|
|
|
return false; |
|
248
|
|
|
} |
|
249
|
|
|
} |
|
250
|
|
|
|
|
251
|
|
|
return true; |
|
252
|
|
|
} |
|
253
|
|
|
|
|
254
|
|
|
/** |
|
255
|
|
|
* Provides jQuery pseudoclass :first. |
|
256
|
|
|
* |
|
257
|
|
|
* @todo |
|
258
|
|
|
* This can be replaced by isNthChild(). |
|
259
|
|
|
*/ |
|
260
|
|
|
protected function isFirst($node): bool |
|
261
|
|
|
{ |
|
262
|
|
|
while (isset($node->previousSibling)) { |
|
263
|
|
|
$node = $node->previousSibling; |
|
264
|
|
|
if ($node->nodeType === XML_ELEMENT_NODE) { |
|
265
|
|
|
return false; |
|
266
|
|
|
} |
|
267
|
|
|
} |
|
268
|
|
|
|
|
269
|
|
|
return true; |
|
270
|
|
|
} |
|
271
|
|
|
|
|
272
|
|
|
/** |
|
273
|
|
|
* Fast version of first-of-type. |
|
274
|
|
|
*/ |
|
275
|
|
|
protected function isFirstOfType($node) |
|
276
|
|
|
{ |
|
277
|
|
|
$type = $node->tagName; |
|
278
|
|
|
while (isset($node->previousSibling)) { |
|
279
|
|
|
$node = $node->previousSibling; |
|
280
|
|
|
if ($node->nodeType === XML_ELEMENT_NODE && $node->tagName === $type) { |
|
281
|
|
|
return false; |
|
282
|
|
|
} |
|
283
|
|
|
} |
|
284
|
|
|
|
|
285
|
|
|
return true; |
|
286
|
|
|
} |
|
287
|
|
|
|
|
288
|
|
|
/** |
|
289
|
|
|
* Fast version of jQuery :last. |
|
290
|
|
|
*/ |
|
291
|
|
|
protected function isLast($node) |
|
292
|
|
|
{ |
|
293
|
|
|
while (isset($node->nextSibling)) { |
|
294
|
|
|
$node = $node->nextSibling; |
|
295
|
|
|
if ($node->nodeType === XML_ELEMENT_NODE) { |
|
296
|
|
|
return false; |
|
297
|
|
|
} |
|
298
|
|
|
} |
|
299
|
|
|
|
|
300
|
|
|
return true; |
|
301
|
|
|
} |
|
302
|
|
|
|
|
303
|
|
|
/** |
|
304
|
|
|
* Provides last-of-type. |
|
305
|
|
|
*/ |
|
306
|
|
|
protected function isLastOfType($node) |
|
307
|
|
|
{ |
|
308
|
|
|
$type = $node->tagName; |
|
309
|
|
|
while (isset($node->nextSibling)) { |
|
310
|
|
|
$node = $node->nextSibling; |
|
311
|
|
|
if ($node->nodeType === XML_ELEMENT_NODE && $node->tagName === $type) { |
|
312
|
|
|
return false; |
|
313
|
|
|
} |
|
314
|
|
|
} |
|
315
|
|
|
|
|
316
|
|
|
return true; |
|
317
|
|
|
} |
|
318
|
|
|
|
|
319
|
|
|
/** |
|
320
|
|
|
* Provides :contains() as the original spec called for. |
|
321
|
|
|
* |
|
322
|
|
|
* This is an INEXACT match. |
|
323
|
|
|
*/ |
|
324
|
|
|
protected function contains($node, $value): bool |
|
325
|
|
|
{ |
|
326
|
|
|
$text = $node->textContent; |
|
327
|
|
|
$value = Util::removeQuotes($value); |
|
328
|
|
|
|
|
329
|
|
|
return isset($text) && (stripos($text, $value) !== false); |
|
330
|
|
|
} |
|
331
|
|
|
|
|
332
|
|
|
/** |
|
333
|
|
|
* Provides :contains-exactly QueryPath pseudoclass. |
|
334
|
|
|
* |
|
335
|
|
|
* This is an EXACT match. |
|
336
|
|
|
*/ |
|
337
|
|
|
protected function containsExactly($node, $value): bool |
|
338
|
|
|
{ |
|
339
|
|
|
$text = $node->textContent; |
|
340
|
|
|
$value = Util::removeQuotes($value); |
|
341
|
|
|
|
|
342
|
|
|
return isset($text) && $text == $value; |
|
343
|
|
|
} |
|
344
|
|
|
|
|
345
|
|
|
/** |
|
346
|
|
|
* Provides :has pseudoclass. |
|
347
|
|
|
* |
|
348
|
|
|
* @throws ParseException |
|
349
|
|
|
*/ |
|
350
|
|
|
protected function has($node, $selector): bool |
|
351
|
|
|
{ |
|
352
|
|
|
$splos = new \SPLObjectStorage(); |
|
353
|
|
|
$splos->attach($node); |
|
354
|
|
|
$traverser = new \QueryPath\CSS\DOMTraverser($splos, true); |
|
355
|
|
|
$results = $traverser->find($selector)->matches(); |
|
356
|
|
|
|
|
357
|
|
|
return count($results) > 0; |
|
358
|
|
|
} |
|
359
|
|
|
|
|
360
|
|
|
/** |
|
361
|
|
|
* Provides :not pseudoclass. |
|
362
|
|
|
* |
|
363
|
|
|
* @throws ParseException |
|
364
|
|
|
*/ |
|
365
|
|
|
protected function isNot($node, $selector): bool |
|
366
|
|
|
{ |
|
367
|
|
|
return !$this->has($node, $selector); |
|
368
|
|
|
} |
|
369
|
|
|
|
|
370
|
|
|
/** |
|
371
|
|
|
* Get the relative position of a node in its sibling set. |
|
372
|
|
|
*/ |
|
373
|
|
|
protected function nodePositionFromStart($node, $byType = false): int |
|
374
|
|
|
{ |
|
375
|
|
|
$i = 1; |
|
376
|
|
|
$tag = $node->tagName; |
|
377
|
|
|
while (isset($node->previousSibling)) { |
|
378
|
|
|
$node = $node->previousSibling; |
|
379
|
|
|
if ($node->nodeType === XML_ELEMENT_NODE && (!$byType || $node->tagName === $tag)) { |
|
380
|
|
|
++$i; |
|
381
|
|
|
} |
|
382
|
|
|
} |
|
383
|
|
|
|
|
384
|
|
|
return $i; |
|
385
|
|
|
} |
|
386
|
|
|
|
|
387
|
|
|
/** |
|
388
|
|
|
* Get the relative position of a node in its sibling set. |
|
389
|
|
|
* |
|
390
|
|
|
* @param $node |
|
391
|
|
|
* @param bool $byType |
|
392
|
|
|
* @return int |
|
393
|
|
|
*/ |
|
394
|
|
|
protected function nodePositionFromEnd($node, $byType = false): int |
|
395
|
|
|
{ |
|
396
|
|
|
$i = 1; |
|
397
|
|
|
$tag = $node->tagName; |
|
398
|
|
|
while (isset($node->nextSibling)) { |
|
399
|
|
|
$node = $node->nextSibling; |
|
400
|
|
|
if ($node->nodeType === XML_ELEMENT_NODE && (!$byType || $node->tagName === $tag)) { |
|
401
|
|
|
++$i; |
|
402
|
|
|
} |
|
403
|
|
|
} |
|
404
|
|
|
|
|
405
|
|
|
return $i; |
|
406
|
|
|
} |
|
407
|
|
|
|
|
408
|
|
|
/** |
|
409
|
|
|
* Provides functionality for all "An+B" rules. |
|
410
|
|
|
* Provides nth-child and also the functionality required for: |
|
411
|
|
|
* |
|
412
|
|
|
*- nth-last-child |
|
413
|
|
|
*- even |
|
414
|
|
|
*- odd |
|
415
|
|
|
*- first |
|
416
|
|
|
*- last |
|
417
|
|
|
*- eq |
|
418
|
|
|
*- nth |
|
419
|
|
|
*- nth-of-type |
|
420
|
|
|
*- first-of-type |
|
421
|
|
|
*- last-of-type |
|
422
|
|
|
*- nth-last-of-type |
|
423
|
|
|
* |
|
424
|
|
|
* See also QueryPath::CSS::DOMTraverser::Util::parseAnB(). |
|
425
|
|
|
* |
|
426
|
|
|
* @param $node |
|
427
|
|
|
* @param $value |
|
428
|
|
|
* @param bool $reverse |
|
429
|
|
|
* @param bool $byType |
|
430
|
|
|
* @return bool |
|
431
|
|
|
*/ |
|
432
|
|
|
protected function isNthChild($node, $value, $reverse = false, $byType = false): bool |
|
433
|
|
|
{ |
|
434
|
|
|
list($groupSize, $elementInGroup) = Util::parseAnB($value); |
|
435
|
|
|
$parent = $node->parentNode; |
|
436
|
|
|
if (empty($parent) |
|
437
|
|
|
|| ($groupSize === 0 && $elementInGroup === 0) |
|
438
|
|
|
|| ($groupSize > 0 && $elementInGroup > $groupSize) |
|
439
|
|
|
) { |
|
440
|
|
|
return false; |
|
441
|
|
|
} |
|
442
|
|
|
|
|
443
|
|
|
// First we need to find the position of $node in other elements. |
|
444
|
|
|
if ($reverse) { |
|
445
|
|
|
$pos = $this->nodePositionFromEnd($node, $byType); |
|
446
|
|
|
} else { |
|
447
|
|
|
$pos = $this->nodePositionFromStart($node, $byType); |
|
448
|
|
|
} |
|
449
|
|
|
|
|
450
|
|
|
// If group size is 0, we just check to see if this |
|
451
|
|
|
// is the nth element: |
|
452
|
|
|
if ($groupSize === 0) { |
|
453
|
|
|
return $pos === $elementInGroup; |
|
454
|
|
|
} |
|
455
|
|
|
|
|
456
|
|
|
// Next, we normalize $elementInGroup |
|
457
|
|
|
if ($elementInGroup < 0) { |
|
458
|
|
|
$elementInGroup = $groupSize + $elementInGroup; |
|
459
|
|
|
} |
|
460
|
|
|
$prod = ($pos - $elementInGroup) / $groupSize; |
|
461
|
|
|
|
|
462
|
|
|
return is_int($prod) && $prod >= 0; |
|
463
|
|
|
} |
|
464
|
|
|
|
|
465
|
|
|
protected function isLocalLink($node): bool |
|
466
|
|
|
{ |
|
467
|
|
|
if (!$node->hasAttribute('href')) { |
|
468
|
|
|
return false; |
|
469
|
|
|
} |
|
470
|
|
|
$url = $node->getAttribute('href'); |
|
471
|
|
|
$scheme = parse_url($url, PHP_URL_SCHEME); |
|
472
|
|
|
|
|
473
|
|
|
return empty($scheme) || $scheme === 'file'; |
|
474
|
|
|
} |
|
475
|
|
|
|
|
476
|
|
|
} |
|
477
|
|
|
|
The
breakstatement is not necessary if it is preceded for example by areturnstatement:If you would like to keep this construct to be consistent with other
casestatements, you can safely mark this issue as a false-positive.