|
1
|
|
|
<?php |
|
2
|
|
|
/** |
|
3
|
|
|
* An implementation of the tree building portion of the HTML5 parsing |
|
4
|
|
|
* spec. |
|
5
|
|
|
* |
|
6
|
|
|
* This program is free software; you can redistribute it and/or modify |
|
7
|
|
|
* it under the terms of the GNU General Public License as published by |
|
8
|
|
|
* the Free Software Foundation; either version 2 of the License, or |
|
9
|
|
|
* (at your option) any later version. |
|
10
|
|
|
* |
|
11
|
|
|
* This program is distributed in the hope that it will be useful, |
|
12
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
13
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
14
|
|
|
* GNU General Public License for more details. |
|
15
|
|
|
* |
|
16
|
|
|
* You should have received a copy of the GNU General Public License along |
|
17
|
|
|
* with this program; if not, write to the Free Software Foundation, Inc., |
|
18
|
|
|
* 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. |
|
19
|
|
|
* http://www.gnu.org/copyleft/gpl.html |
|
20
|
|
|
* |
|
21
|
|
|
* @file |
|
22
|
|
|
* @ingroup Parser |
|
23
|
|
|
* @since 1.27 |
|
24
|
|
|
* @author C. Scott Ananian, 2016 |
|
25
|
|
|
*/ |
|
26
|
|
|
namespace MediaWiki\Tidy; |
|
27
|
|
|
|
|
28
|
|
|
use Wikimedia\Assert\Assert; |
|
29
|
|
|
use Wikimedia\Assert\ParameterAssertionException; |
|
30
|
|
|
use \ExplodeIterator; |
|
31
|
|
|
use \IteratorAggregate; |
|
32
|
|
|
use \ReverseArrayIterator; |
|
33
|
|
|
use \Sanitizer; |
|
34
|
|
|
|
|
35
|
|
|
# A note for future librarization[1] -- this file is a good candidate |
|
36
|
|
|
# for splitting into an independent library, except that it is currently |
|
37
|
|
|
# highly optimized for MediaWiki use. It only implements the portions |
|
38
|
|
|
# of the HTML5 tree builder used by tags supported by MediaWiki, and |
|
39
|
|
|
# does not contain a true tokenizer pass, instead relying on |
|
40
|
|
|
# comment stripping, attribute normalization, and escaping done by |
|
41
|
|
|
# the MediaWiki Sanitizer. It also deliberately avoids building |
|
42
|
|
|
# a true DOM in memory, instead serializing elements to an output string |
|
43
|
|
|
# as soon as possible (usually as soon as the tag is closed) to reduce |
|
44
|
|
|
# its memory footprint. |
|
45
|
|
|
|
|
46
|
|
|
# On the other hand, I've been pretty careful to note with comments in the |
|
47
|
|
|
# code the places where this implementation omits features of the spec or |
|
48
|
|
|
# depends on the MediaWiki Sanitizer. Perhaps in the future we'll want to |
|
49
|
|
|
# implement the missing pieces and make this a standalone PHP HTML5 parser. |
|
50
|
|
|
# In order to do so, some sort of MediaWiki-specific API will need |
|
51
|
|
|
# to be added to (a) allow the Balancer to bypass the tokenizer, |
|
52
|
|
|
# and (b) support on-the-fly flattening instead of DOM node creation. |
|
53
|
|
|
|
|
54
|
|
|
# [1]: https://www.mediawiki.org/wiki/Library_infrastructure_for_MediaWiki |
|
55
|
|
|
|
|
56
|
|
|
/** |
|
57
|
|
|
* Utility constants and sets for the HTML5 tree building algorithm. |
|
58
|
|
|
* Sets are associative arrays indexed first by namespace and then by |
|
59
|
|
|
* lower-cased tag name. |
|
60
|
|
|
* |
|
61
|
|
|
* @ingroup Parser |
|
62
|
|
|
* @since 1.27 |
|
63
|
|
|
*/ |
|
64
|
|
|
class BalanceSets { |
|
65
|
|
|
const HTML_NAMESPACE = 'http://www.w3.org/1999/xhtml'; |
|
66
|
|
|
const MATHML_NAMESPACE = 'http://www.w3.org/1998/Math/MathML'; |
|
67
|
|
|
const SVG_NAMESPACE = 'http://www.w3.org/2000/svg'; |
|
68
|
|
|
|
|
69
|
|
|
public static $unsupportedSet = [ |
|
70
|
|
|
self::HTML_NAMESPACE => [ |
|
71
|
|
|
'html' => true, 'head' => true, 'body' => true, 'frameset' => true, |
|
72
|
|
|
'form' => true, 'frame' => true, |
|
73
|
|
|
'plaintext' => true, 'isindex' => true, 'textarea' => true, |
|
74
|
|
|
'xmp' => true, 'iframe' => true, 'noembed' => true, |
|
75
|
|
|
'noscript' => true, 'select' => true, 'script' => true, |
|
76
|
|
|
'title' => true |
|
77
|
|
|
] |
|
78
|
|
|
]; |
|
79
|
|
|
|
|
80
|
|
|
public static $emptyElementSet = [ |
|
81
|
|
|
self::HTML_NAMESPACE => [ |
|
82
|
|
|
'area' => true, 'base' => true, 'basefont' => true, |
|
83
|
|
|
'bgsound' => true, 'br' => true, 'col' => true, 'command' => true, |
|
84
|
|
|
'embed' => true, 'frame' => true, 'hr' => true, 'img' => true, |
|
85
|
|
|
'input' => true, 'keygen' => true, 'link' => true, 'meta' => true, |
|
86
|
|
|
'param' => true, 'source' => true, 'track' => true, 'wbr' => true |
|
87
|
|
|
] |
|
88
|
|
|
]; |
|
89
|
|
|
|
|
90
|
|
|
public static $headingSet = [ |
|
91
|
|
|
self::HTML_NAMESPACE => [ |
|
92
|
|
|
'h1' => true, 'h2' => true, 'h3' => true, |
|
93
|
|
|
'h4' => true, 'h5' => true, 'h6' => true |
|
94
|
|
|
] |
|
95
|
|
|
]; |
|
96
|
|
|
|
|
97
|
|
|
public static $specialSet = [ |
|
98
|
|
|
self::HTML_NAMESPACE => [ |
|
99
|
|
|
'address' => true, 'applet' => true, 'area' => true, |
|
100
|
|
|
'article' => true, 'aside' => true, 'base' => true, |
|
101
|
|
|
'basefont' => true, 'bgsound' => true, 'blockquote' => true, |
|
102
|
|
|
'body' => true, 'br' => true, 'button' => true, 'caption' => true, |
|
103
|
|
|
'center' => true, 'col' => true, 'colgroup' => true, 'dd' => true, |
|
104
|
|
|
'details' => true, 'dir' => true, 'div' => true, 'dl' => true, |
|
105
|
|
|
'dt' => true, 'embed' => true, 'fieldset' => true, |
|
106
|
|
|
'figcaption' => true, 'figure' => true, 'footer' => true, |
|
107
|
|
|
'form' => true, 'frame' => true, 'frameset' => true, 'h1' => true, |
|
108
|
|
|
'h2' => true, 'h3' => true, 'h4' => true, 'h5' => true, |
|
109
|
|
|
'h6' => true, 'head' => true, 'header' => true, 'hgroup' => true, |
|
110
|
|
|
'hr' => true, 'html' => true, 'iframe' => true, 'img' => true, |
|
111
|
|
|
'input' => true, 'isindex' => true, 'li' => true, 'link' => true, |
|
112
|
|
|
'listing' => true, 'main' => true, 'marquee' => true, |
|
113
|
|
|
'menu' => true, 'menuitem' => true, 'meta' => true, 'nav' => true, |
|
114
|
|
|
'noembed' => true, 'noframes' => true, 'noscript' => true, |
|
115
|
|
|
'object' => true, 'ol' => true, 'p' => true, 'param' => true, |
|
116
|
|
|
'plaintext' => true, 'pre' => true, 'script' => true, |
|
117
|
|
|
'section' => true, 'select' => true, 'source' => true, |
|
118
|
|
|
'style' => true, 'summary' => true, 'table' => true, |
|
119
|
|
|
'tbody' => true, 'td' => true, 'template' => true, |
|
120
|
|
|
'textarea' => true, 'tfoot' => true, 'th' => true, 'thead' => true, |
|
121
|
|
|
'title' => true, 'tr' => true, 'track' => true, 'ul' => true, |
|
122
|
|
|
'wbr' => true, 'xmp' => true |
|
123
|
|
|
], |
|
124
|
|
|
self::SVG_NAMESPACE => [ |
|
125
|
|
|
'foreignobject' => true, 'desc' => true, 'title' => true |
|
126
|
|
|
], |
|
127
|
|
|
self::MATHML_NAMESPACE => [ |
|
128
|
|
|
'mi' => true, 'mo' => true, 'mn' => true, 'ms' => true, |
|
129
|
|
|
'mtext' => true, 'annotation-xml' => true |
|
130
|
|
|
] |
|
131
|
|
|
]; |
|
132
|
|
|
|
|
133
|
|
|
public static $addressDivPSet = [ |
|
134
|
|
|
self::HTML_NAMESPACE => [ |
|
135
|
|
|
'address' => true, 'div' => true, 'p' => true |
|
136
|
|
|
] |
|
137
|
|
|
]; |
|
138
|
|
|
|
|
139
|
|
|
public static $tableSectionRowSet = [ |
|
140
|
|
|
self::HTML_NAMESPACE => [ |
|
141
|
|
|
'table' => true, 'thead' => true, 'tbody' => true, |
|
142
|
|
|
'tfoot' => true, 'tr' => true |
|
143
|
|
|
] |
|
144
|
|
|
]; |
|
145
|
|
|
|
|
146
|
|
|
public static $impliedEndTagsSet = [ |
|
147
|
|
|
self::HTML_NAMESPACE => [ |
|
148
|
|
|
'dd' => true, 'dt' => true, 'li' => true, 'optgroup' => true, |
|
149
|
|
|
'option' => true, 'p' => true, 'rb' => true, 'rp' => true, |
|
150
|
|
|
'rt' => true, 'rtc' => true |
|
151
|
|
|
] |
|
152
|
|
|
]; |
|
153
|
|
|
|
|
154
|
|
|
public static $thoroughImpliedEndTagsSet = [ |
|
155
|
|
|
self::HTML_NAMESPACE => [ |
|
156
|
|
|
'caption' => true, 'colgroup' => true, 'dd' => true, 'dt' => true, |
|
157
|
|
|
'li' => true, 'optgroup' => true, 'option' => true, 'p' => true, |
|
158
|
|
|
'rb' => true, 'rp' => true, 'rt' => true, 'rtc' => true, |
|
159
|
|
|
'tbody' => true, 'td' => true, 'tfoot' => true, 'th' => true, |
|
160
|
|
|
'thead' => true, 'tr' => true |
|
161
|
|
|
] |
|
162
|
|
|
]; |
|
163
|
|
|
|
|
164
|
|
|
public static $tableCellSet = [ |
|
165
|
|
|
self::HTML_NAMESPACE => [ |
|
166
|
|
|
'td' => true, 'th' => true |
|
167
|
|
|
] |
|
168
|
|
|
]; |
|
169
|
|
|
public static $tableContextSet = [ |
|
170
|
|
|
self::HTML_NAMESPACE => [ |
|
171
|
|
|
'table' => true, 'template' => true, 'html' => true |
|
172
|
|
|
] |
|
173
|
|
|
]; |
|
174
|
|
|
|
|
175
|
|
|
public static $tableBodyContextSet = [ |
|
176
|
|
|
self::HTML_NAMESPACE => [ |
|
177
|
|
|
'tbody' => true, 'tfoot' => true, 'thead' => true, |
|
178
|
|
|
'template' => true, 'html' => true |
|
179
|
|
|
] |
|
180
|
|
|
]; |
|
181
|
|
|
|
|
182
|
|
|
public static $tableRowContextSet = [ |
|
183
|
|
|
self::HTML_NAMESPACE => [ |
|
184
|
|
|
'tr' => true, 'template' => true, 'html' => true |
|
185
|
|
|
] |
|
186
|
|
|
]; |
|
187
|
|
|
|
|
188
|
|
|
# OMITTED: formAssociatedSet, since we don't allow <form> |
|
189
|
|
|
|
|
190
|
|
|
public static $inScopeSet = [ |
|
191
|
|
|
self::HTML_NAMESPACE => [ |
|
192
|
|
|
'applet' => true, 'caption' => true, 'html' => true, |
|
193
|
|
|
'marquee' => true, 'object' => true, |
|
194
|
|
|
'table' => true, 'td' => true, 'template' => true, |
|
195
|
|
|
'th' => true |
|
196
|
|
|
], |
|
197
|
|
|
self::SVG_NAMESPACE => [ |
|
198
|
|
|
'foreignobject' => true, 'desc' => true, 'title' => true |
|
199
|
|
|
], |
|
200
|
|
|
self::MATHML_NAMESPACE => [ |
|
201
|
|
|
'mi' => true, 'mo' => true, 'mn' => true, 'ms' => true, |
|
202
|
|
|
'mtext' => true, 'annotation-xml' => true |
|
203
|
|
|
] |
|
204
|
|
|
]; |
|
205
|
|
|
|
|
206
|
|
|
private static $inListItemScopeSet = null; |
|
207
|
|
|
public static function inListItemScopeSet() { |
|
208
|
|
|
if ( self::$inListItemScopeSet === null ) { |
|
209
|
|
|
self::$inListItemScopeSet = self::$inScopeSet; |
|
210
|
|
|
self::$inListItemScopeSet[self::HTML_NAMESPACE]['ol'] = true; |
|
211
|
|
|
self::$inListItemScopeSet[self::HTML_NAMESPACE]['ul'] = true; |
|
212
|
|
|
} |
|
213
|
|
|
return self::$inListItemScopeSet; |
|
214
|
|
|
} |
|
215
|
|
|
|
|
216
|
|
|
private static $inButtonScopeSet = null; |
|
217
|
|
|
public static function inButtonScopeSet() { |
|
218
|
|
|
if ( self::$inButtonScopeSet === null ) { |
|
219
|
|
|
self::$inButtonScopeSet = self::$inScopeSet; |
|
220
|
|
|
self::$inButtonScopeSet[self::HTML_NAMESPACE]['button'] = true; |
|
221
|
|
|
} |
|
222
|
|
|
return self::$inButtonScopeSet; |
|
223
|
|
|
} |
|
224
|
|
|
|
|
225
|
|
|
public static $inTableScopeSet = [ |
|
226
|
|
|
self::HTML_NAMESPACE => [ |
|
227
|
|
|
'html' => true, 'table' => true, 'template' => true |
|
228
|
|
|
] |
|
229
|
|
|
]; |
|
230
|
|
|
|
|
231
|
|
|
public static $mathmlTextIntegrationPointSet = [ |
|
232
|
|
|
self::MATHML_NAMESPACE => [ |
|
233
|
|
|
'mi' => true, 'mo' => true, 'mn' => true, 'ms' => true, |
|
234
|
|
|
'mtext' => true |
|
235
|
|
|
] |
|
236
|
|
|
]; |
|
237
|
|
|
|
|
238
|
|
|
public static $htmlIntegrationPointSet = [ |
|
239
|
|
|
self::SVG_NAMESPACE => [ |
|
240
|
|
|
'foreignobject' => true, |
|
241
|
|
|
'desc' => true, |
|
242
|
|
|
'title' => true |
|
243
|
|
|
] |
|
244
|
|
|
]; |
|
245
|
|
|
|
|
246
|
|
|
// For tidy compatibility. |
|
247
|
|
|
public static $tidyPWrapSet = [ |
|
248
|
|
|
self::HTML_NAMESPACE => [ |
|
249
|
|
|
'body' => true, 'blockquote' => true, |
|
250
|
|
|
// We parse with <body> as the fragment context, but the top-level |
|
251
|
|
|
// element on the stack is actually <html>. We could use the |
|
252
|
|
|
// "adjusted current node" everywhere to work around this, but it's |
|
253
|
|
|
// easier just to add <html> to the p-wrap set. |
|
254
|
|
|
'html' => true, |
|
255
|
|
|
], |
|
256
|
|
|
]; |
|
257
|
|
|
public static $tidyInlineSet = [ |
|
258
|
|
|
self::HTML_NAMESPACE => [ |
|
259
|
|
|
'a' => true, 'abbr' => true, 'acronym' => true, 'applet' => true, |
|
260
|
|
|
'b' => true, 'basefont' => true, 'bdo' => true, 'big' => true, |
|
261
|
|
|
'br' => true, 'button' => true, 'cite' => true, 'code' => true, |
|
262
|
|
|
'dfn' => true, 'em' => true, 'font' => true, 'i' => true, |
|
263
|
|
|
'iframe' => true, 'img' => true, 'input' => true, 'kbd' => true, |
|
264
|
|
|
'label' => true, 'legend' => true, 'map' => true, 'object' => true, |
|
265
|
|
|
'param' => true, 'q' => true, 'rb' => true, 'rbc' => true, |
|
266
|
|
|
'rp' => true, 'rt' => true, 'rtc' => true, 'ruby' => true, |
|
267
|
|
|
's' => true, 'samp' => true, 'select' => true, 'small' => true, |
|
268
|
|
|
'span' => true, 'strike' => true, 'strong' => true, 'sub' => true, |
|
269
|
|
|
'sup' => true, 'textarea' => true, 'tt' => true, 'u' => true, |
|
270
|
|
|
'var' => true, |
|
271
|
|
|
], |
|
272
|
|
|
]; |
|
273
|
|
|
} |
|
274
|
|
|
|
|
275
|
|
|
/** |
|
276
|
|
|
* A BalanceElement is a simplified version of a DOM Node. The main |
|
277
|
|
|
* difference is that we only keep BalanceElements around for nodes |
|
278
|
|
|
* currently on the BalanceStack of open elements. As soon as an |
|
279
|
|
|
* element is closed, with some minor exceptions relating to the |
|
280
|
|
|
* tree builder "adoption agency algorithm", the element and all its |
|
281
|
|
|
* children are serialized to a string using the flatten() method. |
|
282
|
|
|
* This keeps our memory usage low. |
|
283
|
|
|
* |
|
284
|
|
|
* @ingroup Parser |
|
285
|
|
|
* @since 1.27 |
|
286
|
|
|
*/ |
|
287
|
|
|
class BalanceElement { |
|
288
|
|
|
/** |
|
289
|
|
|
* The namespace of the element. |
|
290
|
|
|
* @var string $namespaceURI |
|
291
|
|
|
*/ |
|
292
|
|
|
public $namespaceURI; |
|
293
|
|
|
/** |
|
294
|
|
|
* The lower-cased name of the element. |
|
295
|
|
|
* @var string $localName |
|
296
|
|
|
*/ |
|
297
|
|
|
public $localName; |
|
298
|
|
|
/** |
|
299
|
|
|
* Attributes for the element, in array form |
|
300
|
|
|
* @var array $attribs |
|
301
|
|
|
*/ |
|
302
|
|
|
public $attribs; |
|
303
|
|
|
|
|
304
|
|
|
/** |
|
305
|
|
|
* Parent of this element, or the string "flat" if this element has |
|
306
|
|
|
* already been flattened into its parent. |
|
307
|
|
|
* @var string|null $parent |
|
308
|
|
|
*/ |
|
309
|
|
|
public $parent; |
|
310
|
|
|
|
|
311
|
|
|
/** |
|
312
|
|
|
* An array of children of this element. Typically only the last |
|
313
|
|
|
* child will be an actual BalanceElement object; the rest will |
|
314
|
|
|
* be strings, representing either text nodes or flattened |
|
315
|
|
|
* BalanceElement objects. |
|
316
|
|
|
* @var array $children |
|
317
|
|
|
*/ |
|
318
|
|
|
public $children; |
|
319
|
|
|
|
|
320
|
|
|
/** |
|
321
|
|
|
* A unique string identifier for Noah's Ark purposes, lazy initialized |
|
322
|
|
|
*/ |
|
323
|
|
|
private $noahKey; |
|
324
|
|
|
|
|
325
|
|
|
/** |
|
326
|
|
|
* The next active formatting element in the list, or null if this is the |
|
327
|
|
|
* end of the AFE list or if the element is not in the AFE list. |
|
328
|
|
|
*/ |
|
329
|
|
|
public $nextAFE; |
|
330
|
|
|
|
|
331
|
|
|
/** |
|
332
|
|
|
* The previous active formatting element in the list, or null if this is |
|
333
|
|
|
* the start of the list or if the element is not in the AFE list. |
|
334
|
|
|
*/ |
|
335
|
|
|
public $prevAFE; |
|
336
|
|
|
|
|
337
|
|
|
/** |
|
338
|
|
|
* The next element in the Noah's Ark species bucket. |
|
339
|
|
|
*/ |
|
340
|
|
|
public $nextNoah; |
|
341
|
|
|
|
|
342
|
|
|
/** |
|
343
|
|
|
* Make a new BalanceElement corresponding to the HTML DOM Element |
|
344
|
|
|
* with the given localname, namespace, and attributes. |
|
345
|
|
|
* |
|
346
|
|
|
* @param string $namespaceURI The namespace of the element. |
|
347
|
|
|
* @param string $localName The lowercased name of the tag. |
|
348
|
|
|
* @param array $attribs Attributes of the element |
|
349
|
|
|
*/ |
|
350
|
|
|
public function __construct( $namespaceURI, $localName, array $attribs ) { |
|
351
|
|
|
Assert::parameterType( 'string', $namespaceURI, '$namespaceURI' ); |
|
352
|
|
|
Assert::parameterType( 'string', $localName, '$localName' ); |
|
353
|
|
|
|
|
354
|
|
|
$this->localName = $localName; |
|
355
|
|
|
$this->namespaceURI = $namespaceURI; |
|
356
|
|
|
$this->attribs = $attribs; |
|
357
|
|
|
$this->contents = ''; |
|
|
|
|
|
|
358
|
|
|
$this->parent = null; |
|
359
|
|
|
$this->children = []; |
|
360
|
|
|
} |
|
361
|
|
|
|
|
362
|
|
|
/** |
|
363
|
|
|
* Remove the given child from this element. |
|
364
|
|
|
* @param BalanceElement $elt |
|
365
|
|
|
*/ |
|
366
|
|
|
private function removeChild( $elt ) { |
|
|
|
|
|
|
367
|
|
|
Assert::precondition( |
|
368
|
|
|
$this->parent !== 'flat', "Can't removeChild after flattening $this" |
|
369
|
|
|
); |
|
370
|
|
|
Assert::parameterType( 'MediaWiki\Tidy\BalanceElement', $elt, '$elt' ); |
|
371
|
|
|
Assert::parameter( |
|
372
|
|
|
$elt->parent === $this, 'elt', 'must have $this as a parent' |
|
373
|
|
|
); |
|
374
|
|
|
$idx = array_search( $elt, $this->children, true ); |
|
375
|
|
|
Assert::parameter( $idx !== false, '$elt', 'must be a child of $this' ); |
|
376
|
|
|
$elt->parent = null; |
|
377
|
|
|
array_splice( $this->children, $idx, 1 ); |
|
378
|
|
|
} |
|
379
|
|
|
|
|
380
|
|
|
/** |
|
381
|
|
|
* Find $a in the list of children and insert $b before it. |
|
382
|
|
|
* @param BalanceElement $a |
|
383
|
|
|
* @param BalanceElement|string $b |
|
384
|
|
|
*/ |
|
385
|
|
|
public function insertBefore( $a, $b ) { |
|
386
|
|
|
Assert::precondition( |
|
387
|
|
|
$this->parent !== 'flat', "Can't insertBefore after flattening." |
|
388
|
|
|
); |
|
389
|
|
|
Assert::parameterType( 'MediaWiki\Tidy\BalanceElement', $a, '$a' ); |
|
390
|
|
|
Assert::parameterType( 'MediaWiki\Tidy\BalanceElement|string', $b, '$b' ); |
|
391
|
|
|
$idx = array_search( $a, $this->children, true ); |
|
392
|
|
|
Assert::parameter( $idx !== false, '$a', 'must be a child of $this' ); |
|
393
|
|
|
if ( is_string( $b ) ) { |
|
394
|
|
|
array_splice( $this->children, $idx, 0, [ $b ] ); |
|
395
|
|
|
} else { |
|
396
|
|
|
Assert::parameter( $b->parent !== 'flat', '$b', "Can't be flat" ); |
|
397
|
|
|
if ( $b->parent !== null ) { |
|
398
|
|
|
$b->parent->removeChild( $b ); |
|
|
|
|
|
|
399
|
|
|
} |
|
400
|
|
|
array_splice( $this->children, $idx, 0, [ $b ] ); |
|
401
|
|
|
$b->parent = $this; |
|
|
|
|
|
|
402
|
|
|
} |
|
403
|
|
|
} |
|
404
|
|
|
|
|
405
|
|
|
/** |
|
406
|
|
|
* Append $elt to the end of the list of children. |
|
407
|
|
|
* @param BalanceElement|string $elt |
|
408
|
|
|
*/ |
|
409
|
|
|
public function appendChild( $elt ) { |
|
410
|
|
|
Assert::precondition( |
|
411
|
|
|
$this->parent !== 'flat', "Can't appendChild after flattening." |
|
412
|
|
|
); |
|
413
|
|
|
Assert::parameterType( 'MediaWiki\Tidy\BalanceElement|string', $elt, '$elt' ); |
|
414
|
|
|
if ( is_string( $elt ) ) { |
|
415
|
|
|
array_push( $this->children, $elt ); |
|
416
|
|
|
return; |
|
417
|
|
|
} |
|
418
|
|
|
// Remove $elt from parent, if it had one. |
|
419
|
|
|
if ( $elt->parent !== null ) { |
|
420
|
|
|
$elt->parent->removeChild( $elt ); |
|
|
|
|
|
|
421
|
|
|
} |
|
422
|
|
|
array_push( $this->children, $elt ); |
|
423
|
|
|
$elt->parent = $this; |
|
|
|
|
|
|
424
|
|
|
} |
|
425
|
|
|
|
|
426
|
|
|
/** |
|
427
|
|
|
* Transfer all of the children of $elt to $this. |
|
428
|
|
|
* @param BalanceElement $elt |
|
429
|
|
|
*/ |
|
430
|
|
|
public function adoptChildren( $elt ) { |
|
431
|
|
|
Assert::parameterType( 'MediaWiki\Tidy\BalanceElement', $elt, '$elt' ); |
|
432
|
|
|
Assert::precondition( |
|
433
|
|
|
$elt->parent !== 'flat', "Can't adoptChildren after flattening." |
|
434
|
|
|
); |
|
435
|
|
|
foreach ( $elt->children as $child ) { |
|
436
|
|
|
if ( !is_string( $child ) ) { |
|
437
|
|
|
// This is an optimization which avoids an O(n^2) set of |
|
438
|
|
|
// array_splice operations. |
|
439
|
|
|
$child->parent = null; |
|
440
|
|
|
} |
|
441
|
|
|
$this->appendChild( $child ); |
|
442
|
|
|
} |
|
443
|
|
|
$elt->children = []; |
|
444
|
|
|
} |
|
445
|
|
|
|
|
446
|
|
|
/** |
|
447
|
|
|
* Flatten this node and all of its children into a string, as specified |
|
448
|
|
|
* by the HTML serialization specification, and replace this node |
|
449
|
|
|
* in its parent by that string. |
|
450
|
|
|
* |
|
451
|
|
|
* @see __toString() |
|
452
|
|
|
*/ |
|
453
|
|
|
public function flatten( $tidyCompat = false ) { |
|
454
|
|
|
Assert::parameter( $this->parent !== null, '$this', 'must be a child' ); |
|
455
|
|
|
Assert::parameter( $this->parent !== 'flat', '$this', 'already flat' ); |
|
456
|
|
|
$idx = array_search( $this, $this->parent->children, true ); |
|
457
|
|
|
Assert::parameter( |
|
458
|
|
|
$idx !== false, '$this', 'must be a child of its parent' |
|
459
|
|
|
); |
|
460
|
|
|
if ( $tidyCompat ) { |
|
461
|
|
|
$blank = true; |
|
462
|
|
|
foreach ( $this->children as $elt ) { |
|
463
|
|
|
if ( !is_string( $elt ) ) { |
|
464
|
|
|
$elt = $elt->flatten( $tidyCompat ); |
|
465
|
|
|
} |
|
466
|
|
|
if ( $blank && preg_match( '/[^\t\n\f\r ]/', $elt ) ) { |
|
467
|
|
|
$blank = false; |
|
468
|
|
|
} |
|
469
|
|
|
} |
|
470
|
|
|
if ( $this->isA( 'mw:p-wrap' ) ) { |
|
471
|
|
|
$this->localName = 'p'; |
|
472
|
|
|
} elseif ( $blank ) { |
|
473
|
|
|
// Add 'mw-empty-elt' class so elements can be hidden via CSS |
|
474
|
|
|
// for compatibility with legacy tidy. |
|
475
|
|
|
if ( !count( $this->attribs ) && |
|
476
|
|
|
( $this->localName === 'tr' || $this->localName === 'li' ) |
|
477
|
|
|
) { |
|
478
|
|
|
$this->attribs = [ 'class' => "mw-empty-elt" ]; |
|
479
|
|
|
} |
|
480
|
|
|
$blank = false; |
|
481
|
|
|
} |
|
482
|
|
|
$flat = $blank ? '' : "{$this}"; |
|
483
|
|
|
} else { |
|
484
|
|
|
$flat = "{$this}"; |
|
485
|
|
|
} |
|
486
|
|
|
$this->parent->children[$idx] = $flat; |
|
487
|
|
|
$this->parent = 'flat'; # for assertion checking |
|
488
|
|
|
return $flat; |
|
489
|
|
|
} |
|
490
|
|
|
|
|
491
|
|
|
/** |
|
492
|
|
|
* Serialize this node and all of its children to a string, as specified |
|
493
|
|
|
* by the HTML serialization specification. |
|
494
|
|
|
* |
|
495
|
|
|
* @return string The serialization of the BalanceElement |
|
496
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#serialising-html-fragments |
|
497
|
|
|
*/ |
|
498
|
|
|
public function __toString() { |
|
499
|
|
|
$encAttribs = ''; |
|
500
|
|
|
foreach ( $this->attribs as $name => $value ) { |
|
501
|
|
|
$encValue = Sanitizer::encodeAttribute( $value ); |
|
502
|
|
|
$encAttribs .= " $name=\"$encValue\""; |
|
503
|
|
|
} |
|
504
|
|
|
if ( !$this->isA( BalanceSets::$emptyElementSet ) ) { |
|
505
|
|
|
$out = "<{$this->localName}{$encAttribs}>"; |
|
506
|
|
|
// flatten children |
|
507
|
|
|
foreach ( $this->children as $elt ) { |
|
508
|
|
|
$out .= "{$elt}"; |
|
509
|
|
|
} |
|
510
|
|
|
$out .= "</{$this->localName}>"; |
|
511
|
|
|
} else { |
|
512
|
|
|
$out = "<{$this->localName}{$encAttribs} />"; |
|
513
|
|
|
Assert::invariant( |
|
514
|
|
|
count( $this->children ) === 0, |
|
515
|
|
|
"Empty elements shouldn't have children." |
|
516
|
|
|
); |
|
517
|
|
|
} |
|
518
|
|
|
return $out; |
|
519
|
|
|
} |
|
520
|
|
|
|
|
521
|
|
|
# Utility functions on BalanceElements. |
|
522
|
|
|
|
|
523
|
|
|
/** |
|
524
|
|
|
* Determine if $this represents a specific HTML tag, is a member of |
|
525
|
|
|
* a tag set, or is equal to another BalanceElement. |
|
526
|
|
|
* |
|
527
|
|
|
* @param BalanceElement|array|string $set The target BalanceElement, |
|
528
|
|
|
* set (from the BalanceSets class), or string (HTML tag name). |
|
529
|
|
|
* @return bool |
|
530
|
|
|
*/ |
|
531
|
|
|
public function isA( $set ) { |
|
532
|
|
|
Assert::parameterType( 'MediaWiki\Tidy\BalanceElement|array|string', $set, '$set' ); |
|
533
|
|
|
if ( $set instanceof BalanceElement ) { |
|
534
|
|
|
return $this === $set; |
|
535
|
|
|
} elseif ( is_array( $set ) ) { |
|
536
|
|
|
return isset( $set[$this->namespaceURI] ) && |
|
537
|
|
|
isset( $set[$this->namespaceURI][$this->localName] ); |
|
538
|
|
|
} else { |
|
539
|
|
|
# assume this is an HTML element name. |
|
540
|
|
|
return $this->isHtml() && $this->localName === $set; |
|
541
|
|
|
} |
|
542
|
|
|
} |
|
543
|
|
|
|
|
544
|
|
|
/** |
|
545
|
|
|
* Determine if $this represents an element in the HTML namespace. |
|
546
|
|
|
* |
|
547
|
|
|
* @return bool |
|
548
|
|
|
*/ |
|
549
|
|
|
public function isHtml() { |
|
550
|
|
|
return $this->namespaceURI === BalanceSets::HTML_NAMESPACE; |
|
551
|
|
|
} |
|
552
|
|
|
|
|
553
|
|
|
/** |
|
554
|
|
|
* Determine if $this represents a MathML text integration point, |
|
555
|
|
|
* as defined in the HTML5 specification. |
|
556
|
|
|
* |
|
557
|
|
|
* @return bool |
|
558
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#mathml-text-integration-point |
|
559
|
|
|
*/ |
|
560
|
|
|
public function isMathmlTextIntegrationPoint() { |
|
561
|
|
|
return $this->isA( BalanceSets::$mathmlTextIntegrationPointSet ); |
|
562
|
|
|
} |
|
563
|
|
|
|
|
564
|
|
|
/** |
|
565
|
|
|
* Determine if $this represents an HTML integration point, |
|
566
|
|
|
* as defined in the HTML5 specification. |
|
567
|
|
|
* |
|
568
|
|
|
* @return bool |
|
569
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#html-integration-point |
|
570
|
|
|
*/ |
|
571
|
|
|
public function isHtmlIntegrationPoint() { |
|
572
|
|
|
if ( $this->isA( BalanceSets::$htmlIntegrationPointSet ) ) { |
|
573
|
|
|
return true; |
|
574
|
|
|
} |
|
575
|
|
|
if ( |
|
576
|
|
|
$this->namespaceURI === BalanceSets::MATHML_NAMESPACE && |
|
577
|
|
|
$this->localName === 'annotation-xml' && |
|
578
|
|
|
isset( $this->attribs['encoding'] ) && |
|
579
|
|
|
( strcasecmp( $this->attribs['encoding'], 'text/html' ) == 0 || |
|
580
|
|
|
strcasecmp( $this->attribs['encoding'], 'application/xhtml+xml' ) == 0 ) |
|
581
|
|
|
) { |
|
582
|
|
|
return true; |
|
583
|
|
|
} |
|
584
|
|
|
return false; |
|
585
|
|
|
} |
|
586
|
|
|
|
|
587
|
|
|
/** |
|
588
|
|
|
* Get a string key for the Noah's Ark algorithm |
|
589
|
|
|
*/ |
|
590
|
|
|
public function getNoahKey() { |
|
591
|
|
|
if ( $this->noahKey === null ) { |
|
592
|
|
|
$attribs = $this->attribs; |
|
593
|
|
|
ksort( $attribs ); |
|
594
|
|
|
$this->noahKey = serialize( [ $this->namespaceURI, $this->localName, $attribs ] ); |
|
595
|
|
|
} |
|
596
|
|
|
return $this->noahKey; |
|
597
|
|
|
} |
|
598
|
|
|
} |
|
599
|
|
|
|
|
600
|
|
|
/** |
|
601
|
|
|
* The "stack of open elements" as defined in the HTML5 tree builder |
|
602
|
|
|
* spec. This contains methods to ensure that content (start tags, text) |
|
603
|
|
|
* are inserted at the correct place in the output string, and to |
|
604
|
|
|
* flatten BalanceElements are they are closed to avoid holding onto |
|
605
|
|
|
* a complete DOM tree for the document in memory. |
|
606
|
|
|
* |
|
607
|
|
|
* The stack defines a PHP iterator to traverse it in "reverse order", |
|
608
|
|
|
* that is, the most-recently-added element is visited first in a |
|
609
|
|
|
* foreach loop. |
|
610
|
|
|
* |
|
611
|
|
|
* @ingroup Parser |
|
612
|
|
|
* @since 1.27 |
|
613
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#the-stack-of-open-elements |
|
614
|
|
|
*/ |
|
615
|
|
|
class BalanceStack implements IteratorAggregate { |
|
616
|
|
|
/** |
|
617
|
|
|
* Backing storage for the stack. |
|
618
|
|
|
* @var array $elements |
|
619
|
|
|
*/ |
|
620
|
|
|
private $elements = []; |
|
621
|
|
|
/** |
|
622
|
|
|
* Foster parent mode determines how nodes are inserted into the |
|
623
|
|
|
* stack. |
|
624
|
|
|
* @var bool $fosterParentMode |
|
625
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#foster-parent |
|
626
|
|
|
*/ |
|
627
|
|
|
public $fosterParentMode = false; |
|
628
|
|
|
/** |
|
629
|
|
|
* Tidy compatibility mode, determines behavior of body/blockquote |
|
630
|
|
|
*/ |
|
631
|
|
|
public $tidyCompat = false; |
|
632
|
|
|
|
|
633
|
|
|
/** |
|
634
|
|
|
* Create a new BalanceStack with a single BalanceElement on it, |
|
635
|
|
|
* representing the root <html> node. |
|
636
|
|
|
*/ |
|
637
|
|
|
public function __construct() { |
|
638
|
|
|
# always a root <html> element on the stack |
|
639
|
|
|
array_push( |
|
640
|
|
|
$this->elements, |
|
641
|
|
|
new BalanceElement( BalanceSets::HTML_NAMESPACE, 'html', [] ) |
|
642
|
|
|
); |
|
643
|
|
|
} |
|
644
|
|
|
|
|
645
|
|
|
/** |
|
646
|
|
|
* Return a string representing the output of the tree builder: |
|
647
|
|
|
* all the children of the root <html> node. |
|
648
|
|
|
* @return string |
|
649
|
|
|
*/ |
|
650
|
|
|
public function getOutput() { |
|
651
|
|
|
// Don't include the outer '<html>....</html>' |
|
652
|
|
|
$out = ''; |
|
653
|
|
|
foreach ( $this->elements[0]->children as $elt ) { |
|
654
|
|
|
$out .= is_string( $elt ) ? $elt : |
|
655
|
|
|
$elt->flatten( $this->tidyCompat ); |
|
656
|
|
|
} |
|
657
|
|
|
return $out; |
|
658
|
|
|
} |
|
659
|
|
|
|
|
660
|
|
|
/** |
|
661
|
|
|
* Insert text at the appropriate place for inserting a node. |
|
662
|
|
|
* @param string $value |
|
663
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#appropriate-place-for-inserting-a-node |
|
664
|
|
|
*/ |
|
665
|
|
|
public function insertText( $value ) { |
|
666
|
|
|
Assert::parameterType( 'string', $value, '$value' ); |
|
667
|
|
|
if ( |
|
668
|
|
|
$this->fosterParentMode && |
|
669
|
|
|
$this->currentNode()->isA( BalanceSets::$tableSectionRowSet ) |
|
670
|
|
|
) { |
|
671
|
|
|
$this->fosterParent( $value ); |
|
672
|
|
|
} elseif ( |
|
673
|
|
|
$this->tidyCompat && |
|
674
|
|
|
$this->currentNode()->isA( BalanceSets::$tidyPWrapSet ) |
|
675
|
|
|
) { |
|
676
|
|
|
$this->insertHTMLELement( 'mw:p-wrap', [] ); |
|
677
|
|
|
return $this->insertText( $value ); |
|
678
|
|
|
} else { |
|
679
|
|
|
$this->currentNode()->appendChild( $value ); |
|
680
|
|
|
} |
|
681
|
|
|
} |
|
682
|
|
|
|
|
683
|
|
|
/** |
|
684
|
|
|
* Insert a BalanceElement at the appropriate place, pushing it |
|
685
|
|
|
* on to the open elements stack. |
|
686
|
|
|
* @param string $namespaceURI The element namespace |
|
687
|
|
|
* @param string $tag The tag name |
|
688
|
|
|
* @param string $attribs Normalized attributes, as a string. |
|
689
|
|
|
* @return BalanceElement |
|
690
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#insert-a-foreign-element |
|
691
|
|
|
*/ |
|
692
|
|
|
public function insertForeignElement( $namespaceURI, $tag, $attribs ) { |
|
693
|
|
|
return $this->insertElement( |
|
694
|
|
|
new BalanceElement( $namespaceURI, $tag, $attribs ) |
|
695
|
|
|
); |
|
696
|
|
|
} |
|
697
|
|
|
|
|
698
|
|
|
/** |
|
699
|
|
|
* Insert an HTML element at the appropriate place, pushing it on to |
|
700
|
|
|
* the open elements stack. |
|
701
|
|
|
* @param string $tag The tag name |
|
702
|
|
|
* @param string $attribs Normalized attributes, as a string. |
|
703
|
|
|
* @return BalanceElement |
|
704
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#insert-an-html-element |
|
705
|
|
|
*/ |
|
706
|
|
|
public function insertHTMLElement( $tag, $attribs ) { |
|
707
|
|
|
return $this->insertForeignElement( |
|
708
|
|
|
BalanceSets::HTML_NAMESPACE, $tag, $attribs |
|
709
|
|
|
); |
|
710
|
|
|
} |
|
711
|
|
|
|
|
712
|
|
|
/** |
|
713
|
|
|
* Insert an element at the appropriate place and push it on to the |
|
714
|
|
|
* open elements stack. |
|
715
|
|
|
* @param BalanceElement $elt |
|
716
|
|
|
* @return BalanceElement |
|
717
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#appropriate-place-for-inserting-a-node |
|
718
|
|
|
*/ |
|
719
|
|
|
public function insertElement( $elt ) { |
|
720
|
|
|
Assert::parameterType( 'MediaWiki\Tidy\BalanceElement', $elt, '$elt' ); |
|
721
|
|
|
if ( |
|
722
|
|
|
$this->currentNode()->isA( 'mw:p-wrap' ) && |
|
723
|
|
|
!$elt->isA( BalanceSets::$tidyInlineSet ) |
|
724
|
|
|
) { |
|
725
|
|
|
// Tidy compatibility. |
|
726
|
|
|
$this->pop(); |
|
727
|
|
|
} |
|
728
|
|
|
if ( |
|
729
|
|
|
$this->fosterParentMode && |
|
730
|
|
|
$this->currentNode()->isA( BalanceSets::$tableSectionRowSet ) |
|
731
|
|
|
) { |
|
732
|
|
|
$elt = $this->fosterParent( $elt ); |
|
733
|
|
|
} else { |
|
734
|
|
|
$this->currentNode()->appendChild( $elt ); |
|
735
|
|
|
} |
|
736
|
|
|
Assert::invariant( $elt->parent !== null, "$elt must be in tree" ); |
|
737
|
|
|
Assert::invariant( $elt->parent !== 'flat', "$elt must not have been previous flattened" ); |
|
738
|
|
|
array_push( $this->elements, $elt ); |
|
739
|
|
|
return $elt; |
|
740
|
|
|
} |
|
741
|
|
|
|
|
742
|
|
|
/** |
|
743
|
|
|
* Determine if the stack has $tag in scope. |
|
744
|
|
|
* @param BalanceElement|array|string $tag |
|
745
|
|
|
* @return bool |
|
746
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#has-an-element-in-scope |
|
747
|
|
|
*/ |
|
748
|
|
|
public function inScope( $tag ) { |
|
749
|
|
|
return $this->inSpecificScope( $tag, BalanceSets::$inScopeSet ); |
|
750
|
|
|
} |
|
751
|
|
|
|
|
752
|
|
|
/** |
|
753
|
|
|
* Determine if the stack has $tag in button scope. |
|
754
|
|
|
* @param BalanceElement|array|string $tag |
|
755
|
|
|
* @return bool |
|
756
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#has-an-element-in-button-scope |
|
757
|
|
|
*/ |
|
758
|
|
|
public function inButtonScope( $tag ) { |
|
759
|
|
|
return $this->inSpecificScope( $tag, BalanceSets::inButtonScopeSet() ); |
|
760
|
|
|
} |
|
761
|
|
|
|
|
762
|
|
|
/** |
|
763
|
|
|
* Determine if the stack has $tag in list item scope. |
|
764
|
|
|
* @param BalanceElement|array|string $tag |
|
765
|
|
|
* @return bool |
|
766
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#has-an-element-in-list-item-scope |
|
767
|
|
|
*/ |
|
768
|
|
|
public function inListItemScope( $tag ) { |
|
769
|
|
|
return $this->inSpecificScope( $tag, BalanceSets::inListItemScopeSet() ); |
|
770
|
|
|
} |
|
771
|
|
|
|
|
772
|
|
|
/** |
|
773
|
|
|
* Determine if the stack has $tag in table scope. |
|
774
|
|
|
* @param BalanceElement|array|string $tag |
|
775
|
|
|
* @return bool |
|
776
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#has-an-element-in-table-scope |
|
777
|
|
|
*/ |
|
778
|
|
|
public function inTableScope( $tag ) { |
|
779
|
|
|
return $this->inSpecificScope( $tag, BalanceSets::$inTableScopeSet ); |
|
780
|
|
|
} |
|
781
|
|
|
|
|
782
|
|
|
/** |
|
783
|
|
|
* Determine if the stack has $tag in a specific scope, $set. |
|
784
|
|
|
* @param BalanceElement|array|string $tag |
|
785
|
|
|
* @param BalanceElement|array|string $set |
|
786
|
|
|
* @return bool |
|
787
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#has-an-element-in-the-specific-scope |
|
788
|
|
|
*/ |
|
789
|
|
|
public function inSpecificScope( $tag, $set ) { |
|
790
|
|
|
foreach ( $this as $elt ) { |
|
791
|
|
|
if ( $elt->isA( $tag ) ) { |
|
792
|
|
|
return true; |
|
793
|
|
|
} |
|
794
|
|
|
if ( $elt->isA( $set ) ) { |
|
795
|
|
|
return false; |
|
796
|
|
|
} |
|
797
|
|
|
} |
|
798
|
|
|
return false; |
|
799
|
|
|
} |
|
800
|
|
|
|
|
801
|
|
|
/** |
|
802
|
|
|
* Generate implied end tags. |
|
803
|
|
|
* @param BalanceElement|array|string|null $butnot |
|
804
|
|
|
* @param bool $thorough True if we should generate end tags thoroughly. |
|
805
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#generate-implied-end-tags |
|
806
|
|
|
*/ |
|
807
|
|
|
public function generateImpliedEndTags( $butnot = null, $thorough = false ) { |
|
808
|
|
|
$endTagSet = $thorough ? |
|
809
|
|
|
BalanceSets::$thoroughImpliedEndTagsSet : |
|
810
|
|
|
BalanceSets::$impliedEndTagsSet; |
|
811
|
|
|
while ( $this->length() > 0 ) { |
|
812
|
|
|
if ( $butnot !== null && $this->currentNode()->isA( $butnot ) ) { |
|
813
|
|
|
break; |
|
814
|
|
|
} |
|
815
|
|
|
if ( !$this->currentNode()->isA( $endTagSet ) ) { |
|
816
|
|
|
break; |
|
817
|
|
|
} |
|
818
|
|
|
$this->pop(); |
|
819
|
|
|
} |
|
820
|
|
|
} |
|
821
|
|
|
|
|
822
|
|
|
/** |
|
823
|
|
|
* Return the current node (the element in the stack with the largest |
|
824
|
|
|
* index). |
|
825
|
|
|
* @return BalanceElement |
|
826
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#current-node |
|
827
|
|
|
*/ |
|
828
|
|
|
public function currentNode() { |
|
829
|
|
|
return $this->node( count( $this->elements ) - 1 ); |
|
830
|
|
|
} |
|
831
|
|
|
|
|
832
|
|
|
/** |
|
833
|
|
|
* Return the adjusted current node. |
|
834
|
|
|
*/ |
|
835
|
|
|
public function adjustedCurrentNode( $fragmentContext ) { |
|
836
|
|
|
return ( $fragmentContext && $this->length() === 1 ) ? |
|
837
|
|
|
$fragmentContext : $this->currentNode(); |
|
838
|
|
|
} |
|
839
|
|
|
|
|
840
|
|
|
/** |
|
841
|
|
|
* Return an iterator over this stack which visits the current node |
|
842
|
|
|
* first, and the root node last. |
|
843
|
|
|
* @return Iterator |
|
844
|
|
|
*/ |
|
845
|
|
|
public function getIterator() { |
|
846
|
|
|
return new ReverseArrayIterator( $this->elements ); |
|
847
|
|
|
} |
|
848
|
|
|
|
|
849
|
|
|
/** |
|
850
|
|
|
* Return the BalanceElement at the given position $idx, where |
|
851
|
|
|
* position 0 represents the root element. |
|
852
|
|
|
* @param int $idx |
|
853
|
|
|
* @return BalanceElement |
|
854
|
|
|
*/ |
|
855
|
|
|
public function node( $idx ) { |
|
856
|
|
|
return $this->elements[ $idx ]; |
|
857
|
|
|
} |
|
858
|
|
|
|
|
859
|
|
|
/** |
|
860
|
|
|
* Replace the element at position $idx in the BalanceStack with $elt. |
|
861
|
|
|
* @param int $idx |
|
862
|
|
|
* @param BalanceElement $elt |
|
863
|
|
|
*/ |
|
864
|
|
|
public function replaceAt( $idx, $elt ) { |
|
865
|
|
|
Assert::parameterType( 'MediaWiki\Tidy\BalanceElement', $elt, '$elt' ); |
|
866
|
|
|
Assert::precondition( |
|
867
|
|
|
$this->elements[$idx]->parent !== 'flat', |
|
868
|
|
|
'Replaced element should not have already been flattened.' |
|
869
|
|
|
); |
|
870
|
|
|
Assert::precondition( |
|
871
|
|
|
$elt->parent !== 'flat', |
|
872
|
|
|
'New element should not have already been flattened.' |
|
873
|
|
|
); |
|
874
|
|
|
$this->elements[$idx] = $elt; |
|
875
|
|
|
} |
|
876
|
|
|
|
|
877
|
|
|
/** |
|
878
|
|
|
* Return the position of the given BalanceElement, set, or |
|
879
|
|
|
* HTML tag name string in the BalanceStack. |
|
880
|
|
|
* @param BalanceElement|array|string $tag |
|
881
|
|
|
* @return int |
|
882
|
|
|
*/ |
|
883
|
|
|
public function indexOf( $tag ) { |
|
884
|
|
|
for ( $i = count( $this->elements ) - 1; $i >= 0; $i-- ) { |
|
885
|
|
|
if ( $this->elements[$i]->isA( $tag ) ) { |
|
886
|
|
|
return $i; |
|
887
|
|
|
} |
|
888
|
|
|
} |
|
889
|
|
|
return -1; |
|
890
|
|
|
} |
|
891
|
|
|
|
|
892
|
|
|
/** |
|
893
|
|
|
* Return the number of elements currently in the BalanceStack. |
|
894
|
|
|
* @return int |
|
895
|
|
|
*/ |
|
896
|
|
|
public function length() { |
|
897
|
|
|
return count( $this->elements ); |
|
898
|
|
|
} |
|
899
|
|
|
|
|
900
|
|
|
/** |
|
901
|
|
|
* Remove the current node from the BalanceStack, flattening it |
|
902
|
|
|
* in the process. |
|
903
|
|
|
*/ |
|
904
|
|
|
public function pop() { |
|
905
|
|
|
$elt = array_pop( $this->elements ); |
|
906
|
|
|
if ( !$elt->isA( 'mw:p-wrap' ) ) { |
|
907
|
|
|
$elt->flatten( $this->tidyCompat ); |
|
908
|
|
|
} |
|
909
|
|
|
} |
|
910
|
|
|
|
|
911
|
|
|
/** |
|
912
|
|
|
* Remove all nodes up to and including position $idx from the |
|
913
|
|
|
* BalanceStack, flattening them in the process. |
|
914
|
|
|
* @param int $idx |
|
915
|
|
|
*/ |
|
916
|
|
|
public function popTo( $idx ) { |
|
917
|
|
|
while ( $this->length() > $idx ) { |
|
918
|
|
|
$this->pop(); |
|
919
|
|
|
} |
|
920
|
|
|
} |
|
921
|
|
|
|
|
922
|
|
|
/** |
|
923
|
|
|
* Pop elements off the stack up to and including the first |
|
924
|
|
|
* element with the specified HTML tagname (or matching the given |
|
925
|
|
|
* set). |
|
926
|
|
|
* @param BalanceElement|array|string $tag |
|
927
|
|
|
*/ |
|
928
|
|
|
public function popTag( $tag ) { |
|
929
|
|
|
while ( $this->length() > 0 ) { |
|
930
|
|
|
if ( $this->currentNode()->isA( $tag ) ) { |
|
931
|
|
|
$this->pop(); |
|
932
|
|
|
break; |
|
933
|
|
|
} |
|
934
|
|
|
$this->pop(); |
|
935
|
|
|
} |
|
936
|
|
|
} |
|
937
|
|
|
|
|
938
|
|
|
/** |
|
939
|
|
|
* Pop elements off the stack *not including* the first element |
|
940
|
|
|
* in the specified set. |
|
941
|
|
|
* @param BalanceElement|array|string $set |
|
942
|
|
|
*/ |
|
943
|
|
|
public function clearToContext( $set ) { |
|
944
|
|
|
// Note that we don't loop to 0. Never pop the <html> elt off. |
|
945
|
|
|
while ( $this->length() > 1 ) { |
|
946
|
|
|
if ( $this->currentNode()->isA( $set ) ) { |
|
947
|
|
|
break; |
|
948
|
|
|
} |
|
949
|
|
|
$this->pop(); |
|
950
|
|
|
} |
|
951
|
|
|
} |
|
952
|
|
|
|
|
953
|
|
|
/** |
|
954
|
|
|
* Remove the given $elt from the BalanceStack, optionally |
|
955
|
|
|
* flattening it in the process. |
|
956
|
|
|
* @param BalanceElement $elt The element to remove. |
|
957
|
|
|
* @param bool $flatten Whether to flatten the removed element. |
|
958
|
|
|
*/ |
|
959
|
|
|
public function removeElement( $elt, $flatten = true ) { |
|
960
|
|
|
Assert::parameterType( 'MediaWiki\Tidy\BalanceElement', $elt, '$elt' ); |
|
961
|
|
|
Assert::parameter( |
|
962
|
|
|
$elt->parent !== 'flat', |
|
963
|
|
|
'$elt', |
|
964
|
|
|
'$elt should not already have been flattened.' |
|
965
|
|
|
); |
|
966
|
|
|
Assert::parameter( |
|
967
|
|
|
$elt->parent->parent !== 'flat', |
|
968
|
|
|
'$elt', |
|
969
|
|
|
'The parent of $elt should not already have been flattened.' |
|
970
|
|
|
); |
|
971
|
|
|
$idx = array_search( $elt, $this->elements, true ); |
|
972
|
|
|
Assert::parameter( $idx !== false, '$elt', 'must be in stack' ); |
|
973
|
|
|
array_splice( $this->elements, $idx, 1 ); |
|
974
|
|
|
if ( $flatten ) { |
|
975
|
|
|
// serialize $elt into its parent |
|
976
|
|
|
// otherwise, it will eventually serialize when the parent |
|
977
|
|
|
// is serialized, we just hold onto the memory for its |
|
978
|
|
|
// tree of objects a little longer. |
|
979
|
|
|
$elt->flatten( $this->tidyCompat ); |
|
980
|
|
|
} |
|
981
|
|
|
Assert::postcondition( |
|
982
|
|
|
array_search( $elt, $this->elements, true ) === false, |
|
983
|
|
|
'$elt should no longer be in open elements stack' |
|
984
|
|
|
); |
|
985
|
|
|
} |
|
986
|
|
|
|
|
987
|
|
|
/** |
|
988
|
|
|
* Find $a in the BalanceStack and insert $b after it. |
|
989
|
|
|
* @param BalanceElement $a |
|
990
|
|
|
* @param BalanceElement $b |
|
991
|
|
|
*/ |
|
992
|
|
|
public function insertAfter( $a, $b ) { |
|
993
|
|
|
Assert::parameterType( 'MediaWiki\Tidy\BalanceElement', $a, '$a' ); |
|
994
|
|
|
Assert::parameterType( 'MediaWiki\Tidy\BalanceElement', $b, '$b' ); |
|
995
|
|
|
$idx = $this->indexOf( $a ); |
|
996
|
|
|
Assert::parameter( $idx !== false, '$a', 'must be in stack' ); |
|
997
|
|
|
array_splice( $this->elements, $idx + 1, 0, [ $b ] ); |
|
998
|
|
|
} |
|
999
|
|
|
|
|
1000
|
|
|
# Fostering and adoption. |
|
1001
|
|
|
|
|
1002
|
|
|
/** |
|
1003
|
|
|
* Foster parent the given $elt in the stack of open elements. |
|
1004
|
|
|
* @param BalanceElement|string $elt |
|
1005
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#foster-parent |
|
1006
|
|
|
*/ |
|
1007
|
|
|
private function fosterParent( $elt ) { |
|
1008
|
|
|
Assert::parameterType( 'MediaWiki\Tidy\BalanceElement|string', $elt, '$elt' ); |
|
1009
|
|
|
$lastTable = $this->indexOf( 'table' ); |
|
1010
|
|
|
$lastTemplate = $this->indexOf( 'template' ); |
|
1011
|
|
|
$parent = null; |
|
|
|
|
|
|
1012
|
|
|
$before = null; |
|
1013
|
|
|
|
|
1014
|
|
|
if ( $lastTemplate >= 0 && ( $lastTable < 0 || $lastTemplate > $lastTable ) ) { |
|
1015
|
|
|
$parent = $this->elements[$lastTemplate]; |
|
1016
|
|
|
} elseif ( $lastTable >= 0 ) { |
|
1017
|
|
|
$parent = $this->elements[$lastTable]->parent; |
|
1018
|
|
|
# Assume all tables have parents, since we're not running scripts! |
|
1019
|
|
|
Assert::invariant( |
|
1020
|
|
|
$parent !== null, "All tables should have parents" |
|
1021
|
|
|
); |
|
1022
|
|
|
$before = $this->elements[$lastTable]; |
|
1023
|
|
|
} else { |
|
1024
|
|
|
$parent = $this->elements[0]; // the `html` element. |
|
1025
|
|
|
} |
|
1026
|
|
|
|
|
1027
|
|
|
if ( $this->tidyCompat ) { |
|
1028
|
|
|
if ( is_string( $elt ) ) { |
|
1029
|
|
|
// We're fostering text: do we need a p-wrapper? |
|
1030
|
|
|
if ( $parent->isA( BalanceSets::$tidyPWrapSet ) ) { |
|
1031
|
|
|
$this->insertHTMLElement( 'mw:p-wrap', [] ); |
|
1032
|
|
|
$this->insertText( $elt ); |
|
1033
|
|
|
return $elt; |
|
1034
|
|
|
} |
|
1035
|
|
|
} else { |
|
1036
|
|
|
// We're fostering an element; do we need to merge p-wrappers? |
|
1037
|
|
|
if ( $elt->isA( 'mw:p-wrap' ) ) { |
|
1038
|
|
|
$idx = $before ? |
|
1039
|
|
|
array_search( $before, $parent->children, true ) : |
|
1040
|
|
|
count( $parent->children ); |
|
1041
|
|
|
$after = $idx > 0 ? $parent->children[$idx - 1] : ''; |
|
1042
|
|
|
if ( |
|
1043
|
|
|
$after instanceof BalanceElement && |
|
1044
|
|
|
$after->isA( 'mw:p-wrap' ) |
|
1045
|
|
|
) { |
|
1046
|
|
|
return $after; // Re-use existing p-wrapper. |
|
1047
|
|
|
} |
|
1048
|
|
|
} |
|
1049
|
|
|
} |
|
1050
|
|
|
} |
|
1051
|
|
|
|
|
1052
|
|
|
if ( $before ) { |
|
1053
|
|
|
$parent->insertBefore( $before, $elt ); |
|
1054
|
|
|
} else { |
|
1055
|
|
|
$parent->appendChild( $elt ); |
|
1056
|
|
|
} |
|
1057
|
|
|
return $elt; |
|
1058
|
|
|
} |
|
1059
|
|
|
|
|
1060
|
|
|
/** |
|
1061
|
|
|
* Run the "adoption agency algoritm" (AAA) for the given subject |
|
1062
|
|
|
* tag name. |
|
1063
|
|
|
* @param string $tag The subject tag name. |
|
1064
|
|
|
* @param BalanceActiveFormattingElements $afe The current |
|
1065
|
|
|
* active formatting elements list. |
|
1066
|
|
|
* @return true if the adoption agency algorithm "did something", false |
|
1067
|
|
|
* if more processing is required by the caller. |
|
1068
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#adoption-agency-algorithm |
|
1069
|
|
|
*/ |
|
1070
|
|
|
public function adoptionAgency( $tag, $afe ) { |
|
1071
|
|
|
// If the current node is an HTML element whose tag name is subject, |
|
1072
|
|
|
// and the current node is not in the list of active formatting |
|
1073
|
|
|
// elements, then pop the current node off the stack of open |
|
1074
|
|
|
// elements and abort these steps. |
|
1075
|
|
|
if ( |
|
1076
|
|
|
$this->currentNode()->isA( $tag ) && |
|
1077
|
|
|
!$afe->isInList( $this->currentNode() ) |
|
1078
|
|
|
) { |
|
1079
|
|
|
$this->pop(); |
|
1080
|
|
|
return true; // no more handling required |
|
1081
|
|
|
} |
|
1082
|
|
|
|
|
1083
|
|
|
// Let outer loop counter be zero. |
|
1084
|
|
|
$outer = 0; |
|
1085
|
|
|
|
|
1086
|
|
|
// Outer loop: If outer loop counter is greater than or |
|
1087
|
|
|
// equal to eight, then abort these steps. |
|
1088
|
|
|
while ( $outer < 8 ) { |
|
1089
|
|
|
// Increment outer loop counter by one. |
|
1090
|
|
|
$outer++; |
|
1091
|
|
|
|
|
1092
|
|
|
// Let the formatting element be the last element in the list |
|
1093
|
|
|
// of active formatting elements that: is between the end of |
|
1094
|
|
|
// the list and the last scope marker in the list, if any, or |
|
1095
|
|
|
// the start of the list otherwise, and has the same tag name |
|
1096
|
|
|
// as the token. |
|
1097
|
|
|
$fmtelt = $afe->findElementByTag( $tag ); |
|
1098
|
|
|
|
|
1099
|
|
|
// If there is no such node, then abort these steps and instead |
|
1100
|
|
|
// act as described in the "any other end tag" entry below. |
|
1101
|
|
|
if ( !$fmtelt ) { |
|
1102
|
|
|
return false; // false means handle by the default case |
|
1103
|
|
|
} |
|
1104
|
|
|
|
|
1105
|
|
|
// Otherwise, if there is such a node, but that node is not in |
|
1106
|
|
|
// the stack of open elements, then this is a parse error; |
|
1107
|
|
|
// remove the element from the list, and abort these steps. |
|
1108
|
|
|
$index = $this->indexOf( $fmtelt ); |
|
1109
|
|
|
if ( $index < 0 ) { |
|
1110
|
|
|
$afe->remove( $fmtelt ); |
|
1111
|
|
|
return true; // true means no more handling required |
|
1112
|
|
|
} |
|
1113
|
|
|
|
|
1114
|
|
|
// Otherwise, if there is such a node, and that node is also in |
|
1115
|
|
|
// the stack of open elements, but the element is not in scope, |
|
1116
|
|
|
// then this is a parse error; ignore the token, and abort |
|
1117
|
|
|
// these steps. |
|
1118
|
|
|
if ( !$this->inScope( $fmtelt ) ) { |
|
1119
|
|
|
return true; |
|
1120
|
|
|
} |
|
1121
|
|
|
|
|
1122
|
|
|
// Let the furthest block be the topmost node in the stack of |
|
1123
|
|
|
// open elements that is lower in the stack than the formatting |
|
1124
|
|
|
// element, and is an element in the special category. There |
|
1125
|
|
|
// might not be one. |
|
1126
|
|
|
$furthestblock = null; |
|
1127
|
|
|
$furthestblockindex = -1; |
|
1128
|
|
|
$stacklen = $this->length(); |
|
1129
|
|
|
for ( $i = $index+1; $i < $stacklen; $i++ ) { |
|
1130
|
|
|
if ( $this->node( $i )->isA( BalanceSets::$specialSet ) ) { |
|
1131
|
|
|
$furthestblock = $this->node( $i ); |
|
1132
|
|
|
$furthestblockindex = $i; |
|
1133
|
|
|
break; |
|
1134
|
|
|
} |
|
1135
|
|
|
} |
|
1136
|
|
|
|
|
1137
|
|
|
// If there is no furthest block, then the UA must skip the |
|
1138
|
|
|
// subsequent steps and instead just pop all the nodes from the |
|
1139
|
|
|
// bottom of the stack of open elements, from the current node |
|
1140
|
|
|
// up to and including the formatting element, and remove the |
|
1141
|
|
|
// formatting element from the list of active formatting |
|
1142
|
|
|
// elements. |
|
1143
|
|
|
if ( !$furthestblock ) { |
|
1144
|
|
|
$this->popTag( $fmtelt ); |
|
1145
|
|
|
$afe->remove( $fmtelt ); |
|
1146
|
|
|
return true; |
|
1147
|
|
|
} else { |
|
1148
|
|
|
// Let the common ancestor be the element immediately above |
|
1149
|
|
|
// the formatting element in the stack of open elements. |
|
1150
|
|
|
$ancestor = $this->node( $index-1 ); |
|
1151
|
|
|
|
|
1152
|
|
|
// Let a bookmark note the position of the formatting |
|
1153
|
|
|
// element in the list of active formatting elements |
|
1154
|
|
|
// relative to the elements on either side of it in the |
|
1155
|
|
|
// list. |
|
1156
|
|
|
$BOOKMARK = new BalanceElement( '[bookmark]', '[bookmark]', [] ); |
|
1157
|
|
|
$afe->insertAfter( $fmtelt, $BOOKMARK ); |
|
1158
|
|
|
|
|
1159
|
|
|
// Let node and last node be the furthest block. |
|
1160
|
|
|
$node = $furthestblock; |
|
|
|
|
|
|
1161
|
|
|
$lastnode = $furthestblock; |
|
1162
|
|
|
$nodeindex = $furthestblockindex; |
|
1163
|
|
|
$isAFE = false; |
|
|
|
|
|
|
1164
|
|
|
|
|
1165
|
|
|
// Let inner loop counter be zero. |
|
1166
|
|
|
$inner = 0; |
|
1167
|
|
|
|
|
1168
|
|
|
while ( true ) { |
|
1169
|
|
|
|
|
1170
|
|
|
// Increment inner loop counter by one. |
|
1171
|
|
|
$inner++; |
|
1172
|
|
|
|
|
1173
|
|
|
// Let node be the element immediately above node in |
|
1174
|
|
|
// the stack of open elements, or if node is no longer |
|
1175
|
|
|
// in the stack of open elements (e.g. because it got |
|
1176
|
|
|
// removed by this algorithm), the element that was |
|
1177
|
|
|
// immediately above node in the stack of open elements |
|
1178
|
|
|
// before node was removed. |
|
1179
|
|
|
$node = $this->node( --$nodeindex ); |
|
1180
|
|
|
|
|
1181
|
|
|
// If node is the formatting element, then go |
|
1182
|
|
|
// to the next step in the overall algorithm. |
|
1183
|
|
|
if ( $node === $fmtelt ) break; |
|
1184
|
|
|
|
|
1185
|
|
|
// If the inner loop counter is greater than three and node |
|
1186
|
|
|
// is in the list of active formatting elements, then remove |
|
1187
|
|
|
// node from the list of active formatting elements. |
|
1188
|
|
|
$isAFE = $afe->isInList( $node ); |
|
1189
|
|
|
if ( $inner > 3 && $isAFE ) { |
|
1190
|
|
|
$afe->remove( $node ); |
|
1191
|
|
|
$isAFE = false; |
|
1192
|
|
|
} |
|
1193
|
|
|
|
|
1194
|
|
|
// If node is not in the list of active formatting |
|
1195
|
|
|
// elements, then remove node from the stack of open |
|
1196
|
|
|
// elements and then go back to the step labeled inner |
|
1197
|
|
|
// loop. |
|
1198
|
|
|
if ( !$isAFE ) { |
|
1199
|
|
|
// Don't flatten here, since we're about to relocate |
|
1200
|
|
|
// parts of this $node. |
|
1201
|
|
|
$this->removeElement( $node, false ); |
|
1202
|
|
|
continue; |
|
1203
|
|
|
} |
|
1204
|
|
|
|
|
1205
|
|
|
// Create an element for the token for which the |
|
1206
|
|
|
// element node was created with common ancestor as |
|
1207
|
|
|
// the intended parent, replace the entry for node |
|
1208
|
|
|
// in the list of active formatting elements with an |
|
1209
|
|
|
// entry for the new element, replace the entry for |
|
1210
|
|
|
// node in the stack of open elements with an entry for |
|
1211
|
|
|
// the new element, and let node be the new element. |
|
1212
|
|
|
$newelt = new BalanceElement( |
|
1213
|
|
|
$node->namespaceURI, $node->localName, $node->attribs ); |
|
1214
|
|
|
$afe->replace( $node, $newelt ); |
|
1215
|
|
|
$this->replaceAt( $nodeindex, $newelt ); |
|
1216
|
|
|
$node = $newelt; |
|
1217
|
|
|
|
|
1218
|
|
|
// If last node is the furthest block, then move the |
|
1219
|
|
|
// aforementioned bookmark to be immediately after the |
|
1220
|
|
|
// new node in the list of active formatting elements. |
|
1221
|
|
|
if ( $lastnode === $furthestblock ) { |
|
1222
|
|
|
$afe->remove( $BOOKMARK ); |
|
1223
|
|
|
$afe->insertAfter( $newelt, $BOOKMARK ); |
|
1224
|
|
|
} |
|
1225
|
|
|
|
|
1226
|
|
|
// Insert last node into node, first removing it from |
|
1227
|
|
|
// its previous parent node if any. |
|
1228
|
|
|
$node->appendChild( $lastnode ); |
|
1229
|
|
|
|
|
1230
|
|
|
// Let last node be node. |
|
1231
|
|
|
$lastnode = $node; |
|
1232
|
|
|
} |
|
1233
|
|
|
|
|
1234
|
|
|
// If the common ancestor node is a table, tbody, tfoot, |
|
1235
|
|
|
// thead, or tr element, then, foster parent whatever last |
|
1236
|
|
|
// node ended up being in the previous step, first removing |
|
1237
|
|
|
// it from its previous parent node if any. |
|
1238
|
|
|
if ( |
|
1239
|
|
|
$this->fosterParentMode && |
|
1240
|
|
|
$ancestor->isA( BalanceSets::$tableSectionRowSet ) |
|
1241
|
|
|
) { |
|
1242
|
|
|
$this->fosterParent( $lastnode ); |
|
1243
|
|
|
} else { |
|
1244
|
|
|
// Otherwise, append whatever last node ended up being in |
|
1245
|
|
|
// the previous step to the common ancestor node, first |
|
1246
|
|
|
// removing it from its previous parent node if any. |
|
1247
|
|
|
$ancestor->appendChild( $lastnode ); |
|
1248
|
|
|
} |
|
1249
|
|
|
|
|
1250
|
|
|
// Create an element for the token for which the |
|
1251
|
|
|
// formatting element was created, with furthest block |
|
1252
|
|
|
// as the intended parent. |
|
1253
|
|
|
$newelt2 = new BalanceElement( |
|
1254
|
|
|
$fmtelt->namespaceURI, $fmtelt->localName, $fmtelt->attribs ); |
|
1255
|
|
|
|
|
1256
|
|
|
// Take all of the child nodes of the furthest block and |
|
1257
|
|
|
// append them to the element created in the last step. |
|
1258
|
|
|
$newelt2->adoptChildren( $furthestblock ); |
|
1259
|
|
|
|
|
1260
|
|
|
// Append that new element to the furthest block. |
|
1261
|
|
|
$furthestblock->appendChild( $newelt2 ); |
|
1262
|
|
|
|
|
1263
|
|
|
// Remove the formatting element from the list of active |
|
1264
|
|
|
// formatting elements, and insert the new element into the |
|
1265
|
|
|
// list of active formatting elements at the position of |
|
1266
|
|
|
// the aforementioned bookmark. |
|
1267
|
|
|
$afe->remove( $fmtelt ); |
|
1268
|
|
|
$afe->replace( $BOOKMARK, $newelt2 ); |
|
1269
|
|
|
|
|
1270
|
|
|
// Remove the formatting element from the stack of open |
|
1271
|
|
|
// elements, and insert the new element into the stack of |
|
1272
|
|
|
// open elements immediately below the position of the |
|
1273
|
|
|
// furthest block in that stack. |
|
1274
|
|
|
$this->removeElement( $fmtelt ); |
|
1275
|
|
|
$this->insertAfter( $furthestblock, $newelt2 ); |
|
1276
|
|
|
} |
|
1277
|
|
|
} |
|
1278
|
|
|
|
|
1279
|
|
|
return true; |
|
1280
|
|
|
} |
|
1281
|
|
|
|
|
1282
|
|
|
/** |
|
1283
|
|
|
* Return the contents of the open elements stack as a string for |
|
1284
|
|
|
* debugging. |
|
1285
|
|
|
* @return string |
|
1286
|
|
|
*/ |
|
1287
|
|
|
public function __toString() { |
|
1288
|
|
|
$r = []; |
|
1289
|
|
|
foreach ( $this->elements as $elt ) { |
|
1290
|
|
|
array_push( $r, $elt->localName ); |
|
1291
|
|
|
} |
|
1292
|
|
|
return implode( $r, ' ' ); |
|
1293
|
|
|
} |
|
1294
|
|
|
} |
|
1295
|
|
|
|
|
1296
|
|
|
/** |
|
1297
|
|
|
* A pseudo-element used as a marker in the list of active formatting elements |
|
1298
|
|
|
* |
|
1299
|
|
|
* @ingroup Parser |
|
1300
|
|
|
* @since 1.27 |
|
1301
|
|
|
*/ |
|
1302
|
|
|
class BalanceMarker { |
|
1303
|
|
|
public $nextAFE; |
|
1304
|
|
|
public $prevAFE; |
|
1305
|
|
|
} |
|
1306
|
|
|
|
|
1307
|
|
|
/** |
|
1308
|
|
|
* The list of active formatting elements, which is used to handle |
|
1309
|
|
|
* mis-nested formatting element tags in the HTML5 tree builder |
|
1310
|
|
|
* specification. |
|
1311
|
|
|
* |
|
1312
|
|
|
* @ingroup Parser |
|
1313
|
|
|
* @since 1.27 |
|
1314
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#list-of-active-formatting-elements |
|
1315
|
|
|
*/ |
|
1316
|
|
|
class BalanceActiveFormattingElements { |
|
1317
|
|
|
/** The last (most recent) element in the list */ |
|
1318
|
|
|
private $tail; |
|
1319
|
|
|
|
|
1320
|
|
|
/** The first (least recent) element in the list */ |
|
1321
|
|
|
private $head; |
|
1322
|
|
|
|
|
1323
|
|
|
/** |
|
1324
|
|
|
* An array of arrays representing the population of elements in each bucket |
|
1325
|
|
|
* according to the Noah's Ark clause. The outer array is stack-like, with each |
|
1326
|
|
|
* integer-indexed element representing a segment of the list, bounded by |
|
1327
|
|
|
* markers. The first element represents the segment of the list before the |
|
1328
|
|
|
* first marker. |
|
1329
|
|
|
* |
|
1330
|
|
|
* The inner arrays are indexed by "Noah key", which is a string which uniquely |
|
1331
|
|
|
* identifies each bucket according to the rules in the spec. The value in |
|
1332
|
|
|
* the inner array is the first (least recently inserted) element in the bucket, |
|
1333
|
|
|
* and subsequent members of the bucket can be found by iterating through the |
|
1334
|
|
|
* singly-linked list via $node->nextNoah. |
|
1335
|
|
|
* |
|
1336
|
|
|
* This is optimised for the most common case of inserting into a bucket |
|
1337
|
|
|
* with zero members, and deleting a bucket containing one member. In the |
|
1338
|
|
|
* worst case, iteration through the list is still O(1) in the document |
|
1339
|
|
|
* size, since each bucket can have at most 3 members. |
|
1340
|
|
|
*/ |
|
1341
|
|
|
private $noahTableStack = [ [] ]; |
|
1342
|
|
|
|
|
1343
|
|
|
public function __destruct() { |
|
1344
|
|
|
for ( $node = $this->head; $node; $node = $next ) { |
|
1345
|
|
|
$next = $node->nextAFE; |
|
1346
|
|
|
$node->prevAFE = $node->nextAFE = $node->nextNoah = null; |
|
1347
|
|
|
} |
|
1348
|
|
|
$this->head = $this->tail = $this->noahTableStack = null; |
|
|
|
|
|
|
1349
|
|
|
} |
|
1350
|
|
|
|
|
1351
|
|
|
public function insertMarker() { |
|
1352
|
|
|
$elt = new BalanceMarker; |
|
1353
|
|
View Code Duplication |
if ( $this->tail ) { |
|
1354
|
|
|
$this->tail->nextAFE = $elt; |
|
1355
|
|
|
$elt->prevAFE = $this->tail; |
|
1356
|
|
|
} else { |
|
1357
|
|
|
$this->head = $elt; |
|
1358
|
|
|
} |
|
1359
|
|
|
$this->tail = $elt; |
|
1360
|
|
|
$this->noahTableStack[] = []; |
|
1361
|
|
|
} |
|
1362
|
|
|
|
|
1363
|
|
|
/** |
|
1364
|
|
|
* Follow the steps required when the spec requires us to "push onto the |
|
1365
|
|
|
* list of active formatting elements". |
|
1366
|
|
|
* @param BalanceElement $elt |
|
1367
|
|
|
*/ |
|
1368
|
|
|
public function push( BalanceElement $elt ) { |
|
1369
|
|
|
// Must not be in the list already |
|
1370
|
|
|
if ( $elt->prevAFE !== null || $this->head === $elt ) { |
|
1371
|
|
|
throw new ParameterAssertionException( '$elt', |
|
1372
|
|
|
'Cannot insert a node into the AFE list twice' ); |
|
1373
|
|
|
} |
|
1374
|
|
|
|
|
1375
|
|
|
// "Noah's Ark clause" -- if there are already three copies of |
|
1376
|
|
|
// this element before we encounter a marker, then drop the last |
|
1377
|
|
|
// one. |
|
1378
|
|
|
$noahKey = $elt->getNoahKey(); |
|
1379
|
|
|
$table =& $this->noahTableStack[ count( $this->noahTableStack ) - 1 ]; |
|
1380
|
|
|
if ( !isset( $table[$noahKey] ) ) { |
|
1381
|
|
|
$table[$noahKey] = $elt; |
|
1382
|
|
|
} else { |
|
1383
|
|
|
$count = 1; |
|
1384
|
|
|
$head = $tail = $table[$noahKey]; |
|
1385
|
|
|
while ( $tail->nextNoah ) { |
|
1386
|
|
|
$tail = $tail->nextNoah; |
|
1387
|
|
|
$count++; |
|
1388
|
|
|
} |
|
1389
|
|
|
if ( $count >= 3 ) { |
|
1390
|
|
|
$this->remove( $head ); |
|
1391
|
|
|
} |
|
1392
|
|
|
$tail->nextNoah = $elt; |
|
1393
|
|
|
} |
|
1394
|
|
|
// Add to the main AFE list |
|
1395
|
|
View Code Duplication |
if ( $this->tail ) { |
|
1396
|
|
|
$this->tail->nextAFE = $elt; |
|
1397
|
|
|
$elt->prevAFE = $this->tail; |
|
1398
|
|
|
} else { |
|
1399
|
|
|
$this->head = $elt; |
|
1400
|
|
|
} |
|
1401
|
|
|
$this->tail = $elt; |
|
1402
|
|
|
} |
|
1403
|
|
|
|
|
1404
|
|
|
/** |
|
1405
|
|
|
* Follow the steps required when the spec asks us to "clear the list of |
|
1406
|
|
|
* active formatting elements up to the last marker". |
|
1407
|
|
|
*/ |
|
1408
|
|
|
public function clearToMarker() { |
|
1409
|
|
|
// Iterate back through the list starting from the tail |
|
1410
|
|
|
$tail = $this->tail; |
|
1411
|
|
|
while ( $tail && !( $tail instanceof BalanceMarker ) ) { |
|
1412
|
|
|
// Unlink the element |
|
1413
|
|
|
$prev = $tail->prevAFE; |
|
1414
|
|
|
$tail->prevAFE = null; |
|
1415
|
|
|
if ( $prev ) { |
|
1416
|
|
|
$prev->nextAFE = null; |
|
1417
|
|
|
} |
|
1418
|
|
|
$tail->nextNoah = null; |
|
1419
|
|
|
$tail = $prev; |
|
1420
|
|
|
} |
|
1421
|
|
|
// If we finished on a marker, unlink it and pop it off the Noah table stack |
|
1422
|
|
|
if ( $tail ) { |
|
1423
|
|
|
$prev = $tail->prevAFE; |
|
1424
|
|
|
if ( $prev ) { |
|
1425
|
|
|
$prev->nextAFE = null; |
|
1426
|
|
|
} |
|
1427
|
|
|
$tail = $prev; |
|
1428
|
|
|
array_pop( $this->noahTableStack ); |
|
1429
|
|
|
} else { |
|
1430
|
|
|
// No marker: wipe the top-level Noah table (which is the only one) |
|
1431
|
|
|
$this->noahTableStack[0] = []; |
|
1432
|
|
|
} |
|
1433
|
|
|
// If we removed all the elements, clear the head pointer |
|
1434
|
|
|
if ( !$tail ) { |
|
1435
|
|
|
$this->head = null; |
|
1436
|
|
|
} |
|
1437
|
|
|
$this->tail = $tail; |
|
1438
|
|
|
} |
|
1439
|
|
|
|
|
1440
|
|
|
/** |
|
1441
|
|
|
* Find and return the last element with the specified tag between the |
|
1442
|
|
|
* end of the list and the last marker on the list. |
|
1443
|
|
|
* Used when parsing <a> "in body mode". |
|
1444
|
|
|
*/ |
|
1445
|
|
|
public function findElementByTag( $tag ) { |
|
1446
|
|
|
$elt = $this->tail; |
|
1447
|
|
|
while ( $elt && !( $elt instanceof BalanceMarker ) ) { |
|
1448
|
|
|
if ( $elt->localName === $tag ) { |
|
1449
|
|
|
return $elt; |
|
1450
|
|
|
} |
|
1451
|
|
|
$elt = $elt->prevAFE; |
|
1452
|
|
|
} |
|
1453
|
|
|
return null; |
|
1454
|
|
|
} |
|
1455
|
|
|
|
|
1456
|
|
|
/** |
|
1457
|
|
|
* Determine whether an element is in the list of formatting elements. |
|
1458
|
|
|
* @return boolean |
|
1459
|
|
|
*/ |
|
1460
|
|
|
public function isInList( BalanceElement $elt ) { |
|
1461
|
|
|
return $this->head === $elt || $elt->prevAFE; |
|
1462
|
|
|
} |
|
1463
|
|
|
|
|
1464
|
|
|
/** |
|
1465
|
|
|
* Find the element $elt in the list and remove it. |
|
1466
|
|
|
* Used when parsing <a> in body mode. |
|
1467
|
|
|
*/ |
|
1468
|
|
|
public function remove( BalanceElement $elt ) { |
|
1469
|
|
|
if ( $this->head !== $elt && !$elt->prevAFE ) { |
|
1470
|
|
|
throw new ParameterAssertionException( '$elt', |
|
1471
|
|
|
"Attempted to remove an element which is not in the AFE list" ); |
|
1472
|
|
|
} |
|
1473
|
|
|
// Update head and tail pointers |
|
1474
|
|
|
if ( $this->head === $elt ) { |
|
1475
|
|
|
$this->head = $elt->nextAFE; |
|
1476
|
|
|
} |
|
1477
|
|
|
if ( $this->tail === $elt ) { |
|
1478
|
|
|
$this->tail = $elt->prevAFE; |
|
1479
|
|
|
} |
|
1480
|
|
|
// Update previous element |
|
1481
|
|
|
if ( $elt->prevAFE ) { |
|
1482
|
|
|
$elt->prevAFE->nextAFE = $elt->nextAFE; |
|
1483
|
|
|
} |
|
1484
|
|
|
// Update next element |
|
1485
|
|
|
if ( $elt->nextAFE ) { |
|
1486
|
|
|
$elt->nextAFE->prevAFE = $elt->prevAFE; |
|
1487
|
|
|
} |
|
1488
|
|
|
// Clear pointers so that isInList() etc. will work |
|
1489
|
|
|
$elt->prevAFE = $elt->nextAFE = null; |
|
1490
|
|
|
// Update Noah list |
|
1491
|
|
|
$this->removeFromNoahList( $elt ); |
|
1492
|
|
|
} |
|
1493
|
|
|
|
|
1494
|
|
|
private function addToNoahList( BalanceElement $elt ) { |
|
1495
|
|
|
$noahKey = $elt->getNoahKey(); |
|
1496
|
|
|
$table =& $this->noahTableStack[ count( $this->noahTableStack ) - 1 ]; |
|
1497
|
|
|
if ( !isset( $table[$noahKey] ) ) { |
|
1498
|
|
|
$table[$noahKey] = $elt; |
|
1499
|
|
|
} else { |
|
1500
|
|
|
$tail = $table[$noahKey]; |
|
1501
|
|
|
while ( $tail->nextNoah ) { |
|
1502
|
|
|
$tail = $tail->nextNoah; |
|
1503
|
|
|
} |
|
1504
|
|
|
$tail->nextNoah = $elt; |
|
1505
|
|
|
} |
|
1506
|
|
|
} |
|
1507
|
|
|
|
|
1508
|
|
|
private function removeFromNoahList( BalanceElement $elt ) { |
|
1509
|
|
|
$table =& $this->noahTableStack[ count( $this->noahTableStack ) - 1 ]; |
|
1510
|
|
|
$key = $elt->getNoahKey(); |
|
1511
|
|
|
$noahElt = $table[$key]; |
|
1512
|
|
|
if ( $noahElt === $elt ) { |
|
1513
|
|
|
if ( $noahElt->nextNoah ) { |
|
1514
|
|
|
$table[$key] = $noahElt->nextNoah; |
|
1515
|
|
|
$noahElt->nextNoah = null; |
|
1516
|
|
|
} else { |
|
1517
|
|
|
unset( $table[$key] ); |
|
1518
|
|
|
} |
|
1519
|
|
|
} else { |
|
1520
|
|
|
do { |
|
1521
|
|
|
$prevNoahElt = $noahElt; |
|
1522
|
|
|
$noahElt = $prevNoahElt->nextNoah; |
|
1523
|
|
|
if ( $noahElt === $elt ) { |
|
1524
|
|
|
// Found it, unlink |
|
1525
|
|
|
$prevNoahElt->nextNoah = $elt->nextNoah; |
|
1526
|
|
|
$elt->nextNoah = null; |
|
1527
|
|
|
break; |
|
1528
|
|
|
} |
|
1529
|
|
|
} while ( $noahElt ); |
|
1530
|
|
|
} |
|
1531
|
|
|
} |
|
1532
|
|
|
|
|
1533
|
|
|
/** |
|
1534
|
|
|
* Find element $a in the list and replace it with element $b |
|
1535
|
|
|
*/ |
|
1536
|
|
|
public function replace( BalanceElement $a, BalanceElement $b ) { |
|
1537
|
|
|
if ( $this->head !== $a && !$a->prevAFE ) { |
|
1538
|
|
|
throw new ParameterAssertionException( '$a', |
|
1539
|
|
|
"Attempted to replace an element which is not in the AFE list" ); |
|
1540
|
|
|
} |
|
1541
|
|
|
// Update head and tail pointers |
|
1542
|
|
|
if ( $this->head === $a ) { |
|
1543
|
|
|
$this->head = $b; |
|
1544
|
|
|
} |
|
1545
|
|
|
if ( $this->tail === $a ) { |
|
1546
|
|
|
$this->tail = $b; |
|
1547
|
|
|
} |
|
1548
|
|
|
// Update previous element |
|
1549
|
|
|
if ( $a->prevAFE ) { |
|
1550
|
|
|
$a->prevAFE->nextAFE = $b; |
|
1551
|
|
|
} |
|
1552
|
|
|
// Update next element |
|
1553
|
|
|
if ( $a->nextAFE ) { |
|
1554
|
|
|
$a->nextAFE->prevAFE = $b; |
|
1555
|
|
|
} |
|
1556
|
|
|
$b->prevAFE = $a->prevAFE; |
|
1557
|
|
|
$b->nextAFE = $a->nextAFE; |
|
1558
|
|
|
$a->nextAFE = $a->prevAFE = null; |
|
1559
|
|
|
// Update Noah list |
|
1560
|
|
|
$this->removeFromNoahList( $a ); |
|
1561
|
|
|
$this->addToNoahList( $b ); |
|
1562
|
|
|
} |
|
1563
|
|
|
|
|
1564
|
|
|
/** |
|
1565
|
|
|
* Find $a in the list and insert $b after it. |
|
1566
|
|
|
*/ |
|
1567
|
|
|
public function insertAfter( BalanceElement $a, BalanceElement $b ) { |
|
1568
|
|
|
if ( $this->head !== $a && !$a->prevAFE ) { |
|
1569
|
|
|
throw new ParameterAssertionException( '$a', |
|
1570
|
|
|
"Attempted to insert after an element which is not in the AFE list" ); |
|
1571
|
|
|
} |
|
1572
|
|
|
if ( $this->tail === $a ) { |
|
1573
|
|
|
$this->tail = $b; |
|
1574
|
|
|
} |
|
1575
|
|
|
if ( $a->nextAFE ) { |
|
1576
|
|
|
$a->nextAFE->prevAFE = $b; |
|
1577
|
|
|
} |
|
1578
|
|
|
$b->nextAFE = $a->nextAFE; |
|
1579
|
|
|
$b->prevAFE = $a; |
|
1580
|
|
|
$a->nextAFE = $b; |
|
1581
|
|
|
$this->addToNoahList( $b ); |
|
1582
|
|
|
} |
|
1583
|
|
|
|
|
1584
|
|
|
// @codingStandardsIgnoreStart Generic.Files.LineLength.TooLong |
|
1585
|
|
|
/** |
|
1586
|
|
|
* Reconstruct the active formatting elements. |
|
1587
|
|
|
* @param BalanceStack $stack The open elements stack |
|
1588
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#reconstruct-the-active-formatting-elements |
|
1589
|
|
|
*/ |
|
1590
|
|
|
// @codingStandardsIgnoreEnd |
|
1591
|
|
|
public function reconstruct( $stack ) { |
|
1592
|
|
|
$entry = $this->tail; |
|
1593
|
|
|
// If there are no entries in the list of active formatting elements, |
|
1594
|
|
|
// then there is nothing to reconstruct |
|
1595
|
|
|
if ( !$entry ) { |
|
1596
|
|
|
return; |
|
1597
|
|
|
} |
|
1598
|
|
|
// If the last is a marker, do nothing. |
|
1599
|
|
|
if ( $entry instanceof BalanceMarker ) { |
|
1600
|
|
|
return; |
|
1601
|
|
|
} |
|
1602
|
|
|
// Or if it is an open element, do nothing. |
|
1603
|
|
|
if ( $stack->indexOf( $entry ) >= 0 ) { |
|
1604
|
|
|
return; |
|
1605
|
|
|
} |
|
1606
|
|
|
|
|
1607
|
|
|
// Loop backward through the list until we find a marker or an |
|
1608
|
|
|
// open element |
|
1609
|
|
|
while ( $entry->prevAFE ) { |
|
1610
|
|
|
$entry = $entry->prevAFE; |
|
1611
|
|
|
if ( $entry instanceof BalanceMarker || $stack->indexOf( $entry ) >= 0 ) { |
|
1612
|
|
|
break; |
|
1613
|
|
|
} |
|
1614
|
|
|
} |
|
1615
|
|
|
|
|
1616
|
|
|
// Now loop forward, starting from the element after the current one (or |
|
1617
|
|
|
// the first element if we didn't find a marker or open element), |
|
1618
|
|
|
// recreating formatting elements and pushing them back onto the list |
|
1619
|
|
|
// of open elements. |
|
1620
|
|
|
if ( $entry->prevAFE ) { |
|
1621
|
|
|
$entry = $entry->nextAFE; |
|
1622
|
|
|
} |
|
1623
|
|
|
do { |
|
1624
|
|
|
$newElement = $stack->insertHTMLElement( |
|
1625
|
|
|
$entry->localName, |
|
1626
|
|
|
$entry->attribs ); |
|
1627
|
|
|
$this->replace( $entry, $newElement ); |
|
1628
|
|
|
$entry = $newElement->nextAFE; |
|
1629
|
|
|
} while ( $entry ); |
|
1630
|
|
|
} |
|
1631
|
|
|
|
|
1632
|
|
|
/** |
|
1633
|
|
|
* Get a string representation of the AFE list, for debugging |
|
1634
|
|
|
*/ |
|
1635
|
|
|
public function __toString() { |
|
1636
|
|
|
$prev = null; |
|
1637
|
|
|
$s = ''; |
|
1638
|
|
|
for ( $node = $this->head; $node; $prev = $node, $node = $node->nextAFE ) { |
|
1639
|
|
|
if ( $node instanceof BalanceMarker ) { |
|
1640
|
|
|
$s .= "MARKER\n"; |
|
1641
|
|
|
continue; |
|
1642
|
|
|
} |
|
1643
|
|
|
$s .= $node->localName . '#' . substr( md5( spl_object_hash( $node ) ), 0, 8 ); |
|
1644
|
|
|
if ( $node->nextNoah ) { |
|
1645
|
|
|
$s .= " (noah sibling: {$node->nextNoah->localName}#" . |
|
1646
|
|
|
substr( md5( spl_object_hash( $node->nextNoah ) ), 0, 8 ) . |
|
1647
|
|
|
')'; |
|
1648
|
|
|
} |
|
1649
|
|
|
if ( $node->nextAFE && $node->nextAFE->prevAFE !== $node ) { |
|
1650
|
|
|
$s .= " (reverse link is wrong!)"; |
|
1651
|
|
|
} |
|
1652
|
|
|
$s .= "\n"; |
|
1653
|
|
|
} |
|
1654
|
|
|
if ( $prev !== $this->tail ) { |
|
1655
|
|
|
$s .= "(tail pointer is wrong!)\n"; |
|
1656
|
|
|
} |
|
1657
|
|
|
return $s; |
|
1658
|
|
|
} |
|
1659
|
|
|
} |
|
1660
|
|
|
|
|
1661
|
|
|
/** |
|
1662
|
|
|
* An implementation of the tree building portion of the HTML5 parsing |
|
1663
|
|
|
* spec. |
|
1664
|
|
|
* |
|
1665
|
|
|
* This is used to balance and tidy output so that the result can |
|
1666
|
|
|
* always be cleanly serialized/deserialized by an HTML5 parser. It |
|
1667
|
|
|
* does *not* guarantee "conforming" output -- the HTML5 spec contains |
|
1668
|
|
|
* a number of constraints which are not enforced by the HTML5 parsing |
|
1669
|
|
|
* process. But the result will be free of gross errors: misnested or |
|
1670
|
|
|
* unclosed tags, for example, and will be unchanged by spec-complient |
|
1671
|
|
|
* parsing followed by serialization. |
|
1672
|
|
|
* |
|
1673
|
|
|
* The tree building stage is structured as a state machine. |
|
1674
|
|
|
* When comparing the implementation to |
|
1675
|
|
|
* https://www.w3.org/TR/html5/syntax.html#tree-construction |
|
1676
|
|
|
* note that each state is implemented as a function with a |
|
1677
|
|
|
* name ending in `Mode` (because the HTML spec refers to them |
|
1678
|
|
|
* as insertion modes). The current insertion mode is held by |
|
1679
|
|
|
* the $parseMode property. |
|
1680
|
|
|
* |
|
1681
|
|
|
* The following simplifications have been made: |
|
1682
|
|
|
* - We handle body content only (ie, we start `in body`.) |
|
1683
|
|
|
* - The document is never in "quirks mode". |
|
1684
|
|
|
* - All occurrences of < and > have been entity escaped, so we |
|
1685
|
|
|
* can parse tags by simply splitting on those two characters. |
|
1686
|
|
|
* Similarly, all attributes have been "cleaned" and are double-quoted |
|
1687
|
|
|
* and escaped. |
|
1688
|
|
|
* - All comments and null characters are assumed to have been removed. |
|
1689
|
|
|
* - We don't alter linefeeds after <pre>/<listing>. |
|
1690
|
|
|
* - The following elements are disallowed: <html>, <head>, <body>, <frameset>, |
|
1691
|
|
|
* <form>, <frame>, <plaintext>, <isindex>, <textarea>, <xmp>, <iframe>, |
|
1692
|
|
|
* <noembed>, <noscript>, <select>, <script>, <title>. As a result, |
|
1693
|
|
|
* further simplifications can be made: |
|
1694
|
|
|
* - `frameset-ok` is not tracked. |
|
1695
|
|
|
* - `form element pointer` is not tracked. |
|
1696
|
|
|
* - `head element pointer` is not tracked (but presumed non-null) |
|
1697
|
|
|
* - Tokenizer has only a single mode. |
|
1698
|
|
|
* |
|
1699
|
|
|
* We generally mark places where we omit cases from the spec due to |
|
1700
|
|
|
* disallowed elements with a comment: `# OMITTED: <element-name>`. |
|
1701
|
|
|
* |
|
1702
|
|
|
* The HTML spec keeps a flag during the parsing process to track |
|
1703
|
|
|
* whether or not a "parse error" has been encountered. We don't |
|
1704
|
|
|
* bother to track that flag, we just implement the error-handling |
|
1705
|
|
|
* process as specified. |
|
1706
|
|
|
* |
|
1707
|
|
|
* @ingroup Parser |
|
1708
|
|
|
* @since 1.27 |
|
1709
|
|
|
* @see https://html.spec.whatwg.org/multipage/syntax.html#tree-construction |
|
1710
|
|
|
*/ |
|
1711
|
|
|
class Balancer { |
|
1712
|
|
|
private $parseMode; |
|
1713
|
|
|
private $bitsIterator; |
|
1714
|
|
|
private $allowedHtmlElements; |
|
1715
|
|
|
private $afe; |
|
1716
|
|
|
private $stack; |
|
1717
|
|
|
private $strict; |
|
1718
|
|
|
private $tidyCompat; |
|
1719
|
|
|
|
|
1720
|
|
|
private $textIntegrationMode = false; |
|
1721
|
|
|
private $pendingTableText; |
|
1722
|
|
|
private $originalInsertionMode; |
|
1723
|
|
|
private $fragmentContext; |
|
1724
|
|
|
|
|
1725
|
|
|
/** |
|
1726
|
|
|
* Create a new Balancer. |
|
1727
|
|
|
* @param array $config Balancer configuration. Includes: |
|
1728
|
|
|
* 'strict' : boolean, defaults to false. |
|
1729
|
|
|
* When true, enforces syntactic constraints on input: |
|
1730
|
|
|
* all non-tag '<' must be escaped, all attributes must be |
|
1731
|
|
|
* separated by a single space and double-quoted. This is |
|
1732
|
|
|
* consistent with the output of the Sanitizer. |
|
1733
|
|
|
* 'allowedHtmlElements' : array, defaults to null. |
|
1734
|
|
|
* When present, the keys of this associative array give |
|
1735
|
|
|
* the acceptable HTML tag names. When not present, no |
|
1736
|
|
|
* tag sanitization is done. |
|
1737
|
|
|
* 'tidyCompat' : boolean, defaults to false. |
|
1738
|
|
|
* When true, the serialization algorithm is tweaked to |
|
1739
|
|
|
* provide historical compatibility with the old "tidy" |
|
1740
|
|
|
* program: <p>-wrapping is done to the children of |
|
1741
|
|
|
* <body> and <blockquote> elements, and empty elements |
|
1742
|
|
|
* are removed. |
|
1743
|
|
|
*/ |
|
1744
|
|
|
public function __construct( array $config = [] ) { |
|
1745
|
|
|
$config = $config + [ |
|
1746
|
|
|
'strict' => false, |
|
1747
|
|
|
'allowedHtmlElements' => null, |
|
1748
|
|
|
'tidyCompat' => false, |
|
1749
|
|
|
]; |
|
1750
|
|
|
$this->allowedHtmlElements = $config['allowedHtmlElements']; |
|
1751
|
|
|
$this->strict = $config['strict']; |
|
1752
|
|
|
$this->tidyCompat = $config['tidyCompat']; |
|
1753
|
|
|
if ( $this->allowedHtmlElements !== null ) { |
|
1754
|
|
|
# Sanity check! |
|
1755
|
|
|
$bad = array_uintersect_assoc( |
|
1756
|
|
|
$this->allowedHtmlElements, |
|
1757
|
|
|
BalanceSets::$unsupportedSet[BalanceSets::HTML_NAMESPACE], |
|
1758
|
|
|
function( $a, $b ) { |
|
1759
|
|
|
// Ignore the values (just intersect the keys) by saying |
|
1760
|
|
|
// all values are equal to each other. |
|
1761
|
|
|
return 0; |
|
1762
|
|
|
} |
|
1763
|
|
|
); |
|
1764
|
|
|
if ( count( $bad ) > 0 ) { |
|
1765
|
|
|
$badstr = implode( array_keys( $bad ), ',' ); |
|
1766
|
|
|
throw new ParameterAssertionException( |
|
1767
|
|
|
'$config', |
|
1768
|
|
|
'Balance attempted with sanitization including ' . |
|
1769
|
|
|
"unsupported elements: {$badstr}" |
|
1770
|
|
|
); |
|
1771
|
|
|
} |
|
1772
|
|
|
} |
|
1773
|
|
|
} |
|
1774
|
|
|
|
|
1775
|
|
|
/** |
|
1776
|
|
|
* Return a balanced HTML string for the HTML fragment given by $text, |
|
1777
|
|
|
* subject to the caveats listed in the class description. The result |
|
1778
|
|
|
* will typically be idempotent -- that is, rebalancing the output |
|
1779
|
|
|
* would result in no change. |
|
1780
|
|
|
* |
|
1781
|
|
|
* @param string $text The markup to be balanced |
|
1782
|
|
|
* @param callable $processingCallback Callback to do any variable or |
|
1783
|
|
|
* parameter replacements in HTML attributes values |
|
1784
|
|
|
* @param array|bool $processingArgs Arguments for the processing callback |
|
1785
|
|
|
* @return string The balanced markup |
|
1786
|
|
|
*/ |
|
1787
|
|
|
public function balance( $text, $processingCallback = null, $processingArgs = [] ) { |
|
1788
|
|
|
$this->parseMode = 'inBodyMode'; |
|
1789
|
|
|
$this->bitsIterator = new ExplodeIterator( '<', $text ); |
|
1790
|
|
|
$this->afe = new BalanceActiveFormattingElements(); |
|
1791
|
|
|
$this->stack = new BalanceStack(); |
|
1792
|
|
|
$this->stack->tidyCompat = $this->tidyCompat; |
|
1793
|
|
|
$this->processingCallback = $processingCallback; |
|
|
|
|
|
|
1794
|
|
|
$this->processingArgs = $processingArgs; |
|
|
|
|
|
|
1795
|
|
|
|
|
1796
|
|
|
# The stack is constructed with an <html> element already on it. |
|
1797
|
|
|
# Set this up as a fragment parsed with <body> as the context. |
|
1798
|
|
|
$this->fragmentContext = |
|
1799
|
|
|
new BalanceElement( BalanceSets::HTML_NAMESPACE, 'body', [] ); |
|
1800
|
|
|
$this->resetInsertionMode(); |
|
1801
|
|
|
|
|
1802
|
|
|
// First element is text not tag |
|
1803
|
|
|
$x = $this->bitsIterator->current(); |
|
1804
|
|
|
$this->bitsIterator->next(); |
|
1805
|
|
|
$this->insertToken( 'text', str_replace( '>', '>', $x ) ); |
|
1806
|
|
|
// Now process each tag. |
|
1807
|
|
|
while ( $this->bitsIterator->valid() ) { |
|
1808
|
|
|
$this->advance(); |
|
1809
|
|
|
} |
|
1810
|
|
|
$this->insertToken( 'eof', null ); |
|
1811
|
|
|
$result = $this->stack->getOutput(); |
|
1812
|
|
|
// Free memory before returning. |
|
1813
|
|
|
$this->bitsIterator = null; |
|
1814
|
|
|
$this->afe = null; |
|
1815
|
|
|
$this->stack = null; |
|
1816
|
|
|
$this->fragmentContext = null; |
|
1817
|
|
|
return $result; |
|
1818
|
|
|
} |
|
1819
|
|
|
|
|
1820
|
|
|
/** |
|
1821
|
|
|
* Pass a token to the tree builder. The $token will be one of the |
|
1822
|
|
|
* strings "tag", "endtag", or "text". |
|
1823
|
|
|
*/ |
|
1824
|
|
|
private function insertToken( $token, $value, $attribs = null, $selfclose = false ) { |
|
1825
|
|
|
// validate tags against $unsupportedSet |
|
1826
|
|
|
if ( $token === 'tag' || $token === 'endtag' ) { |
|
1827
|
|
|
if ( isset( BalanceSets::$unsupportedSet[BalanceSets::HTML_NAMESPACE][$value] ) ) { |
|
1828
|
|
|
# As described in "simplifications" above, these tags are |
|
1829
|
|
|
# not supported in the balancer. |
|
1830
|
|
|
Assert::invariant( |
|
1831
|
|
|
!$this->strict, |
|
1832
|
|
|
"Unsupported $token <$value> found." |
|
1833
|
|
|
); |
|
1834
|
|
|
return false; |
|
1835
|
|
|
} |
|
1836
|
|
|
} elseif ( $token === 'text' && $value === '' ) { |
|
1837
|
|
|
# Don't actually inject the empty string as a text token. |
|
1838
|
|
|
return true; |
|
1839
|
|
|
} |
|
1840
|
|
|
// Some hoops we have to jump through |
|
1841
|
|
|
$adjusted = $this->stack->adjustedCurrentNode( $this->fragmentContext ); |
|
1842
|
|
|
|
|
1843
|
|
|
$isForeign = true; |
|
1844
|
|
|
if ( |
|
1845
|
|
|
$this->stack->length() === 0 || |
|
1846
|
|
|
$adjusted->isHtml() || |
|
1847
|
|
|
$token === 'eof' |
|
1848
|
|
|
) { |
|
1849
|
|
|
$isForeign = false; |
|
1850
|
|
|
} elseif ( $adjusted->isMathmlTextIntegrationPoint() ) { |
|
1851
|
|
|
if ( $token === 'text' ) { |
|
1852
|
|
|
$isForeign = false; |
|
1853
|
|
|
} elseif ( |
|
1854
|
|
|
$token === 'tag' && |
|
1855
|
|
|
$value !== 'mglyph' && $value !== 'malignmark' |
|
1856
|
|
|
) { |
|
1857
|
|
|
$isForeign = false; |
|
1858
|
|
|
} |
|
1859
|
|
|
} elseif ( |
|
1860
|
|
|
$adjusted->namespaceURI === BalanceSets::MATHML_NAMESPACE && |
|
1861
|
|
|
$adjusted->localName === 'annotation-xml' && |
|
1862
|
|
|
$token === 'tag' && $value === 'svg' |
|
1863
|
|
|
) { |
|
1864
|
|
|
$isForeign = false; |
|
1865
|
|
|
} elseif ( |
|
1866
|
|
|
$adjusted->isHtmlIntegrationPoint() && |
|
1867
|
|
|
( $token === 'tag' || $token === 'text' ) |
|
1868
|
|
|
) { |
|
1869
|
|
|
$isForeign = false; |
|
1870
|
|
|
} |
|
1871
|
|
|
if ( $isForeign ) { |
|
1872
|
|
|
return $this->insertForeignToken( $token, $value, $attribs, $selfclose ); |
|
1873
|
|
|
} else { |
|
1874
|
|
|
$func = $this->parseMode; |
|
1875
|
|
|
return $this->$func( $token, $value, $attribs, $selfclose ); |
|
1876
|
|
|
} |
|
1877
|
|
|
} |
|
1878
|
|
|
|
|
1879
|
|
|
private function insertForeignToken( $token, $value, $attribs = null, $selfclose = false ) { |
|
1880
|
|
|
if ( $token === 'text' ) { |
|
1881
|
|
|
$this->stack->insertText( $value ); |
|
1882
|
|
|
return true; |
|
1883
|
|
|
} elseif ( $token === 'tag' ) { |
|
1884
|
|
|
switch ( $value ) { |
|
1885
|
|
|
case 'font': |
|
1886
|
|
|
if ( isset( $attribs['color'] ) |
|
1887
|
|
|
|| isset( $attribs['face'] ) |
|
1888
|
|
|
|| isset( $attribs['size'] ) |
|
1889
|
|
|
) { |
|
1890
|
|
|
break; |
|
1891
|
|
|
} |
|
1892
|
|
|
/* otherwise, fall through */ |
|
1893
|
|
|
case 'b': |
|
1894
|
|
|
case 'big': |
|
1895
|
|
|
case 'blockquote': |
|
1896
|
|
|
case 'body': |
|
1897
|
|
|
case 'br': |
|
1898
|
|
|
case 'center': |
|
1899
|
|
|
case 'code': |
|
1900
|
|
|
case 'dd': |
|
1901
|
|
|
case 'div': |
|
1902
|
|
|
case 'dl': |
|
1903
|
|
|
case 'dt': |
|
1904
|
|
|
case 'em': |
|
1905
|
|
|
case 'embed': |
|
1906
|
|
|
case 'h1': |
|
1907
|
|
|
case 'h2': |
|
1908
|
|
|
case 'h3': |
|
1909
|
|
|
case 'h4': |
|
1910
|
|
|
case 'h5': |
|
1911
|
|
|
case 'h6': |
|
1912
|
|
|
case 'head': |
|
1913
|
|
|
case 'hr': |
|
1914
|
|
|
case 'i': |
|
1915
|
|
|
case 'img': |
|
1916
|
|
|
case 'li': |
|
1917
|
|
|
case 'listing': |
|
1918
|
|
|
case 'menu': |
|
1919
|
|
|
case 'meta': |
|
1920
|
|
|
case 'nobr': |
|
1921
|
|
|
case 'ol': |
|
1922
|
|
|
case 'p': |
|
1923
|
|
|
case 'pre': |
|
1924
|
|
|
case 'ruby': |
|
1925
|
|
|
case 's': |
|
1926
|
|
|
case 'small': |
|
1927
|
|
|
case 'span': |
|
1928
|
|
|
case 'strong': |
|
1929
|
|
|
case 'strike': |
|
1930
|
|
|
case 'sub': |
|
1931
|
|
|
case 'sup': |
|
1932
|
|
|
case 'table': |
|
1933
|
|
|
case 'tt': |
|
1934
|
|
|
case 'u': |
|
1935
|
|
|
case 'ul': |
|
1936
|
|
|
case 'var': |
|
1937
|
|
|
if ( $this->fragmentContext ) { |
|
1938
|
|
|
break; |
|
1939
|
|
|
} |
|
1940
|
|
|
while ( true ) { |
|
1941
|
|
|
$this->stack->pop(); |
|
1942
|
|
|
$node = $this->stack->currentNode(); |
|
1943
|
|
|
if ( |
|
1944
|
|
|
$node->isMathmlTextIntegrationPoint() || |
|
1945
|
|
|
$node->isHtmlIntegrationPoint() || |
|
1946
|
|
|
$node->isHtml() |
|
1947
|
|
|
) { |
|
1948
|
|
|
break; |
|
1949
|
|
|
} |
|
1950
|
|
|
} |
|
1951
|
|
|
return $this->insertToken( $token, $value, $attribs, $selfclose ); |
|
1952
|
|
|
} |
|
1953
|
|
|
// "Any other start tag" |
|
1954
|
|
|
$adjusted = ( $this->fragmentContext && $this->stack->length()===1 ) ? |
|
1955
|
|
|
$this->fragmentContext : $this->stack->currentNode(); |
|
1956
|
|
|
$this->stack->insertForeignElement( |
|
1957
|
|
|
$adjusted->namespaceURI, $value, $attribs |
|
1958
|
|
|
); |
|
1959
|
|
|
if ( $selfclose ) { |
|
1960
|
|
|
$this->stack->pop(); |
|
1961
|
|
|
} |
|
1962
|
|
|
return true; |
|
1963
|
|
|
} elseif ( $token === 'endtag' ) { |
|
1964
|
|
|
$first = true; |
|
1965
|
|
|
foreach ( $this->stack as $i => $node ) { |
|
|
|
|
|
|
1966
|
|
|
if ( $node->isHtml() && !$first ) { |
|
1967
|
|
|
// process the end tag as HTML |
|
1968
|
|
|
$func = $this->parseMode; |
|
1969
|
|
|
return $this->$func( $token, $value, $attribs, $selfclose ); |
|
1970
|
|
|
} elseif ( $i === 0 ) { |
|
1971
|
|
|
return true; |
|
1972
|
|
|
} elseif ( $node->localName === $value ) { |
|
1973
|
|
|
$this->stack->popTag( $node ); |
|
1974
|
|
|
return true; |
|
1975
|
|
|
} |
|
1976
|
|
|
$first = false; |
|
1977
|
|
|
} |
|
1978
|
|
|
} |
|
1979
|
|
|
} |
|
1980
|
|
|
|
|
1981
|
|
|
/** |
|
1982
|
|
|
* Grab the next "token" from $bitsIterator. This is either a open/close |
|
1983
|
|
|
* tag or text, depending on whether the Sanitizer approves. |
|
1984
|
|
|
*/ |
|
1985
|
|
|
private function advance() { |
|
1986
|
|
|
$x = $this->bitsIterator->current(); |
|
1987
|
|
|
$this->bitsIterator->next(); |
|
1988
|
|
|
$regs = []; |
|
1989
|
|
|
# $slash: Does the current element start with a '/'? |
|
1990
|
|
|
# $t: Current element name |
|
1991
|
|
|
# $attribStr: String between element name and > |
|
1992
|
|
|
# $brace: Ending '>' or '/>' |
|
1993
|
|
|
# $rest: Everything until the next element from the $bitsIterator |
|
1994
|
|
|
if ( preg_match( Sanitizer::ELEMENT_BITS_REGEX, $x, $regs ) ) { |
|
1995
|
|
|
list( /* $qbar */, $slash, $t, $attribStr, $brace, $rest ) = $regs; |
|
1996
|
|
|
$t = strtolower( $t ); |
|
1997
|
|
|
if ( $this->strict ) { |
|
1998
|
|
|
/* Verify that attributes are all properly double-quoted */ |
|
1999
|
|
|
Assert::invariant( |
|
2000
|
|
|
preg_match( |
|
2001
|
|
|
'/^( [:_A-Z0-9][-.:_A-Z0-9]*="[^"]*")*[ ]*$/i', $attribStr |
|
2002
|
|
|
), |
|
2003
|
|
|
"Bad attribute string found" |
|
2004
|
|
|
); |
|
2005
|
|
|
} |
|
2006
|
|
|
} else { |
|
2007
|
|
|
Assert::invariant( |
|
2008
|
|
|
!$this->strict, "< found which does not start a valid tag" |
|
2009
|
|
|
); |
|
2010
|
|
|
$slash = $t = $attribStr = $brace = $rest = null; |
|
2011
|
|
|
} |
|
2012
|
|
|
$goodtag = $t; |
|
2013
|
|
|
$sanitize = $this->allowedHtmlElements !== null; |
|
2014
|
|
|
if ( $sanitize ) { |
|
2015
|
|
|
$goodtag = $t && isset( $this->allowedHtmlElements[$t] ); |
|
|
|
|
|
|
2016
|
|
|
} |
|
2017
|
|
|
if ( $goodtag ) { |
|
2018
|
|
|
if ( is_callable( $this->processingCallback ) ) { |
|
2019
|
|
|
call_user_func_array( $this->processingCallback, [ &$attribStr, $this->processingArgs ] ); |
|
2020
|
|
|
} |
|
2021
|
|
|
if ( $sanitize ) { |
|
2022
|
|
|
$goodtag = Sanitizer::validateTag( $attribStr, $t ); |
|
2023
|
|
|
} |
|
2024
|
|
|
} |
|
2025
|
|
|
if ( $goodtag ) { |
|
2026
|
|
|
if ( $sanitize ) { |
|
2027
|
|
|
$attribs = Sanitizer::decodeTagAttributes( $attribStr ); |
|
2028
|
|
|
$attribs = Sanitizer::validateTagAttributes( $attribs, $t ); |
|
2029
|
|
|
} else { |
|
2030
|
|
|
$attribs = Sanitizer::decodeTagAttributes( $attribStr ); |
|
2031
|
|
|
} |
|
2032
|
|
|
$goodtag = $this->insertToken( |
|
2033
|
|
|
$slash ? 'endtag' : 'tag', $t, $attribs, $brace === '/>' |
|
2034
|
|
|
); |
|
2035
|
|
|
} |
|
2036
|
|
|
if ( $goodtag ) { |
|
2037
|
|
|
$rest = str_replace( '>', '>', $rest ); |
|
2038
|
|
|
$this->insertToken( 'text', str_replace( '>', '>', $rest ) ); |
|
2039
|
|
|
} else { |
|
2040
|
|
|
# bad tag; serialize entire thing as text. |
|
2041
|
|
|
$this->insertToken( 'text', '<' . str_replace( '>', '>', $x ) ); |
|
2042
|
|
|
} |
|
2043
|
|
|
} |
|
2044
|
|
|
|
|
2045
|
|
|
private function switchMode( $mode ) { |
|
2046
|
|
|
Assert::parameter( |
|
2047
|
|
|
substr( $mode, -4 )==='Mode', '$mode', 'should end in Mode' |
|
2048
|
|
|
); |
|
2049
|
|
|
$oldMode = $this->parseMode; |
|
2050
|
|
|
$this->parseMode = $mode; |
|
2051
|
|
|
return $oldMode; |
|
2052
|
|
|
} |
|
2053
|
|
|
|
|
2054
|
|
|
private function switchModeAndReprocess( $mode, $token, $value, $attribs, $selfclose ) { |
|
2055
|
|
|
$this->switchMode( $mode ); |
|
2056
|
|
|
return $this->insertToken( $token, $value, $attribs, $selfclose ); |
|
2057
|
|
|
} |
|
2058
|
|
|
|
|
2059
|
|
|
private function resetInsertionMode() { |
|
2060
|
|
|
$last = false; |
|
2061
|
|
|
foreach ( $this->stack as $i => $node ) { |
|
|
|
|
|
|
2062
|
|
|
if ( $i === 0 ) { |
|
2063
|
|
|
$last = true; |
|
2064
|
|
|
if ( $this->fragmentContext ) { |
|
2065
|
|
|
$node = $this->fragmentContext; |
|
2066
|
|
|
} |
|
2067
|
|
|
} |
|
2068
|
|
|
if ( $node->isHtml() ) { |
|
2069
|
|
|
switch ( $node->localName ) { |
|
2070
|
|
|
# OMITTED: <select> |
|
2071
|
|
|
/* |
|
2072
|
|
|
case 'select': |
|
2073
|
|
|
$stacklen = $this->stack->length(); |
|
2074
|
|
|
for ( $j = $i + 1; $j < $stacklen-1; $j++ ) { |
|
2075
|
|
|
$ancestor = $this->stack->node( $stacklen-$j-1 ); |
|
2076
|
|
|
if ( $ancestor->isA( 'template' ) ) { |
|
2077
|
|
|
break; |
|
2078
|
|
|
} |
|
2079
|
|
|
if ( $ancestor->isA( 'table' ) ) { |
|
2080
|
|
|
$this->switchMode( 'inSelectInTableMode' ); |
|
2081
|
|
|
return; |
|
2082
|
|
|
} |
|
2083
|
|
|
} |
|
2084
|
|
|
$this->switchMode( 'inSelectMode' ); |
|
2085
|
|
|
return; |
|
2086
|
|
|
*/ |
|
2087
|
|
|
case 'tr': |
|
2088
|
|
|
$this->switchMode( 'inRowMode' ); |
|
2089
|
|
|
return; |
|
2090
|
|
|
case 'tbody': |
|
2091
|
|
|
case 'tfoot': |
|
2092
|
|
|
case 'thead': |
|
2093
|
|
|
$this->switchMode( 'inTableBodyMode' ); |
|
2094
|
|
|
return; |
|
2095
|
|
|
case 'caption': |
|
2096
|
|
|
$this->switchMode( 'inCaptionMode' ); |
|
2097
|
|
|
return; |
|
2098
|
|
|
case 'colgroup': |
|
2099
|
|
|
$this->switchMode( 'inColumnGroupMode' ); |
|
2100
|
|
|
return; |
|
2101
|
|
|
case 'table': |
|
2102
|
|
|
$this->switchMode( 'inTableMode' ); |
|
2103
|
|
|
return; |
|
2104
|
|
|
case 'template': |
|
2105
|
|
|
$this->switchMode( |
|
2106
|
|
|
array_slice( $this->templateInsertionModes, -1 )[0] |
|
|
|
|
|
|
2107
|
|
|
); |
|
2108
|
|
|
return; |
|
2109
|
|
|
case 'body': |
|
2110
|
|
|
$this->switchMode( 'inBodyMode' ); |
|
2111
|
|
|
return; |
|
2112
|
|
|
# OMITTED: <frameset> |
|
2113
|
|
|
# OMITTED: <html> |
|
2114
|
|
|
# OMITTED: <head> |
|
2115
|
|
|
default: |
|
2116
|
|
|
if ( !$last ) { |
|
2117
|
|
|
# OMITTED: <head> |
|
2118
|
|
|
if ( $node->isA( BalanceSets::$tableCellSet ) ) { |
|
2119
|
|
|
$this->switchMode( 'inCellMode' ); |
|
2120
|
|
|
return; |
|
2121
|
|
|
} |
|
2122
|
|
|
} |
|
2123
|
|
|
} |
|
2124
|
|
|
} |
|
2125
|
|
|
if ( $last ) { |
|
2126
|
|
|
$this->switchMode( 'inBodyMode' ); |
|
2127
|
|
|
return; |
|
2128
|
|
|
} |
|
2129
|
|
|
} |
|
2130
|
|
|
} |
|
2131
|
|
|
|
|
2132
|
|
|
private function stopParsing() { |
|
2133
|
|
|
# Most of the spec methods are inapplicable, other than step 2: |
|
2134
|
|
|
# "pop all the nodes off the stack of open elements". |
|
2135
|
|
|
# We're going to keep the top-most <html> element on the stack, though. |
|
2136
|
|
|
|
|
2137
|
|
|
# Clear the AFE list first, otherwise the element objects will stay live |
|
2138
|
|
|
# during serialization, potentially using O(N^2) memory. Note that |
|
2139
|
|
|
# popping the stack will never result in reconstructing the active |
|
2140
|
|
|
# formatting elements. |
|
2141
|
|
|
$this->afe = null; |
|
2142
|
|
|
$this->stack->popTo( 1 ); |
|
2143
|
|
|
} |
|
2144
|
|
|
|
|
2145
|
|
|
private function parseRawText( $value, $attribs = null ) { |
|
2146
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2147
|
|
|
// XXX switch tokenizer to rawtext state? |
|
2148
|
|
|
$this->originalInsertionMode = $this->switchMode( 'inTextMode' ); |
|
2149
|
|
|
return true; |
|
2150
|
|
|
} |
|
2151
|
|
|
|
|
2152
|
|
|
private function inTextMode( $token, $value, $attribs = null, $selfclose = false ) { |
|
2153
|
|
|
if ( $token === 'text' ) { |
|
2154
|
|
|
$this->stack->insertText( $value ); |
|
2155
|
|
|
return true; |
|
2156
|
|
|
} elseif ( $token === 'eof' ) { |
|
2157
|
|
|
$this->stack->pop(); |
|
2158
|
|
|
return $this->switchModeAndReprocess( |
|
2159
|
|
|
$this->originalInsertionMode, $token, $value, $attribs, $selfclose |
|
2160
|
|
|
); |
|
2161
|
|
|
} elseif ( $token === 'endtag' ) { |
|
2162
|
|
|
$this->stack->pop(); |
|
2163
|
|
|
$this->switchMode( $this->originalInsertionMode ); |
|
2164
|
|
|
return true; |
|
2165
|
|
|
} |
|
2166
|
|
|
return true; |
|
2167
|
|
|
} |
|
2168
|
|
|
|
|
2169
|
|
|
private function inHeadMode( $token, $value, $attribs = null, $selfclose = false ) { |
|
2170
|
|
|
if ( $token === 'text' ) { |
|
2171
|
|
View Code Duplication |
if ( preg_match( '/^[\x09\x0A\x0C\x0D\x20]+/', $value, $matches ) ) { |
|
2172
|
|
|
$this->stack->insertText( $matches[0] ); |
|
2173
|
|
|
$value = substr( $value, strlen( $matches[0] ) ); |
|
2174
|
|
|
} |
|
2175
|
|
|
if ( strlen( $value ) === 0 ) { |
|
2176
|
|
|
return true; // All text handled. |
|
2177
|
|
|
} |
|
2178
|
|
|
// Fall through to handle non-whitespace below. |
|
2179
|
|
|
} elseif ( $token === 'tag' ) { |
|
2180
|
|
|
switch ( $value ) { |
|
2181
|
|
|
case 'meta': |
|
2182
|
|
|
# OMITTED: in a full HTML parser, this might change the encoding. |
|
2183
|
|
|
/* falls through */ |
|
2184
|
|
|
# OMITTED: <html> |
|
2185
|
|
|
case 'base': |
|
2186
|
|
|
case 'basefont': |
|
2187
|
|
|
case 'bgsound': |
|
2188
|
|
|
case 'link': |
|
2189
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2190
|
|
|
$this->stack->pop(); |
|
2191
|
|
|
return true; |
|
2192
|
|
|
# OMITTED: <title> |
|
2193
|
|
|
# OMITTED: <noscript> |
|
2194
|
|
|
case 'noframes': |
|
2195
|
|
|
case 'style': |
|
2196
|
|
|
return $this->parseRawText( $value, $attribs ); |
|
2197
|
|
|
# OMITTED: <script> |
|
2198
|
|
|
case 'template': |
|
2199
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2200
|
|
|
$this->afe->insertMarker(); |
|
2201
|
|
|
# OMITTED: frameset_ok |
|
2202
|
|
|
$this->switchMode( 'inTemplateMode' ); |
|
2203
|
|
|
$this->templateInsertionModes[] = $this->parseMode; |
|
2204
|
|
|
return true; |
|
2205
|
|
|
# OMITTED: <head> |
|
2206
|
|
|
} |
|
2207
|
|
|
} elseif ( $token === 'endtag' ) { |
|
2208
|
|
|
switch ( $value ) { |
|
2209
|
|
|
# OMITTED: <head> |
|
2210
|
|
|
# OMITTED: <body> |
|
2211
|
|
|
# OMITTED: <html> |
|
2212
|
|
|
case 'br': |
|
2213
|
|
|
break; // handle at the bottom of the function |
|
2214
|
|
|
case 'template': |
|
2215
|
|
|
if ( $this->stack->indexOf( $value ) < 0 ) { |
|
2216
|
|
|
return true; // Ignore the token. |
|
2217
|
|
|
} |
|
2218
|
|
|
$this->stack->generateImpliedEndTags( null, true /* thorough */ ); |
|
2219
|
|
|
$this->stack->popTag( $value ); |
|
2220
|
|
|
$this->afe->clearToMarker(); |
|
2221
|
|
|
array_pop( $this->templateInsertionModes ); |
|
2222
|
|
|
$this->resetInsertionMode(); |
|
2223
|
|
|
return true; |
|
2224
|
|
|
default: |
|
2225
|
|
|
// ignore any other end tag |
|
2226
|
|
|
return true; |
|
2227
|
|
|
} |
|
2228
|
|
|
} |
|
2229
|
|
|
|
|
2230
|
|
|
// If not handled above |
|
2231
|
|
|
$this->inHeadMode( 'endtag', 'head' ); // synthetic </head> |
|
2232
|
|
|
// Then redo this one |
|
2233
|
|
|
return $this->insertToken( $token, $value, $attribs, $selfclose ); |
|
2234
|
|
|
} |
|
2235
|
|
|
|
|
2236
|
|
|
private function inBodyMode( $token, $value, $attribs = null, $selfclose = false ) { |
|
2237
|
|
|
if ( $token === 'text' ) { |
|
2238
|
|
|
$this->afe->reconstruct( $this->stack ); |
|
2239
|
|
|
$this->stack->insertText( $value ); |
|
2240
|
|
|
return true; |
|
2241
|
|
|
} elseif ( $token === 'eof' ) { |
|
2242
|
|
|
if ( !empty( $this->templateInsertionModes ) ) { |
|
2243
|
|
|
return $this->inTemplateMode( $token, $value, $attribs, $selfclose ); |
|
2244
|
|
|
} |
|
2245
|
|
|
$this->stopParsing(); |
|
2246
|
|
|
return true; |
|
2247
|
|
|
} elseif ( $token === 'tag' ) { |
|
2248
|
|
|
switch ( $value ) { |
|
2249
|
|
|
# OMITTED: <html> |
|
2250
|
|
|
case 'base': |
|
2251
|
|
|
case 'basefont': |
|
2252
|
|
|
case 'bgsound': |
|
2253
|
|
|
case 'link': |
|
2254
|
|
|
case 'meta': |
|
2255
|
|
|
case 'noframes': |
|
2256
|
|
|
# OMITTED: <script> |
|
2257
|
|
|
case 'style': |
|
2258
|
|
|
case 'template': |
|
2259
|
|
|
# OMITTED: <title> |
|
2260
|
|
|
return $this->inHeadMode( $token, $value, $attribs, $selfclose ); |
|
2261
|
|
|
# OMITTED: <body> |
|
2262
|
|
|
# OMITTED: <frameset> |
|
2263
|
|
|
|
|
2264
|
|
|
case 'address': |
|
2265
|
|
|
case 'article': |
|
2266
|
|
|
case 'aside': |
|
2267
|
|
|
case 'blockquote': |
|
2268
|
|
|
case 'center': |
|
2269
|
|
|
case 'details': |
|
2270
|
|
|
case 'dialog': |
|
2271
|
|
|
case 'dir': |
|
2272
|
|
|
case 'div': |
|
2273
|
|
|
case 'dl': |
|
2274
|
|
|
case 'fieldset': |
|
2275
|
|
|
case 'figcaption': |
|
2276
|
|
|
case 'figure': |
|
2277
|
|
|
case 'footer': |
|
2278
|
|
|
case 'header': |
|
2279
|
|
|
case 'hgroup': |
|
2280
|
|
|
case 'main': |
|
2281
|
|
|
case 'menu': |
|
2282
|
|
|
case 'nav': |
|
2283
|
|
|
case 'ol': |
|
2284
|
|
|
case 'p': |
|
2285
|
|
|
case 'section': |
|
2286
|
|
|
case 'summary': |
|
2287
|
|
|
case 'ul': |
|
2288
|
|
|
if ( $this->stack->inButtonScope( 'p' ) ) { |
|
2289
|
|
|
$this->inBodyMode( 'endtag', 'p' ); |
|
2290
|
|
|
} |
|
2291
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2292
|
|
|
return true; |
|
2293
|
|
|
|
|
2294
|
|
|
case 'h1': |
|
2295
|
|
|
case 'h2': |
|
2296
|
|
|
case 'h3': |
|
2297
|
|
|
case 'h4': |
|
2298
|
|
|
case 'h5': |
|
2299
|
|
|
case 'h6': |
|
2300
|
|
|
if ( $this->stack->inButtonScope( 'p' ) ) { |
|
2301
|
|
|
$this->inBodyMode( 'endtag', 'p' ); |
|
2302
|
|
|
} |
|
2303
|
|
|
if ( $this->stack->currentNode()->isA( BalanceSets::$headingSet ) ) { |
|
2304
|
|
|
$this->stack->pop(); |
|
2305
|
|
|
} |
|
2306
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2307
|
|
|
return true; |
|
2308
|
|
|
|
|
2309
|
|
|
case 'pre': |
|
2310
|
|
|
case 'listing': |
|
2311
|
|
|
if ( $this->stack->inButtonScope( 'p' ) ) { |
|
2312
|
|
|
$this->inBodyMode( 'endtag', 'p' ); |
|
2313
|
|
|
} |
|
2314
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2315
|
|
|
# As described in "simplifications" above: |
|
2316
|
|
|
# 1. We don't touch the next token, even if it's a linefeed. |
|
2317
|
|
|
# 2. OMITTED: frameset_ok |
|
2318
|
|
|
return true; |
|
2319
|
|
|
|
|
2320
|
|
|
# OMITTED: <form> |
|
2321
|
|
|
|
|
2322
|
|
|
case 'li': |
|
2323
|
|
|
# OMITTED: frameset_ok |
|
2324
|
|
|
foreach ( $this->stack as $node ) { |
|
|
|
|
|
|
2325
|
|
|
if ( $node->isA( 'li' ) ) { |
|
2326
|
|
|
$this->inBodyMode( 'endtag', 'li' ); |
|
2327
|
|
|
break; |
|
2328
|
|
|
} |
|
2329
|
|
|
if ( |
|
2330
|
|
|
$node->isA( BalanceSets::$specialSet ) && |
|
2331
|
|
|
!$node->isA( BalanceSets::$addressDivPSet ) |
|
2332
|
|
|
) { |
|
2333
|
|
|
break; |
|
2334
|
|
|
} |
|
2335
|
|
|
} |
|
2336
|
|
|
if ( $this->stack->inButtonScope( 'p' ) ) { |
|
2337
|
|
|
$this->inBodyMode( 'endtag', 'p' ); |
|
2338
|
|
|
} |
|
2339
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2340
|
|
|
return true; |
|
2341
|
|
|
|
|
2342
|
|
|
case 'dd': |
|
2343
|
|
|
case 'dt': |
|
2344
|
|
|
# OMITTED: frameset_ok |
|
2345
|
|
|
foreach ( $this->stack as $node ) { |
|
|
|
|
|
|
2346
|
|
|
if ( $node->isA( 'dd' ) ) { |
|
2347
|
|
|
$this->inBodyMode( 'endtag', 'dd' ); |
|
2348
|
|
|
break; |
|
2349
|
|
|
} |
|
2350
|
|
|
if ( $node->isA( 'dt' ) ) { |
|
2351
|
|
|
$this->inBodyMode( 'endtag', 'dt' ); |
|
2352
|
|
|
break; |
|
2353
|
|
|
} |
|
2354
|
|
|
if ( |
|
2355
|
|
|
$node->isA( BalanceSets::$specialSet ) && |
|
2356
|
|
|
!$node->isA( BalanceSets::$addressDivPSet ) |
|
2357
|
|
|
) { |
|
2358
|
|
|
break; |
|
2359
|
|
|
} |
|
2360
|
|
|
} |
|
2361
|
|
|
if ( $this->stack->inButtonScope( 'p' ) ) { |
|
2362
|
|
|
$this->inBodyMode( 'endtag', 'p' ); |
|
2363
|
|
|
} |
|
2364
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2365
|
|
|
return true; |
|
2366
|
|
|
|
|
2367
|
|
|
# OMITTED: <plaintext> |
|
2368
|
|
|
|
|
2369
|
|
View Code Duplication |
case 'button': |
|
2370
|
|
|
if ( $this->stack->inScope( 'button' ) ) { |
|
2371
|
|
|
$this->inBodyMode( 'endtag', 'button' ); |
|
2372
|
|
|
return $this->insertToken( $token, $value, $attribs, $selfclose ); |
|
2373
|
|
|
} |
|
2374
|
|
|
$this->afe->reconstruct( $this->stack ); |
|
2375
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2376
|
|
|
return true; |
|
2377
|
|
|
|
|
2378
|
|
|
case 'a': |
|
2379
|
|
|
$activeElement = $this->afe->findElementByTag( 'a' ); |
|
2380
|
|
|
if ( $activeElement ) { |
|
2381
|
|
|
$this->inBodyMode( 'endtag', 'a' ); |
|
2382
|
|
|
if ( $this->afe->isInList( $activeElement ) ) { |
|
2383
|
|
|
$this->afe->remove( $activeElement ); |
|
2384
|
|
|
// Don't flatten here, since when we fall |
|
2385
|
|
|
// through below we might foster parent |
|
2386
|
|
|
// the new <a> tag inside this one. |
|
2387
|
|
|
$this->stack->removeElement( $activeElement, false ); |
|
2388
|
|
|
} |
|
2389
|
|
|
} |
|
2390
|
|
|
/* Falls through */ |
|
2391
|
|
|
case 'b': |
|
2392
|
|
|
case 'big': |
|
2393
|
|
|
case 'code': |
|
2394
|
|
|
case 'em': |
|
2395
|
|
|
case 'font': |
|
2396
|
|
|
case 'i': |
|
2397
|
|
|
case 's': |
|
2398
|
|
|
case 'small': |
|
2399
|
|
|
case 'strike': |
|
2400
|
|
|
case 'strong': |
|
2401
|
|
|
case 'tt': |
|
2402
|
|
View Code Duplication |
case 'u': |
|
2403
|
|
|
$this->afe->reconstruct( $this->stack ); |
|
2404
|
|
|
$this->afe->push( $this->stack->insertHTMLElement( $value, $attribs ), $attribs ); |
|
2405
|
|
|
return true; |
|
2406
|
|
|
|
|
2407
|
|
|
case 'nobr': |
|
2408
|
|
|
$this->afe->reconstruct( $this->stack ); |
|
2409
|
|
|
if ( $this->stack->inScope( 'nobr' ) ) { |
|
2410
|
|
|
$this->inBodyMode( 'endtag', 'nobr' ); |
|
2411
|
|
|
$this->afe->reconstruct( $this->stack ); |
|
2412
|
|
|
} |
|
2413
|
|
|
$this->afe->push( $this->stack->insertHTMLElement( $value, $attribs ), $attribs ); |
|
2414
|
|
|
return true; |
|
2415
|
|
|
|
|
2416
|
|
|
case 'applet': |
|
2417
|
|
|
case 'marquee': |
|
2418
|
|
|
case 'object': |
|
2419
|
|
|
$this->afe->reconstruct( $this->stack ); |
|
2420
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2421
|
|
|
$this->afe->insertMarker(); |
|
2422
|
|
|
# OMITTED: frameset_ok |
|
2423
|
|
|
return true; |
|
2424
|
|
|
|
|
2425
|
|
View Code Duplication |
case 'table': |
|
2426
|
|
|
# The document is never in "quirks mode"; see simplifications |
|
2427
|
|
|
# above. |
|
2428
|
|
|
if ( $this->stack->inButtonScope( 'p' ) ) { |
|
2429
|
|
|
$this->inBodyMode( 'endtag', 'p' ); |
|
2430
|
|
|
} |
|
2431
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2432
|
|
|
# OMITTED: frameset_ok |
|
2433
|
|
|
$this->switchMode( 'inTableMode' ); |
|
2434
|
|
|
return true; |
|
2435
|
|
|
|
|
2436
|
|
|
case 'area': |
|
2437
|
|
|
case 'br': |
|
2438
|
|
|
case 'embed': |
|
2439
|
|
|
case 'img': |
|
2440
|
|
|
case 'keygen': |
|
2441
|
|
View Code Duplication |
case 'wbr': |
|
2442
|
|
|
$this->afe->reconstruct( $this->stack ); |
|
2443
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2444
|
|
|
$this->stack->pop(); |
|
2445
|
|
|
# OMITTED: frameset_ok |
|
2446
|
|
|
return true; |
|
2447
|
|
|
|
|
2448
|
|
View Code Duplication |
case 'input': |
|
2449
|
|
|
$this->afe->reconstruct( $this->stack ); |
|
2450
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2451
|
|
|
$this->stack->pop(); |
|
2452
|
|
|
# OMITTED: frameset_ok |
|
2453
|
|
|
# (hence we don't need to examine the tag's "type" attribute) |
|
2454
|
|
|
return true; |
|
2455
|
|
|
|
|
2456
|
|
|
case 'menuitem': |
|
2457
|
|
|
case 'param': |
|
2458
|
|
|
case 'source': |
|
2459
|
|
|
case 'track': |
|
2460
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2461
|
|
|
$this->stack->pop(); |
|
2462
|
|
|
return true; |
|
2463
|
|
|
|
|
2464
|
|
View Code Duplication |
case 'hr': |
|
2465
|
|
|
if ( $this->stack->inButtonScope( 'p' ) ) { |
|
2466
|
|
|
$this->inBodyMode( 'endtag', 'p' ); |
|
2467
|
|
|
} |
|
2468
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2469
|
|
|
$this->stack->pop(); |
|
2470
|
|
|
return true; |
|
2471
|
|
|
|
|
2472
|
|
|
case 'image': |
|
2473
|
|
|
# warts! |
|
2474
|
|
|
return $this->inBodyMode( $token, 'img', $attribs, $selfclose ); |
|
2475
|
|
|
|
|
2476
|
|
|
# OMITTED: <isindex> |
|
2477
|
|
|
# OMITTED: <textarea> |
|
2478
|
|
|
# OMITTED: <xmp> |
|
2479
|
|
|
# OMITTED: <iframe> |
|
2480
|
|
|
# OMITTED: <noembed> |
|
2481
|
|
|
# OMITTED: <noscript> |
|
2482
|
|
|
|
|
2483
|
|
|
# OMITTED: <select> |
|
2484
|
|
|
/* |
|
2485
|
|
|
case 'select': |
|
2486
|
|
|
$this->afe->reconstruct( $this->stack ); |
|
2487
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2488
|
|
|
switch ( $this->parseMode ) { |
|
2489
|
|
|
case 'inTableMode': |
|
2490
|
|
|
case 'inCaptionMode': |
|
2491
|
|
|
case 'inTableBodyMode': |
|
2492
|
|
|
case 'inRowMode': |
|
2493
|
|
|
case 'inCellMode': |
|
2494
|
|
|
$this->switchMode( 'inSelectInTableMode' ); |
|
2495
|
|
|
return true; |
|
2496
|
|
|
default: |
|
2497
|
|
|
$this->switchMode( 'inSelectMode' ); |
|
2498
|
|
|
return true; |
|
2499
|
|
|
} |
|
2500
|
|
|
*/ |
|
2501
|
|
|
|
|
2502
|
|
|
case 'optgroup': |
|
2503
|
|
View Code Duplication |
case 'option': |
|
2504
|
|
|
if ( $this->stack->currentNode()->isA( 'option' ) ) { |
|
2505
|
|
|
$this->inBodyMode( 'endtag', 'option' ); |
|
2506
|
|
|
} |
|
2507
|
|
|
$this->afe->reconstruct( $this->stack ); |
|
2508
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2509
|
|
|
return true; |
|
2510
|
|
|
|
|
2511
|
|
|
case 'rb': |
|
2512
|
|
|
case 'rtc': |
|
2513
|
|
|
if ( $this->stack->inScope( 'ruby' ) ) { |
|
2514
|
|
|
$this->stack->generateImpliedEndTags(); |
|
2515
|
|
|
} |
|
2516
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2517
|
|
|
return true; |
|
2518
|
|
|
|
|
2519
|
|
|
case 'rp': |
|
2520
|
|
|
case 'rt': |
|
2521
|
|
|
if ( $this->stack->inScope( 'ruby' ) ) { |
|
2522
|
|
|
$this->stack->generateImpliedEndTags( 'rtc' ); |
|
2523
|
|
|
} |
|
2524
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2525
|
|
|
return true; |
|
2526
|
|
|
|
|
2527
|
|
View Code Duplication |
case 'math': |
|
2528
|
|
|
$this->afe->reconstruct( $this->stack ); |
|
2529
|
|
|
# We skip the spec's "adjust MathML attributes" and |
|
2530
|
|
|
# "adjust foreign attributes" steps, since the browser will |
|
2531
|
|
|
# do this later when it parses the output and it doesn't affect |
|
2532
|
|
|
# balancing. |
|
2533
|
|
|
$this->stack->insertForeignElement( |
|
2534
|
|
|
BalanceSets::MATHML_NAMESPACE, $value, $attribs |
|
2535
|
|
|
); |
|
2536
|
|
|
if ( $selfclose ) { |
|
2537
|
|
|
# emit explicit </math> tag. |
|
2538
|
|
|
$this->stack->pop(); |
|
2539
|
|
|
} |
|
2540
|
|
|
return true; |
|
2541
|
|
|
|
|
2542
|
|
View Code Duplication |
case 'svg': |
|
2543
|
|
|
$this->afe->reconstruct( $this->stack ); |
|
2544
|
|
|
# We skip the spec's "adjust SVG attributes" and |
|
2545
|
|
|
# "adjust foreign attributes" steps, since the browser will |
|
2546
|
|
|
# do this later when it parses the output and it doesn't affect |
|
2547
|
|
|
# balancing. |
|
2548
|
|
|
$this->stack->insertForeignElement( |
|
2549
|
|
|
BalanceSets::SVG_NAMESPACE, $value, $attribs |
|
2550
|
|
|
); |
|
2551
|
|
|
if ( $selfclose ) { |
|
2552
|
|
|
# emit explicit </svg> tag. |
|
2553
|
|
|
$this->stack->pop(); |
|
2554
|
|
|
} |
|
2555
|
|
|
return true; |
|
2556
|
|
|
|
|
2557
|
|
|
case 'caption': |
|
2558
|
|
|
case 'col': |
|
2559
|
|
|
case 'colgroup': |
|
2560
|
|
|
# OMITTED: <frame> |
|
2561
|
|
|
case 'head': |
|
2562
|
|
|
case 'tbody': |
|
2563
|
|
|
case 'td': |
|
2564
|
|
|
case 'tfoot': |
|
2565
|
|
|
case 'th': |
|
2566
|
|
|
case 'thead': |
|
2567
|
|
|
case 'tr': |
|
2568
|
|
|
// Ignore table tags if we're not inTableMode |
|
2569
|
|
|
return true; |
|
2570
|
|
|
} |
|
2571
|
|
|
|
|
2572
|
|
|
// Handle any other start tag here |
|
2573
|
|
|
$this->afe->reconstruct( $this->stack ); |
|
2574
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2575
|
|
|
return true; |
|
2576
|
|
|
} elseif ( $token === 'endtag' ) { |
|
2577
|
|
|
switch ( $value ) { |
|
2578
|
|
|
# </body>,</html> are unsupported. |
|
2579
|
|
|
|
|
2580
|
|
|
case 'template': |
|
2581
|
|
|
return $this->inHeadMode( $token, $value, $attribs, $selfclose ); |
|
2582
|
|
|
|
|
2583
|
|
|
case 'address': |
|
2584
|
|
|
case 'article': |
|
2585
|
|
|
case 'aside': |
|
2586
|
|
|
case 'blockquote': |
|
2587
|
|
|
case 'button': |
|
2588
|
|
|
case 'center': |
|
2589
|
|
|
case 'details': |
|
2590
|
|
|
case 'dialog': |
|
2591
|
|
|
case 'dir': |
|
2592
|
|
|
case 'div': |
|
2593
|
|
|
case 'dl': |
|
2594
|
|
|
case 'fieldset': |
|
2595
|
|
|
case 'figcaption': |
|
2596
|
|
|
case 'figure': |
|
2597
|
|
|
case 'footer': |
|
2598
|
|
|
case 'header': |
|
2599
|
|
|
case 'hgroup': |
|
2600
|
|
|
case 'listing': |
|
2601
|
|
|
case 'main': |
|
2602
|
|
|
case 'menu': |
|
2603
|
|
|
case 'nav': |
|
2604
|
|
|
case 'ol': |
|
2605
|
|
|
case 'pre': |
|
2606
|
|
|
case 'section': |
|
2607
|
|
|
case 'summary': |
|
2608
|
|
|
case 'ul': |
|
2609
|
|
|
// Ignore if there is not a matching open tag |
|
2610
|
|
|
if ( !$this->stack->inScope( $value ) ) { |
|
2611
|
|
|
return true; |
|
2612
|
|
|
} |
|
2613
|
|
|
$this->stack->generateImpliedEndTags(); |
|
2614
|
|
|
$this->stack->popTag( $value ); |
|
2615
|
|
|
return true; |
|
2616
|
|
|
|
|
2617
|
|
|
# OMITTED: <form> |
|
2618
|
|
|
|
|
2619
|
|
View Code Duplication |
case 'p': |
|
2620
|
|
|
if ( !$this->stack->inButtonScope( 'p' ) ) { |
|
2621
|
|
|
$this->inBodyMode( 'tag', 'p', [] ); |
|
2622
|
|
|
return $this->insertToken( $token, $value, $attribs, $selfclose ); |
|
2623
|
|
|
} |
|
2624
|
|
|
$this->stack->generateImpliedEndTags( $value ); |
|
2625
|
|
|
$this->stack->popTag( $value ); |
|
2626
|
|
|
return true; |
|
2627
|
|
|
|
|
2628
|
|
|
case 'li': |
|
2629
|
|
|
if ( !$this->stack->inListItemScope( $value ) ) { |
|
2630
|
|
|
return true; # ignore |
|
2631
|
|
|
} |
|
2632
|
|
|
$this->stack->generateImpliedEndTags( $value ); |
|
2633
|
|
|
$this->stack->popTag( $value ); |
|
2634
|
|
|
return true; |
|
2635
|
|
|
|
|
2636
|
|
|
case 'dd': |
|
2637
|
|
|
case 'dt': |
|
2638
|
|
|
if ( !$this->stack->inScope( $value ) ) { |
|
2639
|
|
|
return true; # ignore |
|
2640
|
|
|
} |
|
2641
|
|
|
$this->stack->generateImpliedEndTags( $value ); |
|
2642
|
|
|
$this->stack->popTag( $value ); |
|
2643
|
|
|
return true; |
|
2644
|
|
|
|
|
2645
|
|
|
case 'h1': |
|
2646
|
|
|
case 'h2': |
|
2647
|
|
|
case 'h3': |
|
2648
|
|
|
case 'h4': |
|
2649
|
|
|
case 'h5': |
|
2650
|
|
|
case 'h6': |
|
2651
|
|
|
if ( !$this->stack->inScope( BalanceSets::$headingSet ) ) { |
|
2652
|
|
|
return; |
|
2653
|
|
|
} |
|
2654
|
|
|
$this->stack->generateImpliedEndTags(); |
|
2655
|
|
|
$this->stack->popTag( BalanceSets::$headingSet ); |
|
2656
|
|
|
return true; |
|
2657
|
|
|
|
|
2658
|
|
|
case 'sarcasm': |
|
2659
|
|
|
# Take a deep breath, then: |
|
2660
|
|
|
break; |
|
2661
|
|
|
|
|
2662
|
|
|
case 'a': |
|
2663
|
|
|
case 'b': |
|
2664
|
|
|
case 'big': |
|
2665
|
|
|
case 'code': |
|
2666
|
|
|
case 'em': |
|
2667
|
|
|
case 'font': |
|
2668
|
|
|
case 'i': |
|
2669
|
|
|
case 'nobr': |
|
2670
|
|
|
case 's': |
|
2671
|
|
|
case 'small': |
|
2672
|
|
|
case 'strike': |
|
2673
|
|
|
case 'strong': |
|
2674
|
|
|
case 'tt': |
|
2675
|
|
|
case 'u': |
|
2676
|
|
|
if ( $this->stack->adoptionAgency( $value, $this->afe ) ) { |
|
|
|
|
|
|
2677
|
|
|
return true; # If we did something, we're done. |
|
2678
|
|
|
} |
|
2679
|
|
|
break; # Go to the "any other end tag" case. |
|
2680
|
|
|
|
|
2681
|
|
|
case 'applet': |
|
2682
|
|
|
case 'marquee': |
|
2683
|
|
View Code Duplication |
case 'object': |
|
2684
|
|
|
if ( !$this->stack->inScope( $value ) ) { |
|
2685
|
|
|
return true; # ignore |
|
2686
|
|
|
} |
|
2687
|
|
|
$this->stack->generateImpliedEndTags(); |
|
2688
|
|
|
$this->stack->popTag( $value ); |
|
2689
|
|
|
$this->afe->clearToMarker(); |
|
2690
|
|
|
return true; |
|
2691
|
|
|
|
|
2692
|
|
|
case 'br': |
|
2693
|
|
|
# Turn </br> into <br> |
|
2694
|
|
|
return $this->inBodyMode( 'tag', $value, [] ); |
|
2695
|
|
|
} |
|
2696
|
|
|
|
|
2697
|
|
|
// Any other end tag goes here |
|
2698
|
|
|
foreach ( $this->stack as $i => $node ) { |
|
|
|
|
|
|
2699
|
|
|
if ( $node->isA( $value ) ) { |
|
2700
|
|
|
$this->stack->generateImpliedEndTags( $value ); |
|
2701
|
|
|
$this->stack->popTo( $i ); # including $i |
|
2702
|
|
|
break; |
|
2703
|
|
|
} elseif ( $node->isA( BalanceSets::$specialSet ) ) { |
|
2704
|
|
|
return true; // ignore this close token. |
|
2705
|
|
|
} |
|
2706
|
|
|
} |
|
2707
|
|
|
return true; |
|
2708
|
|
|
} else { |
|
2709
|
|
|
Assert::invariant( false, "Bad token type: $token" ); |
|
2710
|
|
|
} |
|
2711
|
|
|
} |
|
2712
|
|
|
|
|
2713
|
|
|
private function inTableMode( $token, $value, $attribs = null, $selfclose = false ) { |
|
2714
|
|
|
if ( $token === 'text' ) { |
|
2715
|
|
|
if ( $this->textIntegrationMode ) { |
|
2716
|
|
|
return $this->inBodyMode( $token, $value, $attribs, $selfclose ); |
|
2717
|
|
|
} elseif ( $this->stack->currentNode()->isA( BalanceSets::$tableSectionRowSet ) ) { |
|
2718
|
|
|
$this->pendingTableText = ''; |
|
2719
|
|
|
$this->originalInsertionMode = $this->parseMode; |
|
2720
|
|
|
return $this->switchModeAndReprocess( 'inTableTextMode', $token, $value, $attribs, $selfclose ); |
|
2721
|
|
|
} |
|
2722
|
|
|
// fall through to default case. |
|
2723
|
|
|
} elseif ( $token === 'eof' ) { |
|
2724
|
|
|
$this->stopParsing(); |
|
2725
|
|
|
return true; |
|
2726
|
|
|
} elseif ( $token === 'tag' ) { |
|
2727
|
|
|
switch ( $value ) { |
|
2728
|
|
|
case 'caption': |
|
2729
|
|
|
$this->afe->insertMarker(); |
|
2730
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2731
|
|
|
$this->switchMode( 'inCaptionMode' ); |
|
2732
|
|
|
return true; |
|
2733
|
|
|
case 'colgroup': |
|
2734
|
|
|
$this->stack->clearToContext( BalanceSets::$tableContextSet ); |
|
2735
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2736
|
|
|
$this->switchMode( 'inColumnGroupMode' ); |
|
2737
|
|
|
return true; |
|
2738
|
|
|
case 'col': |
|
2739
|
|
|
$this->inTableMode( 'tag', 'colgroup', [] ); |
|
2740
|
|
|
return $this->insertToken( $token, $value, $attribs, $selfclose ); |
|
2741
|
|
|
case 'tbody': |
|
2742
|
|
|
case 'tfoot': |
|
2743
|
|
|
case 'thead': |
|
2744
|
|
|
$this->stack->clearToContext( BalanceSets::$tableContextSet ); |
|
2745
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2746
|
|
|
$this->switchMode( 'inTableBodyMode' ); |
|
2747
|
|
|
return true; |
|
2748
|
|
|
case 'td': |
|
2749
|
|
|
case 'th': |
|
2750
|
|
|
case 'tr': |
|
2751
|
|
|
$this->inTableMode( 'tag', 'tbody', [] ); |
|
2752
|
|
|
return $this->insertToken( $token, $value, $attribs, $selfclose ); |
|
2753
|
|
|
case 'table': |
|
2754
|
|
|
if ( !$this->stack->inTableScope( $value ) ) { |
|
2755
|
|
|
return true; // Ignore this tag. |
|
2756
|
|
|
} |
|
2757
|
|
|
$this->inTableMode( 'endtag', $value ); |
|
2758
|
|
|
return $this->insertToken( $token, $value, $attribs, $selfclose ); |
|
2759
|
|
|
|
|
2760
|
|
|
case 'style': |
|
2761
|
|
|
# OMITTED: <script> |
|
2762
|
|
|
case 'template': |
|
2763
|
|
|
return $this->inHeadMode( $token, $value, $attribs, $selfclose ); |
|
2764
|
|
|
|
|
2765
|
|
|
case 'input': |
|
2766
|
|
|
if ( !isset( $attribs['type'] ) || strcasecmp( $attribs['type'], 'hidden' ) !== 0 ) { |
|
2767
|
|
|
break; // Handle this as "everything else" |
|
2768
|
|
|
} |
|
2769
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2770
|
|
|
$this->stack->pop(); |
|
2771
|
|
|
return true; |
|
2772
|
|
|
|
|
2773
|
|
|
# OMITTED: <form> |
|
2774
|
|
|
} |
|
2775
|
|
|
// Fall through for "anything else" clause. |
|
2776
|
|
|
} elseif ( $token === 'endtag' ) { |
|
2777
|
|
|
switch ( $value ) { |
|
2778
|
|
|
case 'table': |
|
2779
|
|
|
if ( !$this->stack->inTableScope( $value ) ) { |
|
2780
|
|
|
return true; // Ignore. |
|
2781
|
|
|
} |
|
2782
|
|
|
$this->stack->popTag( $value ); |
|
2783
|
|
|
$this->resetInsertionMode(); |
|
2784
|
|
|
return true; |
|
2785
|
|
|
# OMITTED: <body> |
|
2786
|
|
|
case 'caption': |
|
2787
|
|
|
case 'col': |
|
2788
|
|
|
case 'colgroup': |
|
2789
|
|
|
# OMITTED: <html> |
|
2790
|
|
|
case 'tbody': |
|
2791
|
|
|
case 'td': |
|
2792
|
|
|
case 'tfoot': |
|
2793
|
|
|
case 'th': |
|
2794
|
|
|
case 'thead': |
|
2795
|
|
|
case 'tr': |
|
2796
|
|
|
return true; // Ignore the token. |
|
2797
|
|
|
case 'template': |
|
2798
|
|
|
return $this->inHeadMode( $token, $value, $attribs, $selfclose ); |
|
2799
|
|
|
} |
|
2800
|
|
|
// Fall through for "anything else" clause. |
|
2801
|
|
|
} |
|
2802
|
|
|
// This is the "anything else" case: |
|
2803
|
|
|
$this->stack->fosterParentMode = true; |
|
2804
|
|
|
$this->inBodyMode( $token, $value, $attribs, $selfclose ); |
|
2805
|
|
|
$this->stack->fosterParentMode = false; |
|
2806
|
|
|
return true; |
|
2807
|
|
|
} |
|
2808
|
|
|
|
|
2809
|
|
|
private function inTableTextMode( $token, $value, $attribs = null, $selfclose = false ) { |
|
2810
|
|
|
if ( $token === 'text' ) { |
|
2811
|
|
|
$this->pendingTableText .= $value; |
|
2812
|
|
|
return true; |
|
2813
|
|
|
} |
|
2814
|
|
|
// Non-text token: |
|
2815
|
|
|
$text = $this->pendingTableText; |
|
2816
|
|
|
$this->pendingTableText = ''; |
|
2817
|
|
|
if ( preg_match( '/[^\x09\x0A\x0C\x0D\x20]/', $text ) ) { |
|
2818
|
|
|
// This should match the "anything else" case inTableMode |
|
2819
|
|
|
$this->stack->fosterParentMode = true; |
|
2820
|
|
|
$this->inBodyMode( 'text', $text ); |
|
2821
|
|
|
$this->stack->fosterParentMode = false; |
|
2822
|
|
|
} else { |
|
2823
|
|
|
// Pending text is just whitespace. |
|
2824
|
|
|
$this->stack->insertText( $text ); |
|
2825
|
|
|
} |
|
2826
|
|
|
return $this->switchModeAndReprocess( |
|
2827
|
|
|
$this->originalInsertionMode, $token, $value, $attribs, $selfclose |
|
2828
|
|
|
); |
|
2829
|
|
|
} |
|
2830
|
|
|
|
|
2831
|
|
|
// helper for inCaptionMode |
|
2832
|
|
|
private function endCaption() { |
|
2833
|
|
|
if ( !$this->stack->inTableScope( 'caption' ) ) { |
|
2834
|
|
|
return false; |
|
2835
|
|
|
} |
|
2836
|
|
|
$this->stack->generateImpliedEndTags(); |
|
2837
|
|
|
$this->stack->popTag( 'caption' ); |
|
2838
|
|
|
$this->afe->clearToMarker(); |
|
2839
|
|
|
$this->switchMode( 'inTableMode' ); |
|
2840
|
|
|
return true; |
|
2841
|
|
|
} |
|
2842
|
|
|
|
|
2843
|
|
|
private function inCaptionMode( $token, $value, $attribs = null, $selfclose = false ) { |
|
2844
|
|
|
if ( $token === 'tag' ) { |
|
2845
|
|
View Code Duplication |
switch ( $value ) { |
|
2846
|
|
|
case 'caption': |
|
2847
|
|
|
case 'col': |
|
2848
|
|
|
case 'colgroup': |
|
2849
|
|
|
case 'tbody': |
|
2850
|
|
|
case 'td': |
|
2851
|
|
|
case 'tfoot': |
|
2852
|
|
|
case 'th': |
|
2853
|
|
|
case 'thead': |
|
2854
|
|
|
case 'tr': |
|
2855
|
|
|
if ( $this->endCaption() ) { |
|
2856
|
|
|
$this->insertToken( $token, $value, $attribs, $selfclose ); |
|
2857
|
|
|
} |
|
2858
|
|
|
return true; |
|
2859
|
|
|
} |
|
2860
|
|
|
// Fall through to "anything else" case. |
|
2861
|
|
|
} elseif ( $token === 'endtag' ) { |
|
2862
|
|
|
switch ( $value ) { |
|
2863
|
|
|
case 'caption': |
|
2864
|
|
|
$this->endCaption(); |
|
2865
|
|
|
return true; |
|
2866
|
|
|
case 'table': |
|
2867
|
|
|
if ( $this->endCaption() ) { |
|
2868
|
|
|
$this->insertToken( $token, $value, $attribs, $selfclose ); |
|
2869
|
|
|
} |
|
2870
|
|
|
return true; |
|
2871
|
|
|
case 'body': |
|
2872
|
|
|
case 'col': |
|
2873
|
|
|
case 'colgroup': |
|
2874
|
|
|
# OMITTED: <html> |
|
2875
|
|
|
case 'tbody': |
|
2876
|
|
|
case 'td': |
|
2877
|
|
|
case 'tfoot': |
|
2878
|
|
|
case 'th': |
|
2879
|
|
|
case 'thead': |
|
2880
|
|
|
case 'tr': |
|
2881
|
|
|
// Ignore the token |
|
2882
|
|
|
return true; |
|
2883
|
|
|
} |
|
2884
|
|
|
// Fall through to "anything else" case. |
|
2885
|
|
|
} |
|
2886
|
|
|
// The Anything Else case |
|
2887
|
|
|
return $this->inBodyMode( $token, $value, $attribs, $selfclose ); |
|
2888
|
|
|
} |
|
2889
|
|
|
|
|
2890
|
|
|
private function inColumnGroupMode( $token, $value, $attribs = null, $selfclose = false ) { |
|
2891
|
|
|
if ( $token === 'text' ) { |
|
2892
|
|
View Code Duplication |
if ( preg_match( '/^[\x09\x0A\x0C\x0D\x20]+/', $value, $matches ) ) { |
|
2893
|
|
|
$this->stack->insertText( $matches[0] ); |
|
2894
|
|
|
$value = substr( $value, strlen( $matches[0] ) ); |
|
2895
|
|
|
} |
|
2896
|
|
|
if ( strlen( $value ) === 0 ) { |
|
2897
|
|
|
return true; // All text handled. |
|
2898
|
|
|
} |
|
2899
|
|
|
// Fall through to handle non-whitespace below. |
|
2900
|
|
|
} elseif ( $token === 'tag' ) { |
|
2901
|
|
|
switch ( $value ) { |
|
2902
|
|
|
# OMITTED: <html> |
|
2903
|
|
|
case 'col': |
|
2904
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2905
|
|
|
$this->stack->pop(); |
|
2906
|
|
|
return true; |
|
2907
|
|
|
case 'template': |
|
2908
|
|
|
return $this->inHeadMode( $token, $value, $attribs, $selfclose ); |
|
2909
|
|
|
} |
|
2910
|
|
|
// Fall through for "anything else". |
|
2911
|
|
|
} elseif ( $token === 'endtag' ) { |
|
2912
|
|
|
switch ( $value ) { |
|
2913
|
|
|
case 'colgroup': |
|
2914
|
|
|
if ( !$this->stack->currentNode()->isA( 'colgroup' ) ) { |
|
2915
|
|
|
return true; // Ignore the token. |
|
2916
|
|
|
} |
|
2917
|
|
|
$this->stack->pop(); |
|
2918
|
|
|
$this->switchMode( 'inTableMode' ); |
|
2919
|
|
|
return true; |
|
2920
|
|
|
case 'col': |
|
2921
|
|
|
return true; // Ignore the token. |
|
2922
|
|
|
case 'template': |
|
2923
|
|
|
return $this->inHeadMode( $token, $value, $attribs, $selfclose ); |
|
2924
|
|
|
} |
|
2925
|
|
|
// Fall through for "anything else". |
|
2926
|
|
|
} elseif ( $token === 'eof' ) { |
|
2927
|
|
|
return $this->inBodyMode( $token, $value, $attribs, $selfclose ); |
|
2928
|
|
|
} |
|
2929
|
|
|
|
|
2930
|
|
|
// Anything else |
|
2931
|
|
|
if ( !$this->stack->currentNode()->isA( 'colgroup' ) ) { |
|
2932
|
|
|
return true; // Ignore the token. |
|
2933
|
|
|
} |
|
2934
|
|
|
$this->inColumnGroupMode( 'endtag', 'colgroup' ); |
|
2935
|
|
|
return $this->insertToken( $token, $value, $attribs, $selfclose ); |
|
2936
|
|
|
} |
|
2937
|
|
|
|
|
2938
|
|
|
// Helper function for inTableBodyMode |
|
2939
|
|
|
private function endSection() { |
|
2940
|
|
|
if ( !( |
|
2941
|
|
|
$this->stack->inTableScope( 'tbody' ) || |
|
2942
|
|
|
$this->stack->inTableScope( 'thead' ) || |
|
2943
|
|
|
$this->stack->inTableScope( 'tfoot' ) |
|
2944
|
|
|
) ) { |
|
2945
|
|
|
return false; |
|
2946
|
|
|
} |
|
2947
|
|
|
$this->stack->clearToContext( BalanceSets::$tableBodyContextSet ); |
|
2948
|
|
|
$this->stack->pop(); |
|
2949
|
|
|
$this->switchMode( 'inTableMode' ); |
|
2950
|
|
|
return true; |
|
2951
|
|
|
} |
|
2952
|
|
View Code Duplication |
private function inTableBodyMode( $token, $value, $attribs = null, $selfclose = false ) { |
|
2953
|
|
|
if ( $token === 'tag' ) { |
|
2954
|
|
|
switch ( $value ) { |
|
2955
|
|
|
case 'tr': |
|
2956
|
|
|
$this->stack->clearToContext( BalanceSets::$tableBodyContextSet ); |
|
2957
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
2958
|
|
|
$this->switchMode( 'inRowMode' ); |
|
2959
|
|
|
return true; |
|
2960
|
|
|
case 'th': |
|
2961
|
|
|
case 'td': |
|
2962
|
|
|
$this->inTableBodyMode( 'tag', 'tr', [] ); |
|
2963
|
|
|
$this->insertToken( $token, $value, $attribs, $selfclose ); |
|
2964
|
|
|
return true; |
|
2965
|
|
|
case 'caption': |
|
2966
|
|
|
case 'col': |
|
2967
|
|
|
case 'colgroup': |
|
2968
|
|
|
case 'tbody': |
|
2969
|
|
|
case 'tfoot': |
|
2970
|
|
|
case 'thead': |
|
2971
|
|
|
if ( $this->endSection() ) { |
|
2972
|
|
|
$this->insertToken( $token, $value, $attribs, $selfclose ); |
|
2973
|
|
|
} |
|
2974
|
|
|
return true; |
|
2975
|
|
|
} |
|
2976
|
|
|
} elseif ( $token === 'endtag' ) { |
|
2977
|
|
|
switch ( $value ) { |
|
2978
|
|
|
case 'table': |
|
2979
|
|
|
if ( $this->endSection() ) { |
|
2980
|
|
|
$this->insertToken( $token, $value, $attribs, $selfclose ); |
|
2981
|
|
|
} |
|
2982
|
|
|
return true; |
|
2983
|
|
|
case 'tbody': |
|
2984
|
|
|
case 'tfoot': |
|
2985
|
|
|
case 'thead': |
|
2986
|
|
|
if ( $this->stack->inTableScope( $value ) ) { |
|
2987
|
|
|
$this->endSection(); |
|
2988
|
|
|
} |
|
2989
|
|
|
return true; |
|
2990
|
|
|
# OMITTED: <body> |
|
2991
|
|
|
case 'caption': |
|
2992
|
|
|
case 'col': |
|
2993
|
|
|
case 'colgroup': |
|
2994
|
|
|
# OMITTED: <html> |
|
2995
|
|
|
case 'td': |
|
2996
|
|
|
case 'th': |
|
2997
|
|
|
case 'tr': |
|
2998
|
|
|
return true; // Ignore the token. |
|
2999
|
|
|
} |
|
3000
|
|
|
} |
|
3001
|
|
|
// Anything else: |
|
3002
|
|
|
return $this->inTableMode( $token, $value, $attribs, $selfclose ); |
|
3003
|
|
|
} |
|
3004
|
|
|
|
|
3005
|
|
|
// Helper function for inRowMode |
|
3006
|
|
|
private function endRow() { |
|
3007
|
|
|
if ( !$this->stack->inTableScope( 'tr' ) ) { |
|
3008
|
|
|
return false; |
|
3009
|
|
|
} |
|
3010
|
|
|
$this->stack->clearToContext( BalanceSets::$tableRowContextSet ); |
|
3011
|
|
|
$this->stack->pop(); |
|
3012
|
|
|
$this->switchMode( 'inTableBodyMode' ); |
|
3013
|
|
|
return true; |
|
3014
|
|
|
} |
|
3015
|
|
View Code Duplication |
private function inRowMode( $token, $value, $attribs = null, $selfclose = false ) { |
|
3016
|
|
|
if ( $token === 'tag' ) { |
|
3017
|
|
|
switch ( $value ) { |
|
3018
|
|
|
case 'th': |
|
3019
|
|
|
case 'td': |
|
3020
|
|
|
$this->stack->clearToContext( BalanceSets::$tableRowContextSet ); |
|
3021
|
|
|
$this->stack->insertHTMLElement( $value, $attribs ); |
|
3022
|
|
|
$this->switchMode( 'inCellMode' ); |
|
3023
|
|
|
$this->afe->insertMarker(); |
|
3024
|
|
|
return true; |
|
3025
|
|
|
case 'caption': |
|
3026
|
|
|
case 'col': |
|
3027
|
|
|
case 'colgroup': |
|
3028
|
|
|
case 'tbody': |
|
3029
|
|
|
case 'tfoot': |
|
3030
|
|
|
case 'thead': |
|
3031
|
|
|
case 'tr': |
|
3032
|
|
|
if ( $this->endRow() ) { |
|
3033
|
|
|
$this->insertToken( $token, $value, $attribs, $selfclose ); |
|
3034
|
|
|
} |
|
3035
|
|
|
return true; |
|
3036
|
|
|
} |
|
3037
|
|
|
} elseif ( $token === 'endtag' ) { |
|
3038
|
|
|
switch ( $value ) { |
|
3039
|
|
|
case 'tr': |
|
3040
|
|
|
$this->endRow(); |
|
3041
|
|
|
return true; |
|
3042
|
|
|
case 'table': |
|
3043
|
|
|
if ( $this->endRow() ) { |
|
3044
|
|
|
$this->insertToken( $token, $value, $attribs, $selfclose ); |
|
3045
|
|
|
} |
|
3046
|
|
|
return true; |
|
3047
|
|
|
case 'tbody': |
|
3048
|
|
|
case 'tfoot': |
|
3049
|
|
|
case 'thead': |
|
3050
|
|
|
if ( |
|
3051
|
|
|
$this->stack->inTableScope( $value ) && |
|
3052
|
|
|
$this->endRow() |
|
3053
|
|
|
) { |
|
3054
|
|
|
$this->insertToken( $token, $value, $attribs, $selfclose ); |
|
3055
|
|
|
} |
|
3056
|
|
|
return true; |
|
3057
|
|
|
# OMITTED: <body> |
|
3058
|
|
|
case 'caption': |
|
3059
|
|
|
case 'col': |
|
3060
|
|
|
case 'colgroup': |
|
3061
|
|
|
# OMITTED: <html> |
|
3062
|
|
|
case 'td': |
|
3063
|
|
|
case 'th': |
|
3064
|
|
|
return true; // Ignore the token. |
|
3065
|
|
|
} |
|
3066
|
|
|
} |
|
3067
|
|
|
// Anything else: |
|
3068
|
|
|
return $this->inTableMode( $token, $value, $attribs, $selfclose ); |
|
3069
|
|
|
} |
|
3070
|
|
|
|
|
3071
|
|
|
// Helper for inCellMode |
|
3072
|
|
|
private function endCell() { |
|
3073
|
|
|
if ( $this->stack->inTableScope( 'td' ) ) { |
|
3074
|
|
|
$this->inCellMode( 'endtag', 'td' ); |
|
3075
|
|
|
return true; |
|
3076
|
|
|
} elseif ( $this->stack->inTableScope( 'th' ) ) { |
|
3077
|
|
|
$this->inCellMode( 'endtag', 'th' ); |
|
3078
|
|
|
return true; |
|
3079
|
|
|
} else { |
|
3080
|
|
|
return false; |
|
3081
|
|
|
} |
|
3082
|
|
|
} |
|
3083
|
|
|
private function inCellMode( $token, $value, $attribs = null, $selfclose = false ) { |
|
3084
|
|
|
if ( $token === 'tag' ) { |
|
3085
|
|
View Code Duplication |
switch ( $value ) { |
|
3086
|
|
|
case 'caption': |
|
3087
|
|
|
case 'col': |
|
3088
|
|
|
case 'colgroup': |
|
3089
|
|
|
case 'tbody': |
|
3090
|
|
|
case 'td': |
|
3091
|
|
|
case 'tfoot': |
|
3092
|
|
|
case 'th': |
|
3093
|
|
|
case 'thead': |
|
3094
|
|
|
case 'tr': |
|
3095
|
|
|
if ( $this->endCell() ) { |
|
3096
|
|
|
$this->insertToken( $token, $value, $attribs, $selfclose ); |
|
3097
|
|
|
} |
|
3098
|
|
|
return true; |
|
3099
|
|
|
} |
|
3100
|
|
|
} elseif ( $token === 'endtag' ) { |
|
3101
|
|
|
switch ( $value ) { |
|
3102
|
|
|
case 'td': |
|
3103
|
|
View Code Duplication |
case 'th': |
|
3104
|
|
|
if ( $this->stack->inTableScope( $value ) ) { |
|
3105
|
|
|
$this->stack->generateImpliedEndTags(); |
|
3106
|
|
|
$this->stack->popTag( $value ); |
|
3107
|
|
|
$this->afe->clearToMarker(); |
|
3108
|
|
|
$this->switchMode( 'inRowMode' ); |
|
3109
|
|
|
} |
|
3110
|
|
|
return true; |
|
3111
|
|
|
# OMITTED: <body> |
|
3112
|
|
|
case 'caption': |
|
3113
|
|
|
case 'col': |
|
3114
|
|
|
case 'colgroup': |
|
3115
|
|
|
# OMITTED: <html> |
|
3116
|
|
|
return true; |
|
3117
|
|
|
|
|
3118
|
|
|
case 'table': |
|
3119
|
|
|
case 'tbody': |
|
3120
|
|
|
case 'tfoot': |
|
3121
|
|
|
case 'thead': |
|
3122
|
|
|
case 'tr': |
|
3123
|
|
|
if ( $this->stack->inTableScope( $value ) ) { |
|
3124
|
|
|
$this->stack->generateImpliedEndTags(); |
|
3125
|
|
|
$this->stack->popTag( BalanceSets::$tableCellSet ); |
|
3126
|
|
|
$this->afe->clearToMarker(); |
|
3127
|
|
|
$this->switchMode( 'inRowMode' ); |
|
3128
|
|
|
$this->insertToken( $token, $value, $attribs, $selfclose ); |
|
3129
|
|
|
} |
|
3130
|
|
|
return true; |
|
3131
|
|
|
} |
|
3132
|
|
|
} |
|
3133
|
|
|
// Anything else: |
|
3134
|
|
|
return $this->inBodyMode( $token, $value, $attribs, $selfclose ); |
|
3135
|
|
|
} |
|
3136
|
|
|
|
|
3137
|
|
|
# OMITTED: <select> |
|
3138
|
|
|
/* |
|
3139
|
|
|
private function inSelectMode( $token, $value, $attribs = null, $selfclose = false ) { |
|
3140
|
|
|
Assert::invariant( false, 'Unimplemented' ); |
|
3141
|
|
|
} |
|
3142
|
|
|
|
|
3143
|
|
|
private function inSelectInTableMode( $token, $value, $attribs = null, $selfclose = false ) { |
|
3144
|
|
|
Assert::invariant( false, 'Unimplemented' ); |
|
3145
|
|
|
} |
|
3146
|
|
|
*/ |
|
3147
|
|
|
|
|
3148
|
|
|
private function inTemplateMode( $token, $value, $attribs = null, $selfclose = false ) { |
|
3149
|
|
|
if ( $token === 'text' ) { |
|
3150
|
|
|
return $this->inBodyMode( $token, $value, $attribs, $selfclose ); |
|
3151
|
|
|
} elseif ( $token === 'eof' ) { |
|
3152
|
|
|
if ( $this->stack->indexOf( 'template' ) < 0 ) { |
|
3153
|
|
|
$this->stopParsing(); |
|
3154
|
|
|
} else { |
|
3155
|
|
|
$this->stack->popTag( 'template' ); |
|
3156
|
|
|
$this->afe->clearToMarker(); |
|
3157
|
|
|
array_pop( $this->templateInsertionModes ); |
|
3158
|
|
|
$this->resetInsertionMode(); |
|
3159
|
|
|
$this->insertToken( $token, $value, $attribs, $selfclose ); |
|
3160
|
|
|
} |
|
3161
|
|
|
return true; |
|
3162
|
|
|
} elseif ( $token === 'tag' ) { |
|
3163
|
|
|
switch ( $value ) { |
|
3164
|
|
|
case 'base': |
|
3165
|
|
|
case 'basefont': |
|
3166
|
|
|
case 'bgsound': |
|
3167
|
|
|
case 'link': |
|
3168
|
|
|
case 'meta': |
|
3169
|
|
|
case 'noframes': |
|
3170
|
|
|
# OMITTED: <script> |
|
3171
|
|
|
case 'style': |
|
3172
|
|
|
case 'template': |
|
3173
|
|
|
# OMITTED: <title> |
|
3174
|
|
|
return $this->inHeadMode( $token, $value, $attribs, $selfclose ); |
|
3175
|
|
|
|
|
3176
|
|
|
case 'caption': |
|
3177
|
|
|
case 'colgroup': |
|
3178
|
|
|
case 'tbody': |
|
3179
|
|
|
case 'tfoot': |
|
3180
|
|
|
case 'thead': |
|
3181
|
|
|
return $this->switchModeAndReprocess( |
|
3182
|
|
|
'inTableMode', $token, $value, $attribs, $selfclose |
|
3183
|
|
|
); |
|
3184
|
|
|
|
|
3185
|
|
|
case 'col': |
|
3186
|
|
|
return $this->switchModeAndReprocess( |
|
3187
|
|
|
'inColumnGroupMode', $token, $value, $attribs, $selfclose |
|
3188
|
|
|
); |
|
3189
|
|
|
|
|
3190
|
|
|
case 'tr': |
|
3191
|
|
|
return $this->switchModeAndReprocess( |
|
3192
|
|
|
'inTableBodyMode', $token, $value, $attribs, $selfclose |
|
3193
|
|
|
); |
|
3194
|
|
|
|
|
3195
|
|
|
case 'td': |
|
3196
|
|
|
case 'th': |
|
3197
|
|
|
return $this->switchModeAndReprocess( |
|
3198
|
|
|
'inRowMode', $token, $value, $attribs, $selfclose |
|
3199
|
|
|
); |
|
3200
|
|
|
} |
|
3201
|
|
|
return $this->switchModeAndReprocess( |
|
3202
|
|
|
'inBodyMode', $token, $value, $attribs, $selfclose |
|
3203
|
|
|
); |
|
3204
|
|
|
} elseif ( $token === 'endtag' ) { |
|
3205
|
|
|
switch ( $value ) { |
|
3206
|
|
|
case 'template': |
|
3207
|
|
|
return $this->inHeadMode( $token, $value, $attribs, $selfclose ); |
|
3208
|
|
|
} |
|
3209
|
|
|
return true; |
|
3210
|
|
|
} else { |
|
3211
|
|
|
Assert::invariant( false, "Bad token type: $token" ); |
|
3212
|
|
|
} |
|
3213
|
|
|
} |
|
3214
|
|
|
} |
|
3215
|
|
|
|
In PHP it is possible to write to properties without declaring them. For example, the following is perfectly valid PHP code:
Generally, it is a good practice to explictly declare properties to avoid accidental typos and provide IDE auto-completion: