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.
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()
13
* @todo Allow injectors to request a re-run on their output. This
14
* would help if an operation is recursive.
16
abstract class HTMLPurifier_Injector
20
* Advisory name of injector, this is for friendly error messages
25
* Instance of HTMLPurifier_HTMLDefinition
27
protected $htmlDefinition;
30
* Reference to CurrentNesting variable in Context. This is an array
31
* list of tokens that we are currently "inside"
33
protected $currentNesting;
36
* Reference to InputTokens variable in Context. This is an array
37
* list of the input tokens that are being processed.
39
protected $inputTokens;
42
* Reference to InputIndex variable in Context. This is an integer
43
* array index for $this->inputTokens that indicates what token
44
* is currently being processed.
46
protected $inputIndex;
49
* Array of elements and attributes this injector creates and therefore
50
* need to be allowed by the definition. Takes form of
51
* array('element' => array('attr', 'attr2'), 'element2')
53
public $needed = array();
56
* Index of inputTokens to rewind to.
58
protected $rewind = false;
61
* Rewind to a spot to re-perform processing. This is useful if you
62
* deleted a node, and now need to see if this change affected any
63
* earlier nodes. Rewinding does not affect other injectors, and can
64
* result in infinite loops if not used carefully.
65
* @warning HTML Purifier will prevent you from fast-forwarding with this
68
public function rewind($index) {
69
$this->rewind = $index;
73
* Retrieves rewind, and then unsets it.
75
public function getRewind() {
77
$this->rewind = false;
82
* Prepares the injector by giving it the config and context objects:
83
* this allows references to important variables to be made within
84
* the injector. This function also checks if the HTML environment
85
* will work with the Injector (see checkNeeded()).
86
* @param $config Instance of HTMLPurifier_Config
87
* @param $context Instance of HTMLPurifier_Context
88
* @return Boolean false if success, string of missing needed element/attribute if failure
90
public function prepare($config, $context) {
91
$this->htmlDefinition = $config->getHTMLDefinition();
92
// Even though this might fail, some unit tests ignore this and
93
// still test checkNeeded, so be careful. Maybe get rid of that
95
$result = $this->checkNeeded($config);
96
if ($result !== false) return $result;
97
$this->currentNesting =& $context->get('CurrentNesting');
98
$this->inputTokens =& $context->get('InputTokens');
99
$this->inputIndex =& $context->get('InputIndex');
104
* This function checks if the HTML environment
105
* will work with the Injector: if p tags are not allowed, the
106
* Auto-Paragraphing injector should not be enabled.
107
* @param $config Instance of HTMLPurifier_Config
108
* @param $context Instance of HTMLPurifier_Context
109
* @return Boolean false if success, string of missing needed element/attribute if failure
111
public function checkNeeded($config) {
112
$def = $config->getHTMLDefinition();
113
foreach ($this->needed as $element => $attributes) {
114
if (is_int($element)) $element = $attributes;
115
if (!isset($def->info[$element])) return $element;
116
if (!is_array($attributes)) continue;
117
foreach ($attributes as $name) {
118
if (!isset($def->info[$element]->attr[$name])) return "$element.$name";
125
* Tests if the context node allows a certain element
126
* @param $name Name of element to test for
127
* @return True if element is allowed, false if it is not
129
public function allowsElement($name) {
130
if (!empty($this->currentNesting)) {
131
$parent_token = array_pop($this->currentNesting);
132
$this->currentNesting[] = $parent_token;
133
$parent = $this->htmlDefinition->info[$parent_token->name];
135
$parent = $this->htmlDefinition->info_parent_def;
137
if (!isset($parent->child->elements[$name]) || isset($parent->excludes[$name])) {
140
// check for exclusion
141
for ($i = count($this->currentNesting) - 2; $i >= 0; $i--) {
142
$node = $this->currentNesting[$i];
143
$def = $this->htmlDefinition->info[$node->name];
144
if (isset($def->excludes[$name])) return false;
150
* Iterator function, which starts with the next token and continues until
151
* you reach the end of the input tokens.
152
* @warning Please prevent previous references from interfering with this
153
* functions by setting $i = null beforehand!
154
* @param &$i Current integer index variable for inputTokens
155
* @param &$current Current token variable. Do NOT use $token, as that variable is also a reference
157
protected function forward(&$i, &$current) {
158
if ($i === null) $i = $this->inputIndex + 1;
160
if (!isset($this->inputTokens[$i])) return false;
161
$current = $this->inputTokens[$i];
166
* Similar to _forward, but accepts a third parameter $nesting (which
167
* should be initialized at 0) and stops when we hit the end tag
168
* for the node $this->inputIndex starts in.
170
protected function forwardUntilEndToken(&$i, &$current, &$nesting) {
171
$result = $this->forward($i, $current);
172
if (!$result) return false;
173
if ($nesting === null) $nesting = 0;
174
if ($current instanceof HTMLPurifier_Token_Start) $nesting++;
175
elseif ($current instanceof HTMLPurifier_Token_End) {
176
if ($nesting <= 0) return false;
183
* Iterator function, starts with the previous token and continues until
184
* you reach the beginning of input tokens.
185
* @warning Please prevent previous references from interfering with this
186
* functions by setting $i = null beforehand!
187
* @param &$i Current integer index variable for inputTokens
188
* @param &$current Current token variable. Do NOT use $token, as that variable is also a reference
190
protected function backward(&$i, &$current) {
191
if ($i === null) $i = $this->inputIndex - 1;
193
if ($i < 0) return false;
194
$current = $this->inputTokens[$i];
199
* Initializes the iterator at the current position. Use in a do {} while;
200
* loop to force the _forward and _backward functions to start at the
202
* @warning Please prevent previous references from interfering with this
203
* functions by setting $i = null beforehand!
204
* @param &$i Current integer index variable for inputTokens
205
* @param &$current Current token variable. Do NOT use $token, as that variable is also a reference
207
protected function current(&$i, &$current) {
208
if ($i === null) $i = $this->inputIndex;
209
$current = $this->inputTokens[$i];
213
* Handler that is called when a text token is processed
215
public function handleText(&$token) {}
218
* Handler that is called when a start or empty token is processed
220
public function handleElement(&$token) {}
223
* Handler that is called when an end token is processed
225
public function handleEnd(&$token) {
226
$this->notifyEnd($token);
230
* Notifier that is called when an end token is processed
231
* @note This differs from handlers in that the token is read-only
234
public function notifyEnd($token) {}
239
// vim: et sw=4 sts=4
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.
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()
13
* @todo Allow injectors to request a re-run on their output. This
14
* would help if an operation is recursive.
16
abstract class HTMLPurifier_Injector
20
* Advisory name of injector, this is for friendly error messages.
26
* @type HTMLPurifier_HTMLDefinition
28
protected $htmlDefinition;
31
* Reference to CurrentNesting variable in Context. This is an array
32
* list of tokens that we are currently "inside"
35
protected $currentNesting;
38
* Reference to current token.
39
* @type HTMLPurifier_Token
41
protected $currentToken;
44
* Reference to InputZipper variable in Context.
45
* @type HTMLPurifier_Zipper
47
protected $inputZipper;
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')
55
public $needed = array();
58
* Number of elements to rewind backwards (relative).
61
protected $rewindOffset = false;
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 $offset
69
* @warning HTML Purifier will prevent you from fast-forwarding with this
72
public function rewindOffset($offset)
74
$this->rewindOffset = $offset;
78
* Retrieves rewind offset, and then unsets it.
81
public function getRewindOffset()
83
$r = $this->rewindOffset;
84
$this->rewindOffset = false;
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 $context
95
* @return bool|string Boolean false if success, string of missing needed element/attribute if failure
97
public function prepare($config, $context)
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
103
$result = $this->checkNeeded($config);
104
if ($result !== false) {
107
$this->currentNesting =& $context->get('CurrentNesting');
108
$this->currentToken =& $context->get('CurrentToken');
109
$this->inputZipper =& $context->get('InputZipper');
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 $config
118
* @return bool|string Boolean false if success, string of missing needed element/attribute if failure
120
public function checkNeeded($config)
122
$def = $config->getHTMLDefinition();
123
foreach ($this->needed as $element => $attributes) {
124
if (is_int($element)) {
125
$element = $attributes;
127
if (!isset($def->info[$element])) {
130
if (!is_array($attributes)) {
133
foreach ($attributes as $name) {
134
if (!isset($def->info[$element]->attr[$name])) {
135
return "$element.$name";
143
* Tests if the context node allows a certain element
144
* @param string $name Name of element to test for
145
* @return bool True if element is allowed, false if it is not
147
public function allowsElement($name)
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];
154
$parent = $this->htmlDefinition->info_parent_def;
156
if (!isset($parent->child->elements[$name]) || isset($parent->excludes[$name])) {
159
// check for exclusion
160
for ($i = count($this->currentNesting) - 2; $i >= 0; $i--) {
161
$node = $this->currentNesting[$i];
162
$def = $this->htmlDefinition->info[$node->name];
163
if (isset($def->excludes[$name])) {
171
* Iterator function, which starts with the next token and continues until
172
* you reach the end of the input tokens.
173
* @warning Please prevent previous references from interfering with this
174
* functions by setting $i = null beforehand!
175
* @param int $i Current integer index variable for inputTokens
176
* @param HTMLPurifier_Token $current Current token variable.
177
* Do NOT use $token, as that variable is also a reference
180
protected function forward(&$i, &$current)
183
$i = count($this->inputZipper->back) - 1;
190
$current = $this->inputZipper->back[$i];
195
* Similar to _forward, but accepts a third parameter $nesting (which
196
* should be initialized at 0) and stops when we hit the end tag
197
* for the node $this->inputIndex starts in.
198
* @param int $i Current integer index variable for inputTokens
199
* @param HTMLPurifier_Token $current Current token variable.
200
* Do NOT use $token, as that variable is also a reference
201
* @param int $nesting
204
protected function forwardUntilEndToken(&$i, &$current, &$nesting)
206
$result = $this->forward($i, $current);
210
if ($nesting === null) {
213
if ($current instanceof HTMLPurifier_Token_Start) {
215
} elseif ($current instanceof HTMLPurifier_Token_End) {
225
* Iterator function, starts with the previous token and continues until
226
* you reach the beginning of input tokens.
227
* @warning Please prevent previous references from interfering with this
228
* functions by setting $i = null beforehand!
229
* @param int $i Current integer index variable for inputTokens
230
* @param HTMLPurifier_Token $current Current token variable.
231
* Do NOT use $token, as that variable is also a reference
234
protected function backward(&$i, &$current)
237
$i = count($this->inputZipper->front) - 1;
244
$current = $this->inputZipper->front[$i];
249
* Handler that is called when a text token is processed
251
public function handleText(&$token)
256
* Handler that is called when a start or empty token is processed
258
public function handleElement(&$token)
263
* Handler that is called when an end token is processed
265
public function handleEnd(&$token)
267
$this->notifyEnd($token);
271
* Notifier that is called when an end token is processed
272
* @param HTMLPurifier_Token $token Current token variable.
273
* @note This differs from handlers in that the token is read-only
276
public function notifyEnd($token)
281
// vim: et sw=4 sts=4