[Summary view] [Print] [Text view]

   1  <?php
   2  /**
   3   * HTML API: WP_HTML_Processor_State class
   4   *
   5   * @package WordPress
   6   * @subpackage HTML-API
   7   * @since 6.4.0
   8   */
   9  
  10  /**
  11   * Core class used by the HTML processor during HTML parsing
  12   * for managing the internal parsing state.
  13   *
  14   * This class is designed for internal use by the HTML processor.
  15   *
  16   * @since 6.4.0
  17   *
  18   * @access private
  19   * @ignore
  20   *
  21   * @see WP_HTML_Processor
  22   */
  23  class WP_HTML_Processor_State {
  24      /*
  25       * Insertion mode constants.
  26       *
  27       * These constants exist and are named to make it easier to
  28       * discover and recognize the supported insertion modes in
  29       * the parser.
  30       *
  31       * Out of all the possible insertion modes, only those
  32       * supported by the parser are listed here. As support
  33       * is added to the parser for more modes, add them here
  34       * following the same naming and value pattern.
  35       *
  36       * @see https://html.spec.whatwg.org/#the-insertion-mode
  37       */
  38  
  39      /**
  40       * Initial insertion mode for full HTML parser.
  41       *
  42       * @since 6.4.0
  43       *
  44       * @see https://html.spec.whatwg.org/#the-initial-insertion-mode
  45       * @see WP_HTML_Processor_State::$insertion_mode
  46       *
  47       * @var string
  48       */
  49      const INSERTION_MODE_INITIAL = 'insertion-mode-initial';
  50  
  51      /**
  52       * Before HTML insertion mode for full HTML parser.
  53       *
  54       * @since 6.7.0
  55       *
  56       * @see https://html.spec.whatwg.org/#the-before-html-insertion-mode
  57       * @see WP_HTML_Processor_State::$insertion_mode
  58       *
  59       * @var string
  60       */
  61      const INSERTION_MODE_BEFORE_HTML = 'insertion-mode-before-html';
  62  
  63      /**
  64       * Before head insertion mode for full HTML parser.
  65       *
  66       * @since 6.7.0
  67       *
  68       * @see https://html.spec.whatwg.org/#parsing-main-beforehead
  69       * @see WP_HTML_Processor_State::$insertion_mode
  70       *
  71       * @var string
  72       */
  73      const INSERTION_MODE_BEFORE_HEAD = 'insertion-mode-before-head';
  74  
  75      /**
  76       * In head insertion mode for full HTML parser.
  77       *
  78       * @since 6.7.0
  79       *
  80       * @see https://html.spec.whatwg.org/#parsing-main-inhead
  81       * @see WP_HTML_Processor_State::$insertion_mode
  82       *
  83       * @var string
  84       */
  85      const INSERTION_MODE_IN_HEAD = 'insertion-mode-in-head';
  86  
  87      /**
  88       * In head noscript insertion mode for full HTML parser.
  89       *
  90       * @since 6.7.0
  91       *
  92       * @see https://html.spec.whatwg.org/#parsing-main-inheadnoscript
  93       * @see WP_HTML_Processor_State::$insertion_mode
  94       *
  95       * @var string
  96       */
  97      const INSERTION_MODE_IN_HEAD_NOSCRIPT = 'insertion-mode-in-head-noscript';
  98  
  99      /**
 100       * After head insertion mode for full HTML parser.
 101       *
 102       * @since 6.7.0
 103       *
 104       * @see https://html.spec.whatwg.org/#parsing-main-afterhead
 105       * @see WP_HTML_Processor_State::$insertion_mode
 106       *
 107       * @var string
 108       */
 109      const INSERTION_MODE_AFTER_HEAD = 'insertion-mode-after-head';
 110  
 111      /**
 112       * In body insertion mode for full HTML parser.
 113       *
 114       * @since 6.4.0
 115       *
 116       * @see https://html.spec.whatwg.org/#parsing-main-inbody
 117       * @see WP_HTML_Processor_State::$insertion_mode
 118       *
 119       * @var string
 120       */
 121      const INSERTION_MODE_IN_BODY = 'insertion-mode-in-body';
 122  
 123      /**
 124       * In table insertion mode for full HTML parser.
 125       *
 126       * @since 6.7.0
 127       *
 128       * @see https://html.spec.whatwg.org/#parsing-main-intable
 129       * @see WP_HTML_Processor_State::$insertion_mode
 130       *
 131       * @var string
 132       */
 133      const INSERTION_MODE_IN_TABLE = 'insertion-mode-in-table';
 134  
 135      /**
 136       * In table text insertion mode for full HTML parser.
 137       *
 138       * @since 6.7.0
 139       *
 140       * @see https://html.spec.whatwg.org/#parsing-main-intabletext
 141       * @see WP_HTML_Processor_State::$insertion_mode
 142       *
 143       * @var string
 144       */
 145      const INSERTION_MODE_IN_TABLE_TEXT = 'insertion-mode-in-table-text';
 146  
 147      /**
 148       * In caption insertion mode for full HTML parser.
 149       *
 150       * @since 6.7.0
 151       *
 152       * @see https://html.spec.whatwg.org/#parsing-main-incaption
 153       * @see WP_HTML_Processor_State::$insertion_mode
 154       *
 155       * @var string
 156       */
 157      const INSERTION_MODE_IN_CAPTION = 'insertion-mode-in-caption';
 158  
 159      /**
 160       * In column group insertion mode for full HTML parser.
 161       *
 162       * @since 6.7.0
 163       *
 164       * @see https://html.spec.whatwg.org/#parsing-main-incolumngroup
 165       * @see WP_HTML_Processor_State::$insertion_mode
 166       *
 167       * @var string
 168       */
 169      const INSERTION_MODE_IN_COLUMN_GROUP = 'insertion-mode-in-column-group';
 170  
 171      /**
 172       * In table body insertion mode for full HTML parser.
 173       *
 174       * @since 6.7.0
 175       *
 176       * @see https://html.spec.whatwg.org/#parsing-main-intablebody
 177       * @see WP_HTML_Processor_State::$insertion_mode
 178       *
 179       * @var string
 180       */
 181      const INSERTION_MODE_IN_TABLE_BODY = 'insertion-mode-in-table-body';
 182  
 183      /**
 184       * In row insertion mode for full HTML parser.
 185       *
 186       * @since 6.7.0
 187       *
 188       * @see https://html.spec.whatwg.org/#parsing-main-inrow
 189       * @see WP_HTML_Processor_State::$insertion_mode
 190       *
 191       * @var string
 192       */
 193      const INSERTION_MODE_IN_ROW = 'insertion-mode-in-row';
 194  
 195      /**
 196       * In cell insertion mode for full HTML parser.
 197       *
 198       * @since 6.7.0
 199       *
 200       * @see https://html.spec.whatwg.org/#parsing-main-incell
 201       * @see WP_HTML_Processor_State::$insertion_mode
 202       *
 203       * @var string
 204       */
 205      const INSERTION_MODE_IN_CELL = 'insertion-mode-in-cell';
 206  
 207      /**
 208       * In select insertion mode for full HTML parser.
 209       *
 210       * @since 6.7.0
 211       *
 212       * @see https://html.spec.whatwg.org/#parsing-main-inselect
 213       * @see WP_HTML_Processor_State::$insertion_mode
 214       *
 215       * @var string
 216       */
 217      const INSERTION_MODE_IN_SELECT = 'insertion-mode-in-select';
 218  
 219      /**
 220       * In select in table insertion mode for full HTML parser.
 221       *
 222       * @since 6.7.0
 223       *
 224       * @see https://html.spec.whatwg.org/#parsing-main-inselectintable
 225       * @see WP_HTML_Processor_State::$insertion_mode
 226       *
 227       * @var string
 228       */
 229      const INSERTION_MODE_IN_SELECT_IN_TABLE = 'insertion-mode-in-select-in-table';
 230  
 231      /**
 232       * In template insertion mode for full HTML parser.
 233       *
 234       * @since 6.7.0
 235       *
 236       * @see https://html.spec.whatwg.org/#parsing-main-intemplate
 237       * @see WP_HTML_Processor_State::$insertion_mode
 238       *
 239       * @var string
 240       */
 241      const INSERTION_MODE_IN_TEMPLATE = 'insertion-mode-in-template';
 242  
 243      /**
 244       * After body insertion mode for full HTML parser.
 245       *
 246       * @since 6.7.0
 247       *
 248       * @see https://html.spec.whatwg.org/#parsing-main-afterbody
 249       * @see WP_HTML_Processor_State::$insertion_mode
 250       *
 251       * @var string
 252       */
 253      const INSERTION_MODE_AFTER_BODY = 'insertion-mode-after-body';
 254  
 255      /**
 256       * In frameset insertion mode for full HTML parser.
 257       *
 258       * @since 6.7.0
 259       *
 260       * @see https://html.spec.whatwg.org/#parsing-main-inframeset
 261       * @see WP_HTML_Processor_State::$insertion_mode
 262       *
 263       * @var string
 264       */
 265      const INSERTION_MODE_IN_FRAMESET = 'insertion-mode-in-frameset';
 266  
 267      /**
 268       * After frameset insertion mode for full HTML parser.
 269       *
 270       * @since 6.7.0
 271       *
 272       * @see https://html.spec.whatwg.org/#parsing-main-afterframeset
 273       * @see WP_HTML_Processor_State::$insertion_mode
 274       *
 275       * @var string
 276       */
 277      const INSERTION_MODE_AFTER_FRAMESET = 'insertion-mode-after-frameset';
 278  
 279      /**
 280       * After after body insertion mode for full HTML parser.
 281       *
 282       * @since 6.7.0
 283       *
 284       * @see https://html.spec.whatwg.org/#the-after-after-body-insertion-mode
 285       * @see WP_HTML_Processor_State::$insertion_mode
 286       *
 287       * @var string
 288       */
 289      const INSERTION_MODE_AFTER_AFTER_BODY = 'insertion-mode-after-after-body';
 290  
 291      /**
 292       * After after frameset insertion mode for full HTML parser.
 293       *
 294       * @since 6.7.0
 295       *
 296       * @see https://html.spec.whatwg.org/#the-after-after-frameset-insertion-mode
 297       * @see WP_HTML_Processor_State::$insertion_mode
 298       *
 299       * @var string
 300       */
 301      const INSERTION_MODE_AFTER_AFTER_FRAMESET = 'insertion-mode-after-after-frameset';
 302  
 303      /**
 304       * The stack of template insertion modes.
 305       *
 306       * @since 6.7.0
 307       *
 308       * @see https://html.spec.whatwg.org/#the-insertion-mode:stack-of-template-insertion-modes
 309       *
 310       * @var array<string>
 311       */
 312      public $stack_of_template_insertion_modes = array();
 313  
 314      /**
 315       * Tracks open elements while scanning HTML.
 316       *
 317       * This property is initialized in the constructor and never null.
 318       *
 319       * @since 6.4.0
 320       *
 321       * @see https://html.spec.whatwg.org/#stack-of-open-elements
 322       *
 323       * @var WP_HTML_Open_Elements
 324       */
 325      public $stack_of_open_elements;
 326  
 327      /**
 328       * Tracks open formatting elements, used to handle mis-nested formatting element tags.
 329       *
 330       * This property is initialized in the constructor and never null.
 331       *
 332       * @since 6.4.0
 333       *
 334       * @see https://html.spec.whatwg.org/#list-of-active-formatting-elements
 335       *
 336       * @var WP_HTML_Active_Formatting_Elements
 337       */
 338      public $active_formatting_elements;
 339  
 340      /**
 341       * Refers to the currently-matched tag, if any.
 342       *
 343       * @since 6.4.0
 344       *
 345       * @var WP_HTML_Token|null
 346       */
 347      public $current_token = null;
 348  
 349      /**
 350       * Tree construction insertion mode.
 351       *
 352       * @since 6.4.0
 353       *
 354       * @see https://html.spec.whatwg.org/#insertion-mode
 355       *
 356       * @var string
 357       */
 358      public $insertion_mode = self::INSERTION_MODE_INITIAL;
 359  
 360      /**
 361       * Context node initializing fragment parser, if created as a fragment parser.
 362       *
 363       * @since 6.4.0
 364       * @deprecated 6.8.0 WP_HTML_Processor tracks the context_node internally.
 365       *
 366       * @var null
 367       */
 368      public $context_node = null;
 369  
 370      /**
 371       * The recognized encoding of the input byte stream.
 372       *
 373       * > The stream of code points that comprises the input to the tokenization
 374       * > stage will be initially seen by the user agent as a stream of bytes
 375       * > (typically coming over the network or from the local file system).
 376       * > The bytes encode the actual characters according to a particular character
 377       * > encoding, which the user agent uses to decode the bytes into characters.
 378       *
 379       * @since 6.7.0
 380       *
 381       * @var string|null
 382       */
 383      public $encoding = null;
 384  
 385      /**
 386       * The parser's confidence in the input encoding.
 387       *
 388       * > When the HTML parser is decoding an input byte stream, it uses a character
 389       * > encoding and a confidence. The confidence is either tentative, certain, or
 390       * > irrelevant. The encoding used, and whether the confidence in that encoding
 391       * > is tentative or certain, is used during the parsing to determine whether to
 392       * > change the encoding. If no encoding is necessary, e.g. because the parser is
 393       * > operating on a Unicode stream and doesn't have to use a character encoding
 394       * > at all, then the confidence is irrelevant.
 395       *
 396       * @since 6.7.0
 397       *
 398       * @var string
 399       */
 400      public $encoding_confidence = 'tentative';
 401  
 402      /**
 403       * HEAD element pointer.
 404       *
 405       * @since 6.7.0
 406       *
 407       * @see https://html.spec.whatwg.org/multipage/parsing.html#head-element-pointer
 408       *
 409       * @var WP_HTML_Token|null
 410       */
 411      public $head_element = null;
 412  
 413      /**
 414       * FORM element pointer.
 415       *
 416       * > points to the last form element that was opened and whose end tag has
 417       * > not yet been seen. It is used to make form controls associate with
 418       * > forms in the face of dramatically bad markup, for historical reasons.
 419       * > It is ignored inside template elements.
 420       *
 421       * @todo This may be invalidated by a seek operation.
 422       *
 423       * @see https://html.spec.whatwg.org/#form-element-pointer
 424       *
 425       * @since 6.7.0
 426       *
 427       * @var WP_HTML_Token|null
 428       */
 429      public $form_element = null;
 430  
 431      /**
 432       * The frameset-ok flag indicates if a `FRAMESET` element is allowed in the current state.
 433       *
 434       * > The frameset-ok flag is set to "ok" when the parser is created. It is set to "not ok" after certain tokens are seen.
 435       *
 436       * @since 6.4.0
 437       *
 438       * @see https://html.spec.whatwg.org/#frameset-ok-flag
 439       *
 440       * @var bool
 441       */
 442      public $frameset_ok = true;
 443  
 444      /**
 445       * Constructor - creates a new and empty state value.
 446       *
 447       * @since 6.4.0
 448       *
 449       * @see WP_HTML_Processor
 450       */
 451  	public function __construct() {
 452          $this->stack_of_open_elements     = new WP_HTML_Open_Elements();
 453          $this->active_formatting_elements = new WP_HTML_Active_Formatting_Elements();
 454      }
 455  }