| 1 | <?php |
|---|
| 2 | |
|---|
| 3 | /** |
|---|
| 4 | * Abstract class for a set of proprietary modules that clean up (tidy) |
|---|
| 5 | * poorly written HTML. |
|---|
| 6 | * @todo Figure out how to protect some of these methods/properties |
|---|
| 7 | */ |
|---|
| 8 | class HTMLPurifier_HTMLModule_Tidy extends HTMLPurifier_HTMLModule |
|---|
| 9 | { |
|---|
| 10 | |
|---|
| 11 | /** |
|---|
| 12 | * List of supported levels. Index zero is a special case "no fixes" |
|---|
| 13 | * level. |
|---|
| 14 | */ |
|---|
| 15 | public $levels = array(0 => 'none', 'light', 'medium', 'heavy'); |
|---|
| 16 | |
|---|
| 17 | /** |
|---|
| 18 | * Default level to place all fixes in. Disabled by default |
|---|
| 19 | */ |
|---|
| 20 | public $defaultLevel = null; |
|---|
| 21 | |
|---|
| 22 | /** |
|---|
| 23 | * Lists of fixes used by getFixesForLevel(). Format is: |
|---|
| 24 | * HTMLModule_Tidy->fixesForLevel[$level] = array('fix-1', 'fix-2'); |
|---|
| 25 | */ |
|---|
| 26 | public $fixesForLevel = array( |
|---|
| 27 | 'light' => array(), |
|---|
| 28 | 'medium' => array(), |
|---|
| 29 | 'heavy' => array() |
|---|
| 30 | ); |
|---|
| 31 | |
|---|
| 32 | /** |
|---|
| 33 | * Lazy load constructs the module by determining the necessary |
|---|
| 34 | * fixes to create and then delegating to the populate() function. |
|---|
| 35 | * @todo Wildcard matching and error reporting when an added or |
|---|
| 36 | * subtracted fix has no effect. |
|---|
| 37 | */ |
|---|
| 38 | public function construct($config) { |
|---|
| 39 | |
|---|
| 40 | // create fixes, initialize fixesForLevel |
|---|
| 41 | $fixes = $this->makeFixes(); |
|---|
| 42 | $this->makeFixesForLevel($fixes); |
|---|
| 43 | |
|---|
| 44 | // figure out which fixes to use |
|---|
| 45 | $level = $config->get('HTML', 'TidyLevel'); |
|---|
| 46 | $fixes_lookup = $this->getFixesForLevel($level); |
|---|
| 47 | |
|---|
| 48 | // get custom fix declarations: these need namespace processing |
|---|
| 49 | $add_fixes = $config->get('HTML', 'TidyAdd'); |
|---|
| 50 | $remove_fixes = $config->get('HTML', 'TidyRemove'); |
|---|
| 51 | |
|---|
| 52 | foreach ($fixes as $name => $fix) { |
|---|
| 53 | // needs to be refactored a little to implement globbing |
|---|
| 54 | if ( |
|---|
| 55 | isset($remove_fixes[$name]) || |
|---|
| 56 | (!isset($add_fixes[$name]) && !isset($fixes_lookup[$name])) |
|---|
| 57 | ) { |
|---|
| 58 | unset($fixes[$name]); |
|---|
| 59 | } |
|---|
| 60 | } |
|---|
| 61 | |
|---|
| 62 | // populate this module with necessary fixes |
|---|
| 63 | $this->populate($fixes); |
|---|
| 64 | |
|---|
| 65 | } |
|---|
| 66 | |
|---|
| 67 | /** |
|---|
| 68 | * Retrieves all fixes per a level, returning fixes for that specific |
|---|
| 69 | * level as well as all levels below it. |
|---|
| 70 | * @param $level String level identifier, see $levels for valid values |
|---|
| 71 | * @return Lookup up table of fixes |
|---|
| 72 | */ |
|---|
| 73 | public function getFixesForLevel($level) { |
|---|
| 74 | if ($level == $this->levels[0]) { |
|---|
| 75 | return array(); |
|---|
| 76 | } |
|---|
| 77 | $activated_levels = array(); |
|---|
| 78 | for ($i = 1, $c = count($this->levels); $i < $c; $i++) { |
|---|
| 79 | $activated_levels[] = $this->levels[$i]; |
|---|
| 80 | if ($this->levels[$i] == $level) break; |
|---|
| 81 | } |
|---|
| 82 | if ($i == $c) { |
|---|
| 83 | trigger_error( |
|---|
| 84 | 'Tidy level ' . htmlspecialchars($level) . ' not recognized', |
|---|
| 85 | E_USER_WARNING |
|---|
| 86 | ); |
|---|
| 87 | return array(); |
|---|
| 88 | } |
|---|
| 89 | $ret = array(); |
|---|
| 90 | foreach ($activated_levels as $level) { |
|---|
| 91 | foreach ($this->fixesForLevel[$level] as $fix) { |
|---|
| 92 | $ret[$fix] = true; |
|---|
| 93 | } |
|---|
| 94 | } |
|---|
| 95 | return $ret; |
|---|
| 96 | } |
|---|
| 97 | |
|---|
| 98 | /** |
|---|
| 99 | * Dynamically populates the $fixesForLevel member variable using |
|---|
| 100 | * the fixes array. It may be custom overloaded, used in conjunction |
|---|
| 101 | * with $defaultLevel, or not used at all. |
|---|
| 102 | */ |
|---|
| 103 | public function makeFixesForLevel($fixes) { |
|---|
| 104 | if (!isset($this->defaultLevel)) return; |
|---|
| 105 | if (!isset($this->fixesForLevel[$this->defaultLevel])) { |
|---|
| 106 | trigger_error( |
|---|
| 107 | 'Default level ' . $this->defaultLevel . ' does not exist', |
|---|
| 108 | E_USER_ERROR |
|---|
| 109 | ); |
|---|
| 110 | return; |
|---|
| 111 | } |
|---|
| 112 | $this->fixesForLevel[$this->defaultLevel] = array_keys($fixes); |
|---|
| 113 | } |
|---|
| 114 | |
|---|
| 115 | /** |
|---|
| 116 | * Populates the module with transforms and other special-case code |
|---|
| 117 | * based on a list of fixes passed to it |
|---|
| 118 | * @param $lookup Lookup table of fixes to activate |
|---|
| 119 | */ |
|---|
| 120 | public function populate($fixes) { |
|---|
| 121 | foreach ($fixes as $name => $fix) { |
|---|
| 122 | // determine what the fix is for |
|---|
| 123 | list($type, $params) = $this->getFixType($name); |
|---|
| 124 | switch ($type) { |
|---|
| 125 | case 'attr_transform_pre': |
|---|
| 126 | case 'attr_transform_post': |
|---|
| 127 | $attr = $params['attr']; |
|---|
| 128 | if (isset($params['element'])) { |
|---|
| 129 | $element = $params['element']; |
|---|
| 130 | if (empty($this->info[$element])) { |
|---|
| 131 | $e = $this->addBlankElement($element); |
|---|
| 132 | } else { |
|---|
| 133 | $e = $this->info[$element]; |
|---|
| 134 | } |
|---|
| 135 | } else { |
|---|
| 136 | $type = "info_$type"; |
|---|
| 137 | $e = $this; |
|---|
| 138 | } |
|---|
| 139 | // PHP does some weird parsing when I do |
|---|
| 140 | // $e->$type[$attr], so I have to assign a ref. |
|---|
| 141 | $f =& $e->$type; |
|---|
| 142 | $f[$attr] = $fix; |
|---|
| 143 | break; |
|---|
| 144 | case 'tag_transform': |
|---|
| 145 | $this->info_tag_transform[$params['element']] = $fix; |
|---|
| 146 | break; |
|---|
| 147 | case 'child': |
|---|
| 148 | case 'content_model_type': |
|---|
| 149 | $element = $params['element']; |
|---|
| 150 | if (empty($this->info[$element])) { |
|---|
| 151 | $e = $this->addBlankElement($element); |
|---|
| 152 | } else { |
|---|
| 153 | $e = $this->info[$element]; |
|---|
| 154 | } |
|---|
| 155 | $e->$type = $fix; |
|---|
| 156 | break; |
|---|
| 157 | default: |
|---|
| 158 | trigger_error("Fix type $type not supported", E_USER_ERROR); |
|---|
| 159 | break; |
|---|
| 160 | } |
|---|
| 161 | } |
|---|
| 162 | } |
|---|
| 163 | |
|---|
| 164 | /** |
|---|
| 165 | * Parses a fix name and determines what kind of fix it is, as well |
|---|
| 166 | * as other information defined by the fix |
|---|
| 167 | * @param $name String name of fix |
|---|
| 168 | * @return array(string $fix_type, array $fix_parameters) |
|---|
| 169 | * @note $fix_parameters is type dependant, see populate() for usage |
|---|
| 170 | * of these parameters |
|---|
| 171 | */ |
|---|
| 172 | public function getFixType($name) { |
|---|
| 173 | // parse it |
|---|
| 174 | $property = $attr = null; |
|---|
| 175 | if (strpos($name, '#') !== false) list($name, $property) = explode('#', $name); |
|---|
| 176 | if (strpos($name, '@') !== false) list($name, $attr) = explode('@', $name); |
|---|
| 177 | |
|---|
| 178 | // figure out the parameters |
|---|
| 179 | $params = array(); |
|---|
| 180 | if ($name !== '') $params['element'] = $name; |
|---|
| 181 | if (!is_null($attr)) $params['attr'] = $attr; |
|---|
| 182 | |
|---|
| 183 | // special case: attribute transform |
|---|
| 184 | if (!is_null($attr)) { |
|---|
| 185 | if (is_null($property)) $property = 'pre'; |
|---|
| 186 | $type = 'attr_transform_' . $property; |
|---|
| 187 | return array($type, $params); |
|---|
| 188 | } |
|---|
| 189 | |
|---|
| 190 | // special case: tag transform |
|---|
| 191 | if (is_null($property)) { |
|---|
| 192 | return array('tag_transform', $params); |
|---|
| 193 | } |
|---|
| 194 | |
|---|
| 195 | return array($property, $params); |
|---|
| 196 | |
|---|
| 197 | } |
|---|
| 198 | |
|---|
| 199 | /** |
|---|
| 200 | * Defines all fixes the module will perform in a compact |
|---|
| 201 | * associative array of fix name to fix implementation. |
|---|
| 202 | */ |
|---|
| 203 | public function makeFixes() {} |
|---|
| 204 | |
|---|
| 205 | } |
|---|
| 206 | |
|---|
| 207 | |
|---|