1
|
|
|
<?php |
2
|
|
|
|
3
|
|
|
/** |
4
|
|
|
* Create wrapper P and BR elements in HTML depending on newlines. Useful when |
5
|
|
|
* users use newlines to signal line and paragraph breaks. In all cases output |
6
|
|
|
* should be well-formed markup. |
7
|
|
|
* |
8
|
|
|
* In DIV elements, Ps are only added when there would be at |
9
|
|
|
* least two of them. |
10
|
|
|
* |
11
|
|
|
* @package Elgg.Core |
12
|
|
|
* @subpackage Output |
13
|
|
|
*/ |
14
|
|
|
class ElggAutoP { |
15
|
|
|
|
16
|
|
|
public $encoding = 'UTF-8'; |
17
|
|
|
|
18
|
|
|
/** |
19
|
|
|
* @var DOMDocument |
20
|
|
|
*/ |
21
|
|
|
protected $_doc = null; |
22
|
|
|
|
23
|
|
|
/** |
24
|
|
|
* @var DOMXPath |
25
|
|
|
*/ |
26
|
|
|
protected $_xpath = null; |
27
|
|
|
|
28
|
|
|
protected $_blocks = 'address article area aside blockquote caption col colgroup dd |
29
|
|
|
details div dl dt fieldset figure figcaption footer form h1 h2 h3 h4 h5 h6 header |
30
|
|
|
hr hgroup legend map math menu nav noscript p pre section select style summary |
31
|
|
|
table tbody td tfoot th thead tr ul ol option li'; |
32
|
|
|
|
33
|
|
|
/** |
34
|
|
|
* @var array |
35
|
|
|
*/ |
36
|
|
|
protected $_inlines = 'a abbr audio b button canvas caption cite code command datalist |
37
|
|
|
del dfn em embed i iframe img input ins kbd keygen label map mark meter object |
38
|
|
|
output progress q rp rt ruby s samp script select small source span strong style |
39
|
|
|
sub sup textarea time var video wbr'; |
40
|
|
|
|
41
|
|
|
/** |
42
|
|
|
* Descend into these elements to add Ps |
43
|
|
|
* |
44
|
|
|
* @var array |
45
|
|
|
*/ |
46
|
|
|
protected $_descendList = 'article aside blockquote body details div footer form |
47
|
|
|
header section'; |
48
|
|
|
|
49
|
|
|
/** |
50
|
|
|
* Add Ps inside these elements |
51
|
|
|
* |
52
|
|
|
* @var array |
53
|
|
|
*/ |
54
|
|
|
protected $_alterList = 'article aside blockquote body details div footer header |
55
|
|
|
section'; |
56
|
|
|
|
57
|
|
|
/** @var string */ |
58
|
|
|
protected $_unique = ''; |
59
|
|
|
|
60
|
|
|
/** |
61
|
|
|
* Constructor |
62
|
|
|
*/ |
63
|
7 |
|
public function __construct() { |
64
|
7 |
|
$this->_blocks = preg_split('@\\s+@', $this->_blocks); |
|
|
|
|
65
|
7 |
|
$this->_descendList = preg_split('@\\s+@', $this->_descendList); |
66
|
7 |
|
$this->_alterList = preg_split('@\\s+@', $this->_alterList); |
67
|
7 |
|
$this->_inlines = preg_split('@\\s+@', $this->_inlines); |
68
|
7 |
|
$this->_unique = md5(__FILE__); |
69
|
7 |
|
} |
70
|
|
|
|
71
|
|
|
/** |
72
|
|
|
* Create wrapper P and BR elements in HTML depending on newlines. Useful when |
73
|
|
|
* users use newlines to signal line and paragraph breaks. In all cases output |
74
|
|
|
* should be well-formed markup. |
75
|
|
|
* |
76
|
|
|
* In DIV, LI, TD, and TH elements, Ps are only added when their would be at |
77
|
|
|
* least two of them. |
78
|
|
|
* |
79
|
|
|
* @param string $html snippet |
80
|
|
|
* @return string|false output or false if parse error occurred |
81
|
|
|
*/ |
82
|
5 |
|
public function process($html) { |
83
|
|
|
// normalize whitespace |
84
|
5 |
|
$html = str_replace(array("\r\n", "\r"), "\n", $html); |
85
|
|
|
|
86
|
|
|
// allows preserving entities untouched |
87
|
5 |
|
$html = str_replace('&', $this->_unique . 'AMP', $html); |
88
|
|
|
|
89
|
5 |
|
$this->_doc = new DOMDocument(); |
90
|
|
|
|
91
|
|
|
// parse to DOM, suppressing loadHTML warnings |
92
|
|
|
// http://www.php.net/manual/en/domdocument.loadhtml.php#95463 |
93
|
5 |
|
libxml_use_internal_errors(true); |
94
|
|
|
|
95
|
|
|
// Do not load entities. May be unnecessary, better safe than sorry |
96
|
5 |
|
$disable_load_entities = libxml_disable_entity_loader(true); |
97
|
|
|
|
98
|
5 |
|
if (!$this->_doc->loadHTML("<html><meta http-equiv='content-type' " |
99
|
5 |
|
. "content='text/html; charset={$this->encoding}'><body>{$html}</body>" |
100
|
5 |
|
. "</html>")) { |
101
|
|
|
|
102
|
|
|
libxml_disable_entity_loader($disable_load_entities); |
103
|
|
|
return false; |
104
|
|
|
} |
105
|
|
|
|
106
|
5 |
|
libxml_disable_entity_loader($disable_load_entities); |
107
|
|
|
|
108
|
5 |
|
$this->_xpath = new DOMXPath($this->_doc); |
109
|
|
|
// start processing recursively at the BODY element |
110
|
5 |
|
$nodeList = $this->_xpath->query('//body[1]'); |
111
|
5 |
|
$this->addParagraphs($nodeList->item(0)); |
|
|
|
|
112
|
|
|
|
113
|
|
|
// serialize back to HTML |
114
|
5 |
|
$html = $this->_doc->saveHTML(); |
115
|
|
|
|
116
|
|
|
// Note: we create <autop> elements, which will later be converted to paragraphs |
117
|
|
|
|
118
|
|
|
// split AUTOPs into multiples at /\n\n+/ |
119
|
5 |
|
$html = preg_replace('/(' . $this->_unique . 'NL){2,}/', '</autop><autop>', $html); |
120
|
5 |
|
$html = str_replace(array($this->_unique . 'BR', $this->_unique . 'NL', '<br>'), |
121
|
5 |
|
'<br />', |
122
|
5 |
|
$html); |
123
|
5 |
|
$html = str_replace('<br /></autop>', '</autop>', $html); |
124
|
|
|
|
125
|
|
|
// re-parse so we can handle new AUTOP elements |
126
|
|
|
|
127
|
|
|
// Do not load entities. May be unnecessary, better safe than sorry |
128
|
5 |
|
$disable_load_entities = libxml_disable_entity_loader(true); |
129
|
|
|
|
130
|
5 |
|
if (!$this->_doc->loadHTML($html)) { |
131
|
|
|
libxml_disable_entity_loader($disable_load_entities); |
132
|
|
|
return false; |
133
|
|
|
} |
134
|
|
|
|
135
|
5 |
|
libxml_disable_entity_loader($disable_load_entities); |
136
|
|
|
|
137
|
|
|
// must re-create XPath object after DOM load |
138
|
5 |
|
$this->_xpath = new DOMXPath($this->_doc); |
139
|
|
|
|
140
|
|
|
// strip AUTOPs that only have comments/whitespace |
141
|
5 |
|
foreach ($this->_xpath->query('//autop') as $autop) { |
142
|
|
|
/* @var DOMElement $autop */ |
143
|
5 |
|
$hasContent = false; |
144
|
5 |
|
if (trim($autop->textContent) !== '') { |
145
|
5 |
|
$hasContent = true; |
146
|
5 |
|
} else { |
147
|
2 |
|
foreach ($autop->childNodes as $node) { |
148
|
2 |
|
if ($node->nodeType === XML_ELEMENT_NODE) { |
149
|
2 |
|
$hasContent = true; |
150
|
2 |
|
break; |
151
|
|
|
} |
152
|
2 |
|
} |
153
|
|
|
} |
154
|
5 |
|
if (!$hasContent) { |
155
|
|
|
// mark to be later replaced w/ preg_replace (faster than moving nodes out) |
156
|
2 |
|
$autop->setAttribute("r", "1"); |
157
|
2 |
|
} |
158
|
5 |
|
} |
159
|
|
|
|
160
|
|
|
// If a DIV contains a single AUTOP, remove it |
161
|
5 |
|
foreach ($this->_xpath->query('//div') as $el) { |
162
|
|
|
/* @var DOMElement $el */ |
163
|
1 |
|
$autops = $this->_xpath->query('./autop', $el); |
164
|
1 |
|
if ($autops->length === 1) { |
165
|
1 |
|
$firstAutop = $autops->item(0); |
166
|
|
|
/* @var DOMElement $firstAutop */ |
167
|
1 |
|
$firstAutop->setAttribute("r", "1"); |
168
|
1 |
|
} |
169
|
5 |
|
} |
170
|
|
|
|
171
|
5 |
|
$html = $this->_doc->saveHTML(); |
172
|
|
|
|
173
|
|
|
// trim to the contents of BODY |
174
|
5 |
|
$bodyStart = strpos($html, '<body>'); |
175
|
5 |
|
$bodyEnd = strpos($html, '</body>', $bodyStart + 6); |
176
|
5 |
|
$html = substr($html, $bodyStart + 6, $bodyEnd - $bodyStart - 6); |
177
|
|
|
|
178
|
|
|
// strip AUTOPs that should be removed |
179
|
5 |
|
$html = preg_replace('@<autop r="1">(.*?)</autop>@', '\\1', $html); |
180
|
|
|
|
181
|
|
|
// commit to converting AUTOPs to Ps |
182
|
5 |
|
$html = str_replace('<autop>', "\n<p>", $html); |
183
|
5 |
|
$html = str_replace('</autop>', "</p>\n", $html); |
184
|
|
|
|
185
|
5 |
|
$html = str_replace('<br>', '<br />', $html); |
186
|
5 |
|
$html = str_replace($this->_unique . 'AMP', '&', $html); |
187
|
5 |
|
return $html; |
188
|
|
|
} |
189
|
|
|
|
190
|
|
|
/** |
191
|
|
|
* Add P and BR elements as necessary |
192
|
|
|
* |
193
|
|
|
* @param DOMElement $el DOM element |
194
|
|
|
* @return void |
195
|
|
|
*/ |
196
|
5 |
|
protected function addParagraphs(DOMElement $el) { |
197
|
|
|
// no need to call recursively, just queue up |
198
|
5 |
|
$elsToProcess = array($el); |
199
|
5 |
|
$inlinesToProcess = array(); |
200
|
5 |
|
while ($el = array_shift($elsToProcess)) { |
201
|
|
|
// if true, we can alter all child nodes, if not, we'll just call |
202
|
|
|
// addParagraphs on each element in the descendInto list |
203
|
5 |
|
$alterInline = in_array($el->nodeName, $this->_alterList); |
204
|
|
|
|
205
|
|
|
// inside affected elements, we want to trim leading whitespace from |
206
|
|
|
// the first text node |
207
|
5 |
|
$ltrimFirstTextNode = true; |
208
|
|
|
|
209
|
|
|
// should we open a new AUTOP element to move inline elements into? |
210
|
5 |
|
$openP = true; |
211
|
5 |
|
$autop = null; |
212
|
|
|
|
213
|
|
|
// after BR, ignore a newline |
214
|
5 |
|
$isFollowingBr = false; |
215
|
|
|
|
216
|
5 |
|
$node = $el->firstChild; |
217
|
5 |
|
while (null !== $node) { |
218
|
5 |
|
if ($alterInline) { |
219
|
5 |
|
if ($openP) { |
220
|
5 |
|
$openP = false; |
221
|
|
|
// create a P to move inline content into (this may be removed later) |
222
|
5 |
|
$autop = $el->insertBefore($this->_doc->createElement('autop'), $node); |
223
|
5 |
|
} |
224
|
5 |
|
} |
225
|
|
|
|
226
|
5 |
|
$isElement = ($node->nodeType === XML_ELEMENT_NODE); |
227
|
5 |
|
if ($isElement) { |
228
|
5 |
|
$isBlock = in_array($node->nodeName, $this->_blocks); |
229
|
5 |
|
if (!$isBlock) { |
230
|
|
|
// if we start with an inline element we don't need to do this |
231
|
5 |
|
$ltrimFirstTextNode = false; |
232
|
5 |
|
} |
233
|
5 |
|
} else { |
234
|
5 |
|
$isBlock = false; |
235
|
|
|
} |
236
|
|
|
|
237
|
5 |
|
if ($alterInline) { |
238
|
5 |
|
$isText = ($node->nodeType === XML_TEXT_NODE); |
239
|
5 |
|
$isLastInline = (! $node->nextSibling |
240
|
5 |
|
|| ($node->nextSibling->nodeType === XML_ELEMENT_NODE |
241
|
5 |
|
&& in_array($node->nextSibling->nodeName, $this->_blocks))); |
242
|
5 |
|
if ($isElement) { |
243
|
5 |
|
$isFollowingBr = ($node->nodeName === 'br'); |
244
|
5 |
|
} |
245
|
|
|
|
246
|
5 |
|
if ($isText) { |
247
|
5 |
|
$nodeText = $node->nodeValue; |
248
|
|
|
|
249
|
5 |
|
if ($ltrimFirstTextNode) { |
250
|
|
|
// we're at the beginning of a sequence of text/inline elements |
251
|
4 |
|
$nodeText = ltrim($nodeText); |
252
|
4 |
|
$ltrimFirstTextNode = false; |
253
|
4 |
|
} |
254
|
5 |
|
if ($isFollowingBr && preg_match('@^[ \\t]*\\n[ \\t]*@', $nodeText, $m)) { |
255
|
|
|
// if a user ends a line with <br>, don't add a second BR |
256
|
1 |
|
$nodeText = substr($nodeText, strlen($m[0])); |
257
|
1 |
|
} |
258
|
5 |
|
if ($isLastInline) { |
259
|
|
|
// we're at the end of a sequence of text/inline elements |
260
|
4 |
|
$nodeText = rtrim($nodeText); |
261
|
4 |
|
} |
262
|
5 |
|
$nodeText = str_replace("\n", $this->_unique . 'NL', $nodeText); |
263
|
5 |
|
$tmpNode = $node; |
264
|
5 |
|
$node = $node->nextSibling; // move loop to next node |
265
|
|
|
|
266
|
|
|
// alter node in place, then move into AUTOP |
267
|
5 |
|
$tmpNode->nodeValue = $nodeText; |
268
|
5 |
|
$autop->appendChild($tmpNode); |
269
|
|
|
|
270
|
5 |
|
continue; |
271
|
|
|
} |
272
|
5 |
|
} |
273
|
5 |
|
if ($isBlock || ! $node->nextSibling) { |
274
|
5 |
|
if ($isBlock) { |
275
|
4 |
|
if (in_array($node->nodeName, $this->_descendList)) { |
276
|
3 |
|
$elsToProcess[] = $node; |
277
|
|
|
//$this->addParagraphs($node); |
278
|
3 |
|
} |
279
|
4 |
|
} |
280
|
5 |
|
$openP = true; |
281
|
5 |
|
$ltrimFirstTextNode = true; |
282
|
5 |
|
} |
283
|
5 |
|
if ($alterInline) { |
284
|
5 |
|
if (! $isBlock) { |
285
|
5 |
|
$tmpNode = $node; |
286
|
5 |
|
if ($isElement && false !== strpos($tmpNode->textContent, "\n")) { |
287
|
1 |
|
$inlinesToProcess[] = $tmpNode; |
288
|
1 |
|
} |
289
|
5 |
|
$node = $node->nextSibling; |
290
|
5 |
|
$autop->appendChild($tmpNode); |
291
|
5 |
|
continue; |
292
|
|
|
} |
293
|
4 |
|
} |
294
|
|
|
|
295
|
4 |
|
$node = $node->nextSibling; |
296
|
4 |
|
} |
297
|
5 |
|
} |
298
|
|
|
|
299
|
|
|
// handle inline nodes |
300
|
|
|
// no need to recurse, just queue up |
301
|
5 |
|
while ($el = array_shift($inlinesToProcess)) { |
302
|
1 |
|
$ignoreLeadingNewline = false; |
303
|
1 |
|
foreach ($el->childNodes as $node) { |
304
|
1 |
|
if ($node->nodeType === XML_ELEMENT_NODE) { |
305
|
|
|
if ($node->nodeValue === 'BR') { |
306
|
|
|
$ignoreLeadingNewline = true; |
307
|
|
|
} else { |
308
|
|
|
$ignoreLeadingNewline = false; |
309
|
|
|
if (false !== strpos($node->textContent, "\n")) { |
310
|
|
|
$inlinesToProcess[] = $node; |
311
|
|
|
} |
312
|
|
|
} |
313
|
|
|
continue; |
314
|
1 |
|
} elseif ($node->nodeType === XML_TEXT_NODE) { |
315
|
1 |
|
$text = $node->nodeValue; |
316
|
1 |
|
if ($text[0] === "\n" && $ignoreLeadingNewline) { |
317
|
|
|
$text = substr($text, 1); |
318
|
|
|
$ignoreLeadingNewline = false; |
319
|
|
|
} |
320
|
1 |
|
$node->nodeValue = str_replace("\n", $this->_unique . 'BR', $text); |
321
|
1 |
|
} |
322
|
1 |
|
} |
323
|
1 |
|
} |
324
|
5 |
|
} |
325
|
|
|
} |
326
|
|
|
|
Our type inference engine has found an assignment to a property that is incompatible with the declared type of that property.
Either this assignment is in error or the assigned type should be added to the documentation/type hint for that property..