library/HTMLPurifier/Injector.php

   1 <?php
   2
   3 /**
   4  * Injects tokens into the document while parsing for well-formedness.
   5  * This enables "formatter-like" functionality such as auto-paragraphing,
   6  * smiley-ification and linkification to take place.
   7  *
   8  * A note on how handlers create changes; this is done by assigning a new
   9  * value to the $token reference. These values can take a variety of forms and
  10  * are best described HTMLPurifier_Strategy_MakeWellFormed->processToken()
  11  * documentation.
  12  *
  13  * @todo Allow injectors to request a re-run on their output. This
  14  *       would help if an operation is recursive.
  15  */
  16 abstract class HTMLPurifier_Injector
  17 {
  18
  19     /**
  20      * Advisory name of injector, this is for friendly error messages.
  21      * @type string
  22      */
  23     public $name;
  24
  25     /**
  26      * @type HTMLPurifier_HTMLDefinition
  27      */
  28     protected $htmlDefinition;
  29
  30     /**
  31      * Reference to CurrentNesting variable in Context. This is an array
  32      * list of tokens that we are currently "inside"
  33      * @type array
  34      */
  35     protected $currentNesting;
  36
  37     /**
  38      * Reference to current token.
  39      * @type HTMLPurifier_Token
  40      */
  41     protected $currentToken;
  42
  43     /**
  44      * Reference to InputZipper variable in Context.
  45      * @type HTMLPurifier_Zipper
  46      */
  47     protected $inputZipper;
  48
  49     /**
  50      * Array of elements and attributes this injector creates and therefore
  51      * need to be allowed by the definition. Takes form of
  52      * array('element' => array('attr', 'attr2'), 'element2')
  53      * @type array
  54      */
  55     public $needed = array();
  56
  57     /**
  58      * Number of elements to rewind backwards (relative).
  59      * @type bool|int
  60      */
  61     protected $rewindOffset = false;
  62
  63     /**
  64      * Rewind to a spot to re-perform processing. This is useful if you
  65      * deleted a node, and now need to see if this change affected any
  66      * earlier nodes. Rewinding does not affect other injectors, and can
  67      * result in infinite loops if not used carefully.
  68      * @param bool|int $offset
  69      * @warning HTML Purifier will prevent you from fast-forwarding with this
  70      *          function.
  71      */
  72     public function rewindOffset($offset)
  73     {
  74         $this->rewindOffset = $offset;
  75     }
  76
  77     /**
  78      * Retrieves rewind offset, and then unsets it.
  79      * @return bool|int
  80      */
  81     public function getRewindOffset()
  82     {
  83         $r = $this->rewindOffset;
  84         $this->rewindOffset = false;
  85         return $r;
  86     }
  87
  88     /**
  89      * Prepares the injector by giving it the config and context objects:
  90      * this allows references to important variables to be made within
  91      * the injector. This function also checks if the HTML environment
  92      * will work with the Injector (see checkNeeded()).
  93      * @param HTMLPurifier_Config $config
  94      * @param HTMLPurifier_Context $context
  95      * @return bool|string Boolean false if success, string of missing needed element/attribute if failure
  96      */
  97     public function prepare($config, $context)
  98     {
  99         $this->htmlDefinition = $config->getHTMLDefinition();
 100         // Even though this might fail, some unit tests ignore this and
 101         // still test checkNeeded, so be careful. Maybe get rid of that
 102         // dependency.
 103         $result = $this->checkNeeded($config);
 104         if ($result !== false) {
 105             return $result;
 106         }
 107         $this->currentNesting =& $context->get('CurrentNesting');
 108         $this->currentToken   =& $context->get('CurrentToken');
 109         $this->inputZipper    =& $context->get('InputZipper');
 110         return false;
 111     }
 112
 113     /**
 114      * This function checks if the HTML environment
 115      * will work with the Injector: if p tags are not allowed, the
 116      * Auto-Paragraphing injector should not be enabled.
 117      * @param HTMLPurifier_Config $config
 118      * @return bool|string Boolean false if success, string of missing needed element/attribute if failure
 119      */
 120     public function checkNeeded($config)
 121     {
 122         $def = $config->getHTMLDefinition();
 123         foreach ($this->needed as $element => $attributes) {
 124             if (is_int($element)) {
 125                 $element = $attributes;
 126             }
 127             if (!isset($def->info[$element])) {
 128                 return $element;
 129             }
 130             if (!is_array($attributes)) {
 131                 continue;
 132             }
 133             foreach ($attributes as $name) {
 134                 if (!isset($def->info[$element]->attr[$name])) {
 135                     return "$element.$name";
 136                 }
 137             }
 138         }
 139         return false;
 140     }
 141
 142     /**
 143      * Tests if the context node allows a certain element
 144      * @param string $name Name of element to test for
 145      * @return bool True if element is allowed, false if it is not
 146      */
 147     public function allowsElement($name)
 148     {
 149         if (!empty($this->currentNesting)) {
 150             $parent_token = array_pop($this->currentNesting);
 151             $this->currentNesting[] = $parent_token;
 152             $parent = $this->htmlDefinition->info[$parent_token->name];
 153         } else {
 154             $parent = $this->htmlDefinition->info_parent_def;
 155         }
 156         if (!isset($parent->child->elements[$name]) || isset($parent->excludes[$name])) {
 157             return false;
 158         }
 159         // check for exclusion
 160         if (!empty($this->currentNesting)) {
 161             for ($i = count($this->currentNesting) - 2; $i >= 0; $i--) {
 162                 $node = $this->currentNesting[$i];
 163                 $def  = $this->htmlDefinition->info[$node->name];
 164                 if (isset($def->excludes[$name])) {
 165                     return false;
 166                 }
 167             }
 168         }
 169         return true;
 170     }
 171
 172     /**
 173      * Iterator function, which starts with the next token and continues until
 174      * you reach the end of the input tokens.
 175      * @warning Please prevent previous references from interfering with this
 176      *          functions by setting $i = null beforehand!
 177      * @param int $i Current integer index variable for inputTokens
 178      * @param HTMLPurifier_Token $current Current token variable.
 179      *          Do NOT use $token, as that variable is also a reference
 180      * @return bool
 181      */
 182     protected function forward(&$i, &$current)
 183     {
 184         if ($i === null) {
 185             $i = count($this->inputZipper->back) - 1;
 186         } else {
 187             $i--;
 188         }
 189         if ($i < 0) {
 190             return false;
 191         }
 192         $current = $this->inputZipper->back[$i];
 193         return true;
 194     }
 195
 196     /**
 197      * Similar to _forward, but accepts a third parameter $nesting (which
 198      * should be initialized at 0) and stops when we hit the end tag
 199      * for the node $this->inputIndex starts in.
 200      * @param int $i Current integer index variable for inputTokens
 201      * @param HTMLPurifier_Token $current Current token variable.
 202      *          Do NOT use $token, as that variable is also a reference
 203      * @param int $nesting
 204      * @return bool
 205      */
 206     protected function forwardUntilEndToken(&$i, &$current, &$nesting)
 207     {
 208         $result = $this->forward($i, $current);
 209         if (!$result) {
 210             return false;
 211         }
 212         if ($nesting === null) {
 213             $nesting = 0;
 214         }
 215         if ($current instanceof HTMLPurifier_Token_Start) {
 216             $nesting++;
 217         } elseif ($current instanceof HTMLPurifier_Token_End) {
 218             if ($nesting <= 0) {
 219                 return false;
 220             }
 221             $nesting--;
 222         }
 223         return true;
 224     }
 225
 226     /**
 227      * Iterator function, starts with the previous token and continues until
 228      * you reach the beginning of input tokens.
 229      * @warning Please prevent previous references from interfering with this
 230      *          functions by setting $i = null beforehand!
 231      * @param int $i Current integer index variable for inputTokens
 232      * @param HTMLPurifier_Token $current Current token variable.
 233      *          Do NOT use $token, as that variable is also a reference
 234      * @return bool
 235      */
 236     protected function backward(&$i, &$current)
 237     {
 238         if ($i === null) {
 239             $i = count($this->inputZipper->front) - 1;
 240         } else {
 241             $i--;
 242         }
 243         if ($i < 0) {
 244             return false;
 245         }
 246         $current = $this->inputZipper->front[$i];
 247         return true;
 248     }
 249
 250     /**
 251      * Handler that is called when a text token is processed
 252      */
 253     public function handleText(&$token)
 254     {
 255     }
 256
 257     /**
 258      * Handler that is called when a start or empty token is processed
 259      */
 260     public function handleElement(&$token)
 261     {
 262     }
 263
 264     /**
 265      * Handler that is called when an end token is processed
 266      */
 267     public function handleEnd(&$token)
 268     {
 269         $this->notifyEnd($token);
 270     }
 271
 272     /**
 273      * Notifier that is called when an end token is processed
 274      * @param HTMLPurifier_Token $token Current token variable.
 275      * @note This differs from handlers in that the token is read-only
 276      * @deprecated
 277      */
 278     public function notifyEnd($token)
 279     {
 280     }
 281 }
 282
 283 // vim: et sw=4 sts=4