]>
Commit | Line | Data |
---|---|---|
d4949327 NL |
1 | <?php\r |
2 | \r | |
3 | /**\r | |
4 | * Definition of the purified HTML that describes allowed children,\r | |
5 | * attributes, and many other things.\r | |
6 | *\r | |
7 | * Conventions:\r | |
8 | *\r | |
9 | * All member variables that are prefixed with info\r | |
10 | * (including the main $info array) are used by HTML Purifier internals\r | |
11 | * and should not be directly edited when customizing the HTMLDefinition.\r | |
12 | * They can usually be set via configuration directives or custom\r | |
13 | * modules.\r | |
14 | *\r | |
15 | * On the other hand, member variables without the info prefix are used\r | |
16 | * internally by the HTMLDefinition and MUST NOT be used by other HTML\r | |
17 | * Purifier internals. Many of them, however, are public, and may be\r | |
18 | * edited by userspace code to tweak the behavior of HTMLDefinition.\r | |
19 | *\r | |
20 | * @note This class is inspected by Printer_HTMLDefinition; please\r | |
21 | * update that class if things here change.\r | |
22 | *\r | |
23 | * @warning Directives that change this object's structure must be in\r | |
24 | * the HTML or Attr namespace!\r | |
25 | */\r | |
26 | class HTMLPurifier_HTMLDefinition extends HTMLPurifier_Definition\r | |
27 | {\r | |
28 | \r | |
29 | // FULLY-PUBLIC VARIABLES ---------------------------------------------\r | |
30 | \r | |
31 | /**\r | |
32 | * Associative array of element names to HTMLPurifier_ElementDef.\r | |
33 | * @type HTMLPurifier_ElementDef[]\r | |
34 | */\r | |
35 | public $info = array();\r | |
36 | \r | |
37 | /**\r | |
38 | * Associative array of global attribute name to attribute definition.\r | |
39 | * @type array\r | |
40 | */\r | |
41 | public $info_global_attr = array();\r | |
42 | \r | |
43 | /**\r | |
44 | * String name of parent element HTML will be going into.\r | |
45 | * @type string\r | |
46 | */\r | |
47 | public $info_parent = 'div';\r | |
48 | \r | |
49 | /**\r | |
50 | * Definition for parent element, allows parent element to be a\r | |
51 | * tag that's not allowed inside the HTML fragment.\r | |
52 | * @type HTMLPurifier_ElementDef\r | |
53 | */\r | |
54 | public $info_parent_def;\r | |
55 | \r | |
56 | /**\r | |
57 | * String name of element used to wrap inline elements in block context.\r | |
58 | * @type string\r | |
59 | * @note This is rarely used except for BLOCKQUOTEs in strict mode\r | |
60 | */\r | |
61 | public $info_block_wrapper = 'p';\r | |
62 | \r | |
63 | /**\r | |
64 | * Associative array of deprecated tag name to HTMLPurifier_TagTransform.\r | |
65 | * @type array\r | |
66 | */\r | |
67 | public $info_tag_transform = array();\r | |
68 | \r | |
69 | /**\r | |
70 | * Indexed list of HTMLPurifier_AttrTransform to be performed before validation.\r | |
71 | * @type HTMLPurifier_AttrTransform[]\r | |
72 | */\r | |
73 | public $info_attr_transform_pre = array();\r | |
74 | \r | |
75 | /**\r | |
76 | * Indexed list of HTMLPurifier_AttrTransform to be performed after validation.\r | |
77 | * @type HTMLPurifier_AttrTransform[]\r | |
78 | */\r | |
79 | public $info_attr_transform_post = array();\r | |
80 | \r | |
81 | /**\r | |
82 | * Nested lookup array of content set name (Block, Inline) to\r | |
83 | * element name to whether or not it belongs in that content set.\r | |
84 | * @type array\r | |
85 | */\r | |
86 | public $info_content_sets = array();\r | |
87 | \r | |
88 | /**\r | |
89 | * Indexed list of HTMLPurifier_Injector to be used.\r | |
90 | * @type HTMLPurifier_Injector[]\r | |
91 | */\r | |
92 | public $info_injector = array();\r | |
93 | \r | |
94 | /**\r | |
95 | * Doctype object\r | |
96 | * @type HTMLPurifier_Doctype\r | |
97 | */\r | |
98 | public $doctype;\r | |
99 | \r | |
100 | \r | |
101 | \r | |
102 | // RAW CUSTOMIZATION STUFF --------------------------------------------\r | |
103 | \r | |
104 | /**\r | |
105 | * Adds a custom attribute to a pre-existing element\r | |
106 | * @note This is strictly convenience, and does not have a corresponding\r | |
107 | * method in HTMLPurifier_HTMLModule\r | |
108 | * @param string $element_name Element name to add attribute to\r | |
109 | * @param string $attr_name Name of attribute\r | |
110 | * @param mixed $def Attribute definition, can be string or object, see\r | |
111 | * HTMLPurifier_AttrTypes for details\r | |
112 | */\r | |
113 | public function addAttribute($element_name, $attr_name, $def)\r | |
114 | {\r | |
115 | $module = $this->getAnonymousModule();\r | |
116 | if (!isset($module->info[$element_name])) {\r | |
117 | $element = $module->addBlankElement($element_name);\r | |
118 | } else {\r | |
119 | $element = $module->info[$element_name];\r | |
120 | }\r | |
121 | $element->attr[$attr_name] = $def;\r | |
122 | }\r | |
123 | \r | |
124 | /**\r | |
125 | * Adds a custom element to your HTML definition\r | |
126 | * @see HTMLPurifier_HTMLModule::addElement() for detailed\r | |
127 | * parameter and return value descriptions.\r | |
128 | */\r | |
129 | public function addElement($element_name, $type, $contents, $attr_collections, $attributes = array())\r | |
130 | {\r | |
131 | $module = $this->getAnonymousModule();\r | |
132 | // assume that if the user is calling this, the element\r | |
133 | // is safe. This may not be a good idea\r | |
134 | $element = $module->addElement($element_name, $type, $contents, $attr_collections, $attributes);\r | |
135 | return $element;\r | |
136 | }\r | |
137 | \r | |
138 | /**\r | |
139 | * Adds a blank element to your HTML definition, for overriding\r | |
140 | * existing behavior\r | |
141 | * @param string $element_name\r | |
142 | * @return HTMLPurifier_ElementDef\r | |
143 | * @see HTMLPurifier_HTMLModule::addBlankElement() for detailed\r | |
144 | * parameter and return value descriptions.\r | |
145 | */\r | |
146 | public function addBlankElement($element_name)\r | |
147 | {\r | |
148 | $module = $this->getAnonymousModule();\r | |
149 | $element = $module->addBlankElement($element_name);\r | |
150 | return $element;\r | |
151 | }\r | |
152 | \r | |
153 | /**\r | |
154 | * Retrieves a reference to the anonymous module, so you can\r | |
155 | * bust out advanced features without having to make your own\r | |
156 | * module.\r | |
157 | * @return HTMLPurifier_HTMLModule\r | |
158 | */\r | |
159 | public function getAnonymousModule()\r | |
160 | {\r | |
161 | if (!$this->_anonModule) {\r | |
162 | $this->_anonModule = new HTMLPurifier_HTMLModule();\r | |
163 | $this->_anonModule->name = 'Anonymous';\r | |
164 | }\r | |
165 | return $this->_anonModule;\r | |
166 | }\r | |
167 | \r | |
168 | private $_anonModule = null;\r | |
169 | \r | |
170 | // PUBLIC BUT INTERNAL VARIABLES --------------------------------------\r | |
171 | \r | |
172 | /**\r | |
173 | * @type string\r | |
174 | */\r | |
175 | public $type = 'HTML';\r | |
176 | \r | |
177 | /**\r | |
178 | * @type HTMLPurifier_HTMLModuleManager\r | |
179 | */\r | |
180 | public $manager;\r | |
181 | \r | |
182 | /**\r | |
183 | * Performs low-cost, preliminary initialization.\r | |
184 | */\r | |
185 | public function __construct()\r | |
186 | {\r | |
187 | $this->manager = new HTMLPurifier_HTMLModuleManager();\r | |
188 | }\r | |
189 | \r | |
190 | /**\r | |
191 | * @param HTMLPurifier_Config $config\r | |
192 | */\r | |
193 | protected function doSetup($config)\r | |
194 | {\r | |
195 | $this->processModules($config);\r | |
196 | $this->setupConfigStuff($config);\r | |
197 | unset($this->manager);\r | |
198 | \r | |
199 | // cleanup some of the element definitions\r | |
200 | foreach ($this->info as $k => $v) {\r | |
201 | unset($this->info[$k]->content_model);\r | |
202 | unset($this->info[$k]->content_model_type);\r | |
203 | }\r | |
204 | }\r | |
205 | \r | |
206 | /**\r | |
207 | * Extract out the information from the manager\r | |
208 | * @param HTMLPurifier_Config $config\r | |
209 | */\r | |
210 | protected function processModules($config)\r | |
211 | {\r | |
212 | if ($this->_anonModule) {\r | |
213 | // for user specific changes\r | |
214 | // this is late-loaded so we don't have to deal with PHP4\r | |
215 | // reference wonky-ness\r | |
216 | $this->manager->addModule($this->_anonModule);\r | |
217 | unset($this->_anonModule);\r | |
218 | }\r | |
219 | \r | |
220 | $this->manager->setup($config);\r | |
221 | $this->doctype = $this->manager->doctype;\r | |
222 | \r | |
223 | foreach ($this->manager->modules as $module) {\r | |
224 | foreach ($module->info_tag_transform as $k => $v) {\r | |
225 | if ($v === false) {\r | |
226 | unset($this->info_tag_transform[$k]);\r | |
227 | } else {\r | |
228 | $this->info_tag_transform[$k] = $v;\r | |
229 | }\r | |
230 | }\r | |
231 | foreach ($module->info_attr_transform_pre as $k => $v) {\r | |
232 | if ($v === false) {\r | |
233 | unset($this->info_attr_transform_pre[$k]);\r | |
234 | } else {\r | |
235 | $this->info_attr_transform_pre[$k] = $v;\r | |
236 | }\r | |
237 | }\r | |
238 | foreach ($module->info_attr_transform_post as $k => $v) {\r | |
239 | if ($v === false) {\r | |
240 | unset($this->info_attr_transform_post[$k]);\r | |
241 | } else {\r | |
242 | $this->info_attr_transform_post[$k] = $v;\r | |
243 | }\r | |
244 | }\r | |
245 | foreach ($module->info_injector as $k => $v) {\r | |
246 | if ($v === false) {\r | |
247 | unset($this->info_injector[$k]);\r | |
248 | } else {\r | |
249 | $this->info_injector[$k] = $v;\r | |
250 | }\r | |
251 | }\r | |
252 | }\r | |
253 | $this->info = $this->manager->getElements();\r | |
254 | $this->info_content_sets = $this->manager->contentSets->lookup;\r | |
255 | }\r | |
256 | \r | |
257 | /**\r | |
258 | * Sets up stuff based on config. We need a better way of doing this.\r | |
259 | * @param HTMLPurifier_Config $config\r | |
260 | */\r | |
261 | protected function setupConfigStuff($config)\r | |
262 | {\r | |
263 | $block_wrapper = $config->get('HTML.BlockWrapper');\r | |
264 | if (isset($this->info_content_sets['Block'][$block_wrapper])) {\r | |
265 | $this->info_block_wrapper = $block_wrapper;\r | |
266 | } else {\r | |
267 | trigger_error(\r | |
268 | 'Cannot use non-block element as block wrapper',\r | |
269 | E_USER_ERROR\r | |
270 | );\r | |
271 | }\r | |
272 | \r | |
273 | $parent = $config->get('HTML.Parent');\r | |
274 | $def = $this->manager->getElement($parent, true);\r | |
275 | if ($def) {\r | |
276 | $this->info_parent = $parent;\r | |
277 | $this->info_parent_def = $def;\r | |
278 | } else {\r | |
279 | trigger_error(\r | |
280 | 'Cannot use unrecognized element as parent',\r | |
281 | E_USER_ERROR\r | |
282 | );\r | |
283 | $this->info_parent_def = $this->manager->getElement($this->info_parent, true);\r | |
284 | }\r | |
285 | \r | |
286 | // support template text\r | |
287 | $support = "(for information on implementing this, see the support forums) ";\r | |
288 | \r | |
289 | // setup allowed elements -----------------------------------------\r | |
290 | \r | |
291 | $allowed_elements = $config->get('HTML.AllowedElements');\r | |
292 | $allowed_attributes = $config->get('HTML.AllowedAttributes'); // retrieve early\r | |
293 | \r | |
294 | if (!is_array($allowed_elements) && !is_array($allowed_attributes)) {\r | |
295 | $allowed = $config->get('HTML.Allowed');\r | |
296 | if (is_string($allowed)) {\r | |
297 | list($allowed_elements, $allowed_attributes) = $this->parseTinyMCEAllowedList($allowed);\r | |
298 | }\r | |
299 | }\r | |
300 | \r | |
301 | if (is_array($allowed_elements)) {\r | |
302 | foreach ($this->info as $name => $d) {\r | |
303 | if (!isset($allowed_elements[$name])) {\r | |
304 | unset($this->info[$name]);\r | |
305 | }\r | |
306 | unset($allowed_elements[$name]);\r | |
307 | }\r | |
308 | // emit errors\r | |
309 | foreach ($allowed_elements as $element => $d) {\r | |
310 | $element = htmlspecialchars($element); // PHP doesn't escape errors, be careful!\r | |
311 | trigger_error("Element '$element' is not supported $support", E_USER_WARNING);\r | |
312 | }\r | |
313 | }\r | |
314 | \r | |
315 | // setup allowed attributes ---------------------------------------\r | |
316 | \r | |
317 | $allowed_attributes_mutable = $allowed_attributes; // by copy!\r | |
318 | if (is_array($allowed_attributes)) {\r | |
319 | // This actually doesn't do anything, since we went away from\r | |
320 | // global attributes. It's possible that userland code uses\r | |
321 | // it, but HTMLModuleManager doesn't!\r | |
322 | foreach ($this->info_global_attr as $attr => $x) {\r | |
323 | $keys = array($attr, "*@$attr", "*.$attr");\r | |
324 | $delete = true;\r | |
325 | foreach ($keys as $key) {\r | |
326 | if ($delete && isset($allowed_attributes[$key])) {\r | |
327 | $delete = false;\r | |
328 | }\r | |
329 | if (isset($allowed_attributes_mutable[$key])) {\r | |
330 | unset($allowed_attributes_mutable[$key]);\r | |
331 | }\r | |
332 | }\r | |
333 | if ($delete) {\r | |
334 | unset($this->info_global_attr[$attr]);\r | |
335 | }\r | |
336 | }\r | |
337 | \r | |
338 | foreach ($this->info as $tag => $info) {\r | |
339 | foreach ($info->attr as $attr => $x) {\r | |
340 | $keys = array("$tag@$attr", $attr, "*@$attr", "$tag.$attr", "*.$attr");\r | |
341 | $delete = true;\r | |
342 | foreach ($keys as $key) {\r | |
343 | if ($delete && isset($allowed_attributes[$key])) {\r | |
344 | $delete = false;\r | |
345 | }\r | |
346 | if (isset($allowed_attributes_mutable[$key])) {\r | |
347 | unset($allowed_attributes_mutable[$key]);\r | |
348 | }\r | |
349 | }\r | |
350 | if ($delete) {\r | |
351 | if ($this->info[$tag]->attr[$attr]->required) {\r | |
352 | trigger_error(\r | |
353 | "Required attribute '$attr' in element '$tag' " .\r | |
354 | "was not allowed, which means '$tag' will not be allowed either",\r | |
355 | E_USER_WARNING\r | |
356 | );\r | |
357 | }\r | |
358 | unset($this->info[$tag]->attr[$attr]);\r | |
359 | }\r | |
360 | }\r | |
361 | }\r | |
362 | // emit errors\r | |
363 | foreach ($allowed_attributes_mutable as $elattr => $d) {\r | |
364 | $bits = preg_split('/[.@]/', $elattr, 2);\r | |
365 | $c = count($bits);\r | |
366 | switch ($c) {\r | |
367 | case 2:\r | |
368 | if ($bits[0] !== '*') {\r | |
369 | $element = htmlspecialchars($bits[0]);\r | |
370 | $attribute = htmlspecialchars($bits[1]);\r | |
371 | if (!isset($this->info[$element])) {\r | |
372 | trigger_error(\r | |
373 | "Cannot allow attribute '$attribute' if element " .\r | |
374 | "'$element' is not allowed/supported $support"\r | |
375 | );\r | |
376 | } else {\r | |
377 | trigger_error(\r | |
378 | "Attribute '$attribute' in element '$element' not supported $support",\r | |
379 | E_USER_WARNING\r | |
380 | );\r | |
381 | }\r | |
382 | break;\r | |
383 | }\r | |
384 | // otherwise fall through\r | |
385 | case 1:\r | |
386 | $attribute = htmlspecialchars($bits[0]);\r | |
387 | trigger_error(\r | |
388 | "Global attribute '$attribute' is not ".\r | |
389 | "supported in any elements $support",\r | |
390 | E_USER_WARNING\r | |
391 | );\r | |
392 | break;\r | |
393 | }\r | |
394 | }\r | |
395 | }\r | |
396 | \r | |
397 | // setup forbidden elements ---------------------------------------\r | |
398 | \r | |
399 | $forbidden_elements = $config->get('HTML.ForbiddenElements');\r | |
400 | $forbidden_attributes = $config->get('HTML.ForbiddenAttributes');\r | |
401 | \r | |
402 | foreach ($this->info as $tag => $info) {\r | |
403 | if (isset($forbidden_elements[$tag])) {\r | |
404 | unset($this->info[$tag]);\r | |
405 | continue;\r | |
406 | }\r | |
407 | foreach ($info->attr as $attr => $x) {\r | |
408 | if (isset($forbidden_attributes["$tag@$attr"]) ||\r | |
409 | isset($forbidden_attributes["*@$attr"]) ||\r | |
410 | isset($forbidden_attributes[$attr])\r | |
411 | ) {\r | |
412 | unset($this->info[$tag]->attr[$attr]);\r | |
413 | continue;\r | |
414 | } elseif (isset($forbidden_attributes["$tag.$attr"])) { // this segment might get removed eventually\r | |
415 | // $tag.$attr are not user supplied, so no worries!\r | |
416 | trigger_error(\r | |
417 | "Error with $tag.$attr: tag.attr syntax not supported for " .\r | |
418 | "HTML.ForbiddenAttributes; use tag@attr instead",\r | |
419 | E_USER_WARNING\r | |
420 | );\r | |
421 | }\r | |
422 | }\r | |
423 | }\r | |
424 | foreach ($forbidden_attributes as $key => $v) {\r | |
425 | if (strlen($key) < 2) {\r | |
426 | continue;\r | |
427 | }\r | |
428 | if ($key[0] != '*') {\r | |
429 | continue;\r | |
430 | }\r | |
431 | if ($key[1] == '.') {\r | |
432 | trigger_error(\r | |
433 | "Error with $key: *.attr syntax not supported for HTML.ForbiddenAttributes; use attr instead",\r | |
434 | E_USER_WARNING\r | |
435 | );\r | |
436 | }\r | |
437 | }\r | |
438 | \r | |
439 | // setup injectors -----------------------------------------------------\r | |
440 | foreach ($this->info_injector as $i => $injector) {\r | |
441 | if ($injector->checkNeeded($config) !== false) {\r | |
442 | // remove injector that does not have it's required\r | |
443 | // elements/attributes present, and is thus not needed.\r | |
444 | unset($this->info_injector[$i]);\r | |
445 | }\r | |
446 | }\r | |
447 | }\r | |
448 | \r | |
449 | /**\r | |
450 | * Parses a TinyMCE-flavored Allowed Elements and Attributes list into\r | |
451 | * separate lists for processing. Format is element[attr1|attr2],element2...\r | |
452 | * @warning Although it's largely drawn from TinyMCE's implementation,\r | |
453 | * it is different, and you'll probably have to modify your lists\r | |
454 | * @param array $list String list to parse\r | |
455 | * @return array\r | |
456 | * @todo Give this its own class, probably static interface\r | |
457 | */\r | |
458 | public function parseTinyMCEAllowedList($list)\r | |
459 | {\r | |
460 | $list = str_replace(array(' ', "\t"), '', $list);\r | |
461 | \r | |
462 | $elements = array();\r | |
463 | $attributes = array();\r | |
464 | \r | |
465 | $chunks = preg_split('/(,|[\n\r]+)/', $list);\r | |
466 | foreach ($chunks as $chunk) {\r | |
467 | if (empty($chunk)) {\r | |
468 | continue;\r | |
469 | }\r | |
470 | // remove TinyMCE element control characters\r | |
471 | if (!strpos($chunk, '[')) {\r | |
472 | $element = $chunk;\r | |
473 | $attr = false;\r | |
474 | } else {\r | |
475 | list($element, $attr) = explode('[', $chunk);\r | |
476 | }\r | |
477 | if ($element !== '*') {\r | |
478 | $elements[$element] = true;\r | |
479 | }\r | |
480 | if (!$attr) {\r | |
481 | continue;\r | |
482 | }\r | |
483 | $attr = substr($attr, 0, strlen($attr) - 1); // remove trailing ]\r | |
484 | $attr = explode('|', $attr);\r | |
485 | foreach ($attr as $key) {\r | |
486 | $attributes["$element.$key"] = true;\r | |
487 | }\r | |
488 | }\r | |
489 | return array($elements, $attributes);\r | |
490 | }\r | |
491 | }\r | |
492 | \r | |
493 | // vim: et sw=4 sts=4\r |