Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
70.31% |
161 / 229 |
|
64.71% |
11 / 17 |
CRAP | |
0.00% |
0 / 1 |
| Block_Scanner | |
70.31% |
161 / 229 |
|
64.71% |
11 / 17 |
285.18 | |
0.00% |
0 / 1 |
| create | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| next_delimiter | |
83.62% |
97 / 116 |
|
0.00% |
0 / 1 |
40.38 | |||
| __construct | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| find_html_comment_end | |
66.67% |
12 / 18 |
|
0.00% |
0 / 1 |
12.00 | |||
| get_last_error | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| get_last_json_error | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| get_delimiter_type | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| has_void_flag | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| is_block_type | |
88.89% |
16 / 18 |
|
0.00% |
0 / 1 |
10.14 | |||
| opens_block | |
100.00% |
8 / 8 |
|
100.00% |
1 / 1 |
5 | |||
| is_freeform | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| is_non_whitespace_freeform | |
66.67% |
2 / 3 |
|
0.00% |
0 / 1 |
2.15 | |||
| get_block_type | |
83.33% |
5 / 6 |
|
0.00% |
0 / 1 |
3.04 | |||
| get_attributes | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| allocate_and_return_parsed_attributes | |
100.00% |
12 / 12 |
|
100.00% |
1 / 1 |
5 | |||
| get_span | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| debug_print_structure | |
0.00% |
0 / 39 |
|
0.00% |
0 / 1 |
90 | |||
| 1 | <?php |
| 2 | /** |
| 3 | * Efficiently scan through block structure in document without parsing |
| 4 | * the entire block tree and all of its JSON attributes into memory. |
| 5 | * |
| 6 | * @package automattic/block-delimiter |
| 7 | */ |
| 8 | |
| 9 | declare( strict_types = 1 ); |
| 10 | |
| 11 | namespace Automattic; |
| 12 | |
| 13 | use Exception; |
| 14 | use WP_HTML_Span; |
| 15 | |
| 16 | /** |
| 17 | * Class for efficiently scanning through block structure in a document |
| 18 | * without parsing the entire block tree and JSON attributes into memory. |
| 19 | * |
| 20 | * This class follows design values of the HTML API: |
| 21 | * - minimize allocations and strive for zero memory overhead |
| 22 | * - make costs explicit; pay only for what you need |
| 23 | * - follow a streaming, re-entrant design for pausing and aborting |
| 24 | * |
| 25 | * For usage, jump straight to {@see self::next_delimiter}. |
| 26 | */ |
| 27 | class Block_Scanner { |
| 28 | /** |
| 29 | * Indicates if the last operation failed, otherwise |
| 30 | * will be `null` for success. |
| 31 | * |
| 32 | * @var string|null |
| 33 | */ |
| 34 | private $last_error = null; |
| 35 | |
| 36 | /** |
| 37 | * Indicates failures from decoding JSON attributes. |
| 38 | * |
| 39 | * @var int |
| 40 | */ |
| 41 | private $last_json_error = JSON_ERROR_NONE; |
| 42 | |
| 43 | /** |
| 44 | * Holds a reference to the original source text from which to |
| 45 | * extract the parsed spans of the delimiter. |
| 46 | * |
| 47 | * @var string |
| 48 | */ |
| 49 | private $source_text; |
| 50 | |
| 51 | /** |
| 52 | * Byte offset into source text where entire delimiter begins. |
| 53 | * |
| 54 | * @var int |
| 55 | */ |
| 56 | private $delimiter_at = 0; |
| 57 | |
| 58 | /** |
| 59 | * Byte length of full span of delimiter. |
| 60 | * |
| 61 | * @var int |
| 62 | */ |
| 63 | private $delimiter_length = 0; |
| 64 | |
| 65 | /** |
| 66 | * Byte offset where namespace span begins. |
| 67 | * |
| 68 | * @var int |
| 69 | */ |
| 70 | private $namespace_at = 0; |
| 71 | |
| 72 | /** |
| 73 | * Byte length of namespace span, or `0` if implicitly in the "core" namespace. |
| 74 | * |
| 75 | * @var int |
| 76 | */ |
| 77 | private $namespace_length = 0; |
| 78 | |
| 79 | /** |
| 80 | * Byte offset where block name span begins. |
| 81 | * |
| 82 | * @var int |
| 83 | */ |
| 84 | private $name_at = 0; |
| 85 | |
| 86 | /** |
| 87 | * Byte length of block name span. |
| 88 | * |
| 89 | * @var int |
| 90 | */ |
| 91 | private $name_length = 0; |
| 92 | |
| 93 | /** |
| 94 | * Whether the delimiter contains the block self-closing flag. |
| 95 | * |
| 96 | * This may be erroneous if present within a block closer, |
| 97 | * therefore the {@see self::has_void_flag} can be used by |
| 98 | * calling code to perform appropriate error-handling. |
| 99 | * |
| 100 | * @var bool |
| 101 | */ |
| 102 | private $has_void_flag = false; |
| 103 | |
| 104 | /** |
| 105 | * Byte offset where JSON attributes span begins. |
| 106 | * |
| 107 | * @var int |
| 108 | */ |
| 109 | private $json_at; |
| 110 | |
| 111 | /** |
| 112 | * Byte length of JSON attributes span, or `0` if none are present. |
| 113 | * |
| 114 | * @var int |
| 115 | */ |
| 116 | private $json_length; |
| 117 | |
| 118 | /** |
| 119 | * Indicates what kind of block comment delimiter this represents. |
| 120 | * |
| 121 | * One of: |
| 122 | * |
| 123 | * - `static::OPENER` If the delimiter is opening a block. |
| 124 | * - `static::CLOSER` If the delimiter is closing an open block. |
| 125 | * - `static::VOID` If the delimiter represents a void block with no inner content. |
| 126 | * |
| 127 | * If a parsed comment delimiter contains both the closing and the void |
| 128 | * flags then it will be interpreted as a void block to match the behavior |
| 129 | * of the official block parser, however, this is a mistake and probably |
| 130 | * the block ought to close an open block of the same name, if one is open. |
| 131 | * |
| 132 | * @var string |
| 133 | */ |
| 134 | private $type; |
| 135 | |
| 136 | /** |
| 137 | * Creates a new block scanner. |
| 138 | * |
| 139 | * Example: |
| 140 | * |
| 141 | * $scanner = Block_Scanner::create( $html ); |
| 142 | * while ( $scanner->next_delimiter() ) { |
| 143 | * if ( $scanner->opens_block( 'core/image' ) ) { |
| 144 | * echo "Found an image!\n"; |
| 145 | * } |
| 146 | * } |
| 147 | * |
| 148 | * This function is currently a stub so that future improvements can add configuration |
| 149 | * options and reject creation, which cannot occur directly inside class constructors. |
| 150 | * |
| 151 | * @see self::next_delimiter |
| 152 | * |
| 153 | * @param string $source_text Input document potentially containing block content. |
| 154 | * @return ?self Created block scanner, if successfully created. |
| 155 | */ |
| 156 | public static function create( string $source_text ): ?self { |
| 157 | return new self( $source_text ); |
| 158 | } |
| 159 | |
| 160 | /** |
| 161 | * Scan to the next block delimiter in a document, indicating if one was found. |
| 162 | * |
| 163 | * Block comment delimiters must be valid HTML comments and may contain JSON. |
| 164 | * This search does not determine, however, if the JSON is valid. |
| 165 | * |
| 166 | * Example delimiters: |
| 167 | * |
| 168 | * `<!-- wp:paragraph {"dropCap": true} -->` |
| 169 | * `<!-- wp:separator /-->` |
| 170 | * `<!-- /wp:paragraph -->` |
| 171 | * |
| 172 | * In the case that a block comment delimiter contains both the void indicator and |
| 173 | * also the closing indicator, it will be treated as a void block. |
| 174 | * |
| 175 | * Example: |
| 176 | * |
| 177 | * // Find all image block opening delimiters. |
| 178 | * $images = array(); |
| 179 | * $scanner = Block_Scanner::create( $html ); |
| 180 | * while ( $scanner->next_delimiter() ) { |
| 181 | * if ( $scanner->opens_block( 'core/image' ) ) { |
| 182 | * $images[] = $scanner->get_span(); |
| 183 | * } |
| 184 | * } |
| 185 | * |
| 186 | * Not all blocks have explicit delimiters. Non-block content at the top-level of |
| 187 | * a document (so-called “HTML soup”) forms implicit blocks containing neither a |
| 188 | * block name nor block attributes. Because this content often comprises only |
| 189 | * HTML whitespace and adds undo performance burden, it is skipped by default. |
| 190 | * To scan the implicit freeform blocks, pass the `$freeform_blocks` argument. |
| 191 | * |
| 192 | * Example: |
| 193 | * |
| 194 | * $html = '<!-- wp:void /-->\n<!-- wp:void /-->'; |
| 195 | * $blocks = [ |
| 196 | * [ 'blockName' => 'core/void' ], |
| 197 | * [ 'blockName' => null ], |
| 198 | * [ 'blockName' => 'core/void' ], |
| 199 | * ]; |
| 200 | * $scanner = Block_Scanner::create( $html ); |
| 201 | * while ( $scanner->next_delimiter( freeform_blocks: 'visit' ) { |
| 202 | * ... |
| 203 | * } |
| 204 | * |
| 205 | * In some cases it may be useful to conditionally visit the implicit freeform |
| 206 | * blocks, such as when determining if a post contains freeform content that |
| 207 | * isn’t purely whitespace. |
| 208 | * |
| 209 | * Example: |
| 210 | * |
| 211 | * $seen_block_types = []; |
| 212 | * $freeform_blocks = 'visit'; |
| 213 | * $scanner = Block_Scanner::create( $html ); |
| 214 | * while ( $scanner->next_delimiter( freeform_blocks: $freeform_blocks ) { |
| 215 | * if ( ! $scanner->opens_block() ) { |
| 216 | * continue; |
| 217 | * } |
| 218 | * |
| 219 | * // Stop wasting time visiting freeform blocks after one has been found. |
| 220 | * if ('visit' === $freeform_blocks ) { |
| 221 | * if ( $scanner->is_non_whitespace_freeform() ) { |
| 222 | * $freeform_blocks = 'skip'; |
| 223 | * $seen_block_types['core/freeform'] = true; |
| 224 | * } |
| 225 | * continue; |
| 226 | * } |
| 227 | * |
| 228 | * $seen_block_types[ $scanner->get_block_type() ] = true; |
| 229 | * } |
| 230 | * |
| 231 | * @param string $freeform_blocks Optional. Pass `visit` to match freeform HTML content |
| 232 | * not surrounded by block delimiters. Defaults to `skip`. |
| 233 | * @return bool Whether a block delimiter was matched. |
| 234 | */ |
| 235 | public function next_delimiter( string $freeform_blocks = 'skip' ): bool { // phpcs:ignore VariableAnalysis.CodeAnalysis.VariableAnalysis.UnusedVariable |
| 236 | if ( $this->last_error ) { |
| 237 | return false; |
| 238 | } |
| 239 | |
| 240 | $text = $this->source_text; |
| 241 | $end = strlen( $text ); |
| 242 | $at = $this->delimiter_at + $this->delimiter_length; |
| 243 | $found_one = false; |
| 244 | |
| 245 | while ( $at < $end ) { |
| 246 | /* |
| 247 | * Find the next possible opening. |
| 248 | * |
| 249 | * This follows the behavior in the official block parser, which treats a post |
| 250 | * as a list of blocks with nested HTML. If HTML comment syntax appears within |
| 251 | * an HTML attribute value, SCRIPT or STYLE element, or in other select places, |
| 252 | * which it can do inside of HTML, then the block parsing may break. |
| 253 | * |
| 254 | * For a more robust parse scan through the document with the HTML API. In |
| 255 | * practice, this has not been a problem in the entire history of blocks. |
| 256 | */ |
| 257 | $comment_opening_at = strpos( $text, '<!--', $at ); |
| 258 | if ( false === $comment_opening_at ) { |
| 259 | return false; |
| 260 | } |
| 261 | |
| 262 | $opening_whitespace_at = $comment_opening_at + 4; |
| 263 | $opening_whitespace_length = strspn( $text, " \t\f\r\n", $opening_whitespace_at ); |
| 264 | if ( 0 === $opening_whitespace_length ) { |
| 265 | $at = $this->find_html_comment_end( $comment_opening_at, $end ); |
| 266 | continue; |
| 267 | } |
| 268 | |
| 269 | $wp_prefix_at = $opening_whitespace_at + $opening_whitespace_length; |
| 270 | if ( $wp_prefix_at >= $end ) { |
| 271 | $this->last_error = self::INCOMPLETE_INPUT; |
| 272 | return false; |
| 273 | } |
| 274 | |
| 275 | $has_closer = false; |
| 276 | if ( '/' === $text[ $wp_prefix_at ] ) { |
| 277 | $has_closer = true; |
| 278 | ++$wp_prefix_at; |
| 279 | } |
| 280 | |
| 281 | if ( 0 !== substr_compare( $text, 'wp:', $wp_prefix_at, 3 ) ) { |
| 282 | $at = $this->find_html_comment_end( $comment_opening_at, $end ); |
| 283 | continue; |
| 284 | } |
| 285 | |
| 286 | $namespace_at = $wp_prefix_at + 3; |
| 287 | if ( $namespace_at >= $end ) { |
| 288 | $this->last_error = self::INCOMPLETE_INPUT; |
| 289 | return false; |
| 290 | } |
| 291 | |
| 292 | $start_of_namespace = $text[ $namespace_at ]; |
| 293 | |
| 294 | // The namespace must start with a-z. |
| 295 | if ( 'a' > $start_of_namespace || 'z' < $start_of_namespace ) { |
| 296 | $at = $this->find_html_comment_end( $comment_opening_at, $end ); |
| 297 | continue; |
| 298 | } |
| 299 | |
| 300 | $namespace_length = 1 + strspn( $text, 'abcdefghijklmnopqrstuvwxyz0123456789-_', $namespace_at + 1 ); |
| 301 | $separator_at = $namespace_at + $namespace_length; |
| 302 | if ( $separator_at >= $end ) { |
| 303 | $this->last_error = self::INCOMPLETE_INPUT; |
| 304 | return false; |
| 305 | } |
| 306 | |
| 307 | $has_separator = '/' === $text[ $separator_at ]; |
| 308 | if ( $has_separator ) { |
| 309 | $name_at = $separator_at + 1; |
| 310 | $start_of_name = $text[ $name_at ]; |
| 311 | if ( 'a' > $start_of_name || 'z' < $start_of_name ) { |
| 312 | $at = $this->find_html_comment_end( $comment_opening_at, $end ); |
| 313 | continue; |
| 314 | } |
| 315 | |
| 316 | $name_length = 1 + strspn( $text, 'abcdefghijklmnopqrstuvwxyz0123456789-_', $name_at + 1 ); |
| 317 | } else { |
| 318 | $name_at = $namespace_at; |
| 319 | $name_length = $namespace_length; |
| 320 | $namespace_length = 0; |
| 321 | } |
| 322 | |
| 323 | $after_name_whitespace_at = $name_at + $name_length; |
| 324 | $after_name_whitespace_length = strspn( $text, " \t\f\r\n", $after_name_whitespace_at ); |
| 325 | if ( 0 === $after_name_whitespace_length ) { |
| 326 | $at = $this->find_html_comment_end( $comment_opening_at, $end ); |
| 327 | continue; |
| 328 | } |
| 329 | |
| 330 | $json_at = $after_name_whitespace_at + $after_name_whitespace_length; |
| 331 | if ( $json_at >= $end ) { |
| 332 | $this->last_error = self::INCOMPLETE_INPUT; |
| 333 | return false; |
| 334 | } |
| 335 | $has_json = '{' === $text[ $json_at ]; |
| 336 | $json_length = 0; |
| 337 | |
| 338 | /* |
| 339 | * For the final span of the delimiter it's most efficient to find the end |
| 340 | * of the HTML comment and work backwards. This prevents complicated parsing |
| 341 | * inside the JSON span, which cannot contain the HTML comment terminator. |
| 342 | * |
| 343 | * This also matches the behavior in the official block parser, though it |
| 344 | * allows for matching invalid JSON content. |
| 345 | */ |
| 346 | $comment_closing_at = strpos( $text, '-->', $json_at ); |
| 347 | if ( false === $comment_closing_at ) { |
| 348 | $this->last_error = self::INCOMPLETE_INPUT; |
| 349 | return false; |
| 350 | } |
| 351 | |
| 352 | /* |
| 353 | * It looks like this logic leaves an error in here, when the position |
| 354 | * overlaps the JSON or block name. However, for neither of those is it |
| 355 | * possible to parse a valid block if that last overlapping character |
| 356 | * is the void flag. This, therefore, will be valid regardless of how |
| 357 | * the rest of the comment delimiter is written. |
| 358 | */ |
| 359 | if ( '/' === $text[ $comment_closing_at - 1 ] ) { |
| 360 | $has_void_flag = true; |
| 361 | $void_flag_length = 1; |
| 362 | } else { |
| 363 | $has_void_flag = false; |
| 364 | $void_flag_length = 0; |
| 365 | } |
| 366 | |
| 367 | /* |
| 368 | * If there's no JSON, then the span of text after the name |
| 369 | * until the comment closing must be completely whitespace. |
| 370 | */ |
| 371 | if ( ! $has_json ) { |
| 372 | $max_whitespace_length = $comment_closing_at - $json_at - $void_flag_length; |
| 373 | |
| 374 | // This shouldn't be possible, but it can't be allowed regardless. |
| 375 | if ( $max_whitespace_length < 0 ) { |
| 376 | $at = $this->find_html_comment_end( $comment_opening_at, $end ); |
| 377 | continue; |
| 378 | } |
| 379 | |
| 380 | $closing_whitespace_length = strspn( $text, " \t\f\r\n", $json_at, $comment_closing_at - $json_at - $void_flag_length ); |
| 381 | if ( 0 === $after_name_whitespace_length + $closing_whitespace_length ) { |
| 382 | $at = $this->find_html_comment_end( $comment_opening_at, $end ); |
| 383 | continue; |
| 384 | } |
| 385 | |
| 386 | // This must be a block delimiter! |
| 387 | $found_one = true; |
| 388 | break; |
| 389 | } |
| 390 | |
| 391 | // There's no JSON, so attempt to find its boundary. |
| 392 | $after_json_whitespace_length = 0; |
| 393 | for ( $char_at = $comment_closing_at - $void_flag_length - 1; $char_at > $json_at; $char_at-- ) { |
| 394 | $char = $text[ $char_at ]; |
| 395 | |
| 396 | switch ( $char ) { |
| 397 | case ' ': |
| 398 | case "\t": |
| 399 | case "\f": |
| 400 | case "\r": |
| 401 | case "\n": |
| 402 | ++$after_json_whitespace_length; |
| 403 | continue 2; |
| 404 | |
| 405 | case '}': |
| 406 | $json_length = $char_at - $json_at + 1; |
| 407 | break 2; |
| 408 | |
| 409 | default: |
| 410 | ++$at; |
| 411 | continue 3; |
| 412 | } |
| 413 | } |
| 414 | |
| 415 | if ( 0 === $json_length || 0 === $after_json_whitespace_length ) { |
| 416 | $at = $this->find_html_comment_end( $comment_opening_at, $end ); |
| 417 | continue; |
| 418 | } |
| 419 | |
| 420 | // This must be a block delimiter! |
| 421 | $found_one = true; |
| 422 | break; |
| 423 | } |
| 424 | |
| 425 | if ( ! $found_one ) { |
| 426 | return false; |
| 427 | } |
| 428 | |
| 429 | $this->delimiter_at = $comment_opening_at; |
| 430 | $this->delimiter_length = $comment_closing_at + 3 - $comment_opening_at; |
| 431 | |
| 432 | $this->namespace_at = $namespace_at; |
| 433 | $this->namespace_length = $namespace_length; |
| 434 | |
| 435 | $this->name_at = $name_at; |
| 436 | $this->name_length = $name_length; |
| 437 | |
| 438 | $this->json_at = $json_at; |
| 439 | $this->json_length = $json_length; |
| 440 | |
| 441 | $this->type = $has_closer |
| 442 | ? static::CLOSER |
| 443 | : ( $has_void_flag ? static::VOID : static::OPENER ); |
| 444 | |
| 445 | $this->has_void_flag = $has_void_flag; |
| 446 | |
| 447 | return true; |
| 448 | } |
| 449 | |
| 450 | /** |
| 451 | * Constructor function. |
| 452 | * |
| 453 | * @param string $source_text Input document potentially containing block content. |
| 454 | */ |
| 455 | private function __construct( string $source_text ) { |
| 456 | $this->source_text = $source_text; |
| 457 | } |
| 458 | |
| 459 | /** |
| 460 | * Returns the byte-offset after the ending character of an HTML comment, |
| 461 | * assuming the proper starting byte offset. |
| 462 | * |
| 463 | * @param int $comment_starting_at Where the HTML comment started, the leading `<`. |
| 464 | * @param int $search_end Last offset in which to search, for limiting search span. |
| 465 | * @return int Offset after the current HTML comment ends, or `$end` if no end was found. |
| 466 | */ |
| 467 | private function find_html_comment_end( int $comment_starting_at, int $search_end ): int { |
| 468 | $text = $this->source_text; |
| 469 | |
| 470 | // Find span-of-dashes comments which look like `<!----->`. |
| 471 | $span_of_dashes = strspn( $text, '-', $comment_starting_at + 2 ); |
| 472 | if ( |
| 473 | $comment_starting_at + 2 + $span_of_dashes < $search_end && |
| 474 | '>' === $text[ $comment_starting_at + 2 + $span_of_dashes ] |
| 475 | ) { |
| 476 | return $comment_starting_at + $span_of_dashes + 1; |
| 477 | } |
| 478 | |
| 479 | // Otherwise, there are other characters inside the comment, find the first `-->` or `--!>`. |
| 480 | $now_at = $comment_starting_at + 4; |
| 481 | while ( $now_at < $search_end ) { |
| 482 | $dashes_at = strpos( $text, '--', $now_at ); |
| 483 | if ( false === $dashes_at ) { |
| 484 | $this->last_error = self::INCOMPLETE_INPUT; |
| 485 | return $search_end; |
| 486 | } |
| 487 | |
| 488 | $closer_must_be_at = $dashes_at + 2 + strspn( $text, '-', $dashes_at + 2 ); |
| 489 | if ( $closer_must_be_at < $search_end && '!' === $text[ $closer_must_be_at ] ) { |
| 490 | $closer_must_be_at++; |
| 491 | } |
| 492 | |
| 493 | if ( $closer_must_be_at < $search_end && '>' === $text[ $closer_must_be_at ] ) { |
| 494 | return $closer_must_be_at + 1; |
| 495 | } |
| 496 | |
| 497 | $now_at++; |
| 498 | } |
| 499 | |
| 500 | return $search_end; |
| 501 | } |
| 502 | |
| 503 | /** |
| 504 | * Indicates if the last attempt to parse a block comment delimiter |
| 505 | * failed, if set, otherwise `null` if the last attempt succeeded. |
| 506 | * |
| 507 | * @return string|null |
| 508 | */ |
| 509 | public function get_last_error() { |
| 510 | return $this->last_error; |
| 511 | } |
| 512 | |
| 513 | /** |
| 514 | * Indicates if the last attempt to parse a block’s JSON attributes failed. |
| 515 | * |
| 516 | * @see JSON_ERROR_NONE, JSON_ERROR_DEPTH, etc… |
| 517 | * |
| 518 | * @return int JSON_ERROR_ code from last attempt to parse block JSON attributes. |
| 519 | */ |
| 520 | public function get_last_json_error(): int { |
| 521 | return $this->last_json_error; |
| 522 | } |
| 523 | |
| 524 | /** |
| 525 | * Returns the type of the block comment delimiter. |
| 526 | * |
| 527 | * One of: |
| 528 | * |
| 529 | * - `static::OPENER` |
| 530 | * - `static::CLOSER` |
| 531 | * - `static::VOID` |
| 532 | * |
| 533 | * @return string type of the block comment delimiter. |
| 534 | */ |
| 535 | public function get_delimiter_type(): string { |
| 536 | return $this->type; |
| 537 | } |
| 538 | |
| 539 | /** |
| 540 | * Returns whether the delimiter contains the void flag. |
| 541 | * |
| 542 | * This should be avoided except in cases of handling errors with |
| 543 | * block closers containing the void flag. For normative use, |
| 544 | * {@see self::get_delimiter_type}. |
| 545 | * |
| 546 | * @return bool |
| 547 | */ |
| 548 | public function has_void_flag(): bool { |
| 549 | return $this->has_void_flag; |
| 550 | } |
| 551 | |
| 552 | /** |
| 553 | * Indicates if the block delimiter represents a block of the given type. |
| 554 | * |
| 555 | * Since the "core" namespace may be implicit, it's allowable to pass |
| 556 | * either the fully-qualified block type with namespace and block name |
| 557 | * as well as the shorthand version only containing the block name, if |
| 558 | * the desired block is in the "core" namespace. |
| 559 | * |
| 560 | * Example: |
| 561 | * |
| 562 | * $is_core_paragraph = $scanner->is_block_type( 'paragraph' ); |
| 563 | * $is_core_paragraph = $scanner->is_block_type( 'core/paragraph' ); |
| 564 | * $is_formula = $scanner->is_block_type( 'math-block/formula' ); |
| 565 | * |
| 566 | * @param string $block_type Block type name for the desired block. |
| 567 | * E.g. "paragraph", "core/paragraph", "math-blocks/formula". |
| 568 | * @return bool Whether this delimiter represents a block of the given type. |
| 569 | */ |
| 570 | public function is_block_type( string $block_type ): bool { |
| 571 | // This is a core/freeform text block, it’s special. |
| 572 | if ( 0 === $this->name_length ) { |
| 573 | return 'core/freeform' === $block_type || 'freeform' === $block_type; |
| 574 | } |
| 575 | |
| 576 | $slash_at = strpos( $block_type, '/' ); |
| 577 | if ( false === $slash_at ) { |
| 578 | $namespace = 'core'; |
| 579 | $block_name = $block_type; |
| 580 | } else { |
| 581 | // @todo Get lengths but avoid the allocation, use substr_compare below. |
| 582 | $namespace = substr( $block_type, 0, $slash_at ); |
| 583 | $block_name = substr( $block_type, $slash_at + 1 ); |
| 584 | } |
| 585 | |
| 586 | // Only the 'core' namespace is allowed to be omitted. |
| 587 | if ( 0 === $this->namespace_length && 'core' !== $namespace ) { |
| 588 | return false; |
| 589 | } |
| 590 | |
| 591 | // If given an explicit namespace, they must match. |
| 592 | if ( |
| 593 | 0 !== $this->namespace_length && ( |
| 594 | strlen( $namespace ) !== $this->namespace_length || |
| 595 | 0 !== substr_compare( $this->source_text, $namespace, $this->namespace_at, $this->namespace_length ) |
| 596 | ) |
| 597 | ) { |
| 598 | return false; |
| 599 | } |
| 600 | |
| 601 | // The block name must match. |
| 602 | return ( |
| 603 | strlen( $block_name ) === $this->name_length && |
| 604 | 0 === substr_compare( $this->source_text, $block_name, $this->name_at, $this->name_length ) |
| 605 | ); |
| 606 | } |
| 607 | |
| 608 | /** |
| 609 | * Indicates if the matched delimiter is an opening or void delimiter |
| 610 | * (i.e. it opens the block) of the given type, if a type is provided. |
| 611 | * |
| 612 | * This is a helper method to ease handling of code inspecting where |
| 613 | * blocks start, and of checking if the blocks are of a given type. |
| 614 | * The function is variadic to allow for checking if the delimiter |
| 615 | * opens one of many possible block types. |
| 616 | * |
| 617 | * Example: |
| 618 | * |
| 619 | * $scanner = Block_Scanner::create( $html ); |
| 620 | * while ( $scanner->next_delimiter() ) { |
| 621 | * if ( $scanner->opens_block( 'core/code', 'syntaxhighlighter/code' ) ) { |
| 622 | * echo "Found code!"; |
| 623 | * continue; |
| 624 | * } |
| 625 | * |
| 626 | * if ( $scanner->opens_block( 'core/image' ) ) { |
| 627 | * echo "Found an image!"; |
| 628 | * continue; |
| 629 | * } |
| 630 | * |
| 631 | * if ( $scanner->opens_block() ) { |
| 632 | * echo "Found a new block!"; |
| 633 | * } |
| 634 | * } |
| 635 | * |
| 636 | * @see self::is_block_type |
| 637 | * |
| 638 | * @param string|null ...$block_type Optional. Is the matched block type one of these? |
| 639 | * If none are provided, will not test block type. |
| 640 | * @return bool Whether the matched block delimiter opens a block, and whether it |
| 641 | * opens a block of one of the given block types, if provided. |
| 642 | */ |
| 643 | public function opens_block( ...$block_type ): bool { |
| 644 | if ( static::CLOSER === $this->type ) { |
| 645 | return false; |
| 646 | } |
| 647 | |
| 648 | if ( count( $block_type ) === 0 ) { |
| 649 | return true; |
| 650 | } |
| 651 | |
| 652 | foreach ( $block_type as $block ) { |
| 653 | if ( $this->is_block_type( $block ) ) { |
| 654 | return true; |
| 655 | } |
| 656 | } |
| 657 | |
| 658 | return false; |
| 659 | } |
| 660 | |
| 661 | /** |
| 662 | * Indicates if the matched delimiter is implied due to top-level |
| 663 | * non-block content in the post. |
| 664 | * |
| 665 | * @see self::is_non_whitespace_freeform |
| 666 | * |
| 667 | * @return bool Whether or not the matched delimiter is implied as `core/freeform`. |
| 668 | */ |
| 669 | public function is_freeform(): bool { |
| 670 | return 0 === $this->name_length; |
| 671 | } |
| 672 | |
| 673 | /** |
| 674 | * Indicates if the matched delimiter is implicit and surrounding |
| 675 | * top-level non-block content that contains non-whitespace text. |
| 676 | * |
| 677 | * Many block serializers introduce newlines between block delimiters, |
| 678 | * so the presence of top-level non-block content does not imply that |
| 679 | * there are “real” freeform HTML blocks. Checking if there is content |
| 680 | * beyond whitespace is a more certain check, such as for determining |
| 681 | * whether to load CSS for the freeform or fallback block type. |
| 682 | * |
| 683 | * @see self::is_freeform |
| 684 | * |
| 685 | * @return bool |
| 686 | */ |
| 687 | public function is_non_whitespace_freeform(): bool { |
| 688 | if ( 0 !== $this->name_length ) { |
| 689 | return false; |
| 690 | } |
| 691 | |
| 692 | // For now, return false as this method is not yet fully implemented. |
| 693 | // @todo Implement logic to check if freeform content contains non-whitespace text. |
| 694 | return false; |
| 695 | } |
| 696 | |
| 697 | /** |
| 698 | * Allocates a substring for the block type and returns the |
| 699 | * fully-qualified name, including the namespace. |
| 700 | * |
| 701 | * This function allocates a substring for the given block type. This |
| 702 | * allocation will be small and likely fine in most cases, but it's |
| 703 | * preferable to call {@link self::is_block_type} if only needing |
| 704 | * to know whether the delimiter is for a given block type, as that |
| 705 | * function is more efficient for this purpose and avoids the allocation. |
| 706 | * |
| 707 | * Example: |
| 708 | * |
| 709 | * // Avoid. |
| 710 | * 'core/paragraph' = $scanner->get_block_type(); |
| 711 | * |
| 712 | * // Prefer. |
| 713 | * $scanner->is_block_type( 'core/paragraph' ); |
| 714 | * $scanner->is_block_type( 'paragraph' ); |
| 715 | * |
| 716 | * @return string Fully-qualified block namespace and type, e.g. "core/paragraph". |
| 717 | */ |
| 718 | public function get_block_type(): string { |
| 719 | // This is a core/freeform text block, it’s special. |
| 720 | if ( 0 === $this->name_length ) { |
| 721 | return 'core/freeform'; |
| 722 | } |
| 723 | |
| 724 | // This is implicitly in the "core" namespace. |
| 725 | if ( 0 === $this->namespace_length ) { |
| 726 | $block_name = substr( $this->source_text, $this->name_at, $this->name_length ); |
| 727 | return "core/{$block_name}"; |
| 728 | } |
| 729 | |
| 730 | return substr( $this->source_text, $this->namespace_at, $this->namespace_length + $this->name_length + 1 ); |
| 731 | } |
| 732 | |
| 733 | /** |
| 734 | * Returns a lazy wrapper around the block attributes, which can be used |
| 735 | * for efficiently interacting with the JSON attributes. |
| 736 | * |
| 737 | * @throws Exception This function is not yet implemented. |
| 738 | * |
| 739 | * @todo Create a lazy JSON wrapper so specific attributes can be |
| 740 | * efficiently queried without parsing everything and loading |
| 741 | * the entire object into memory. |
| 742 | * @todo After realistic benchmarking, see if JsonStreamingParser\Parser |
| 743 | * could be used — it would need to be fast enough for the reduction |
| 744 | * in memory use to be worth it, compared to {@see \json_decode}. |
| 745 | * |
| 746 | * @see \JsonStreamingParser\Parser |
| 747 | * |
| 748 | * @return never |
| 749 | */ |
| 750 | public function get_attributes(): void { |
| 751 | throw new Exception( 'Lazy attribute parsing not yet supported' ); |
| 752 | } |
| 753 | |
| 754 | /** |
| 755 | * Attempts to parse and return the entire JSON attributes from the delimiter, |
| 756 | * allocating memory and processing the JSON span in the process. |
| 757 | * |
| 758 | * This does not return any parsed attributes for a closing block delimiter |
| 759 | * even if there is a span of JSON content; this JSON is a parsing error. |
| 760 | * |
| 761 | * Consider calling {@link self::get_attributes} instead if it's not |
| 762 | * necessary to read all the attributes at the same time, as that provides |
| 763 | * a more efficient mechanism for typical use cases. |
| 764 | * |
| 765 | * Since the JSON span inside the comment delimiter may not be valid JSON, |
| 766 | * this function will return `null` if it cannot parse the span and set the |
| 767 | * {@see self::get_last_json_error} to the appropriate JSON_ERROR_ constant. |
| 768 | * |
| 769 | * If the delimiter contains no JSON span, it will also return `null`, |
| 770 | * but the last error will be set to {@see JSON_ERROR_NONE}. |
| 771 | * |
| 772 | * Example: |
| 773 | * |
| 774 | * $scanner = Block_Scanner::create( '<!-- wp:image {"url": "https://wordpress.org/favicon.ico"} -->' ); |
| 775 | * $scanner->next_delimiter(); |
| 776 | * $memory_hungry_and_slow_attributes = $scanner->allocate_and_return_parsed_attributes(); |
| 777 | * $memory_hungry_and_slow_attributes === array( 'url' => 'https://wordpress.org/favicon.ico' ); |
| 778 | * |
| 779 | * $scanner = Block_Scanner::create( '<!-- /wp:image {"url": "https://wordpress.org/favicon.ico"} -->' ); |
| 780 | * $scanner->next_delimiter(); |
| 781 | * null = $scanner->allocate_and_return_parsed_attributes(); |
| 782 | * JSON_ERROR_NONE = $scanner->get_last_json_error(); |
| 783 | * |
| 784 | * $scanner = Block_Scanner::create( '<!-- wp:separator {} /-->' ); |
| 785 | * $scanner->next_delimiter(); |
| 786 | * array() === $scanner->allocate_and_return_parsed_attributes(); |
| 787 | * |
| 788 | * $scanner = Block_Scanner::create( '<!-- wp:separator /-->' ); |
| 789 | * $scanner->next_delimiter(); |
| 790 | * null = $scanner->allocate_and_return_parsed_attributes(); |
| 791 | * |
| 792 | * $scanner = Block_Scanner::create( '<!-- wp:image {"url} -->' ); |
| 793 | * $scanner->next_delimiter(); |
| 794 | * null = $scanner->allocate_and_return_parsed_attributes(); |
| 795 | * JSON_ERROR_CTRL_CHAR = $scanner->get_last_json_error(); |
| 796 | * |
| 797 | * @return array|null Parsed JSON attributes, if present and valid, otherwise `null`. |
| 798 | */ |
| 799 | public function allocate_and_return_parsed_attributes(): ?array { |
| 800 | $this->last_json_error = JSON_ERROR_NONE; |
| 801 | |
| 802 | if ( static::CLOSER === $this->type ) { |
| 803 | return null; |
| 804 | } |
| 805 | |
| 806 | if ( 0 === $this->json_length ) { |
| 807 | return null; |
| 808 | } |
| 809 | |
| 810 | $json_span = substr( $this->source_text, $this->json_at, $this->json_length ); |
| 811 | $parsed = json_decode( $json_span, null, 512, JSON_OBJECT_AS_ARRAY | JSON_INVALID_UTF8_SUBSTITUTE ); |
| 812 | |
| 813 | $last_error = json_last_error(); |
| 814 | $this->last_json_error = $last_error; |
| 815 | |
| 816 | return ( JSON_ERROR_NONE === $last_error && is_array( $parsed ) ) |
| 817 | ? $parsed |
| 818 | : null; |
| 819 | } |
| 820 | |
| 821 | /** |
| 822 | * Returns the span representing the currently-matched delimiter, |
| 823 | * if matched, otherwise `null`. |
| 824 | * |
| 825 | * Note that for freeform blocks this will return a span of length |
| 826 | * zero, since there is no explicit block delimiter. |
| 827 | * |
| 828 | * Example: |
| 829 | * |
| 830 | * $scanner = Block_Scanner::create( '<!-- wp:void /-->' ); |
| 831 | * null === $scanner->get_span(); |
| 832 | * |
| 833 | * $scanner->next_delimiter(); |
| 834 | * WP_HTML_Span( 0, 17 ) === $scanner->get_span(); |
| 835 | * |
| 836 | * @return WP_HTML_Span|null Span of text in source text spanning matched delimiter. |
| 837 | */ |
| 838 | public function get_span(): ?WP_HTML_Span { |
| 839 | return new WP_HTML_Span( $this->delimiter_at, $this->delimiter_length ); |
| 840 | } |
| 841 | |
| 842 | // Debugging methods not meant for production use. |
| 843 | |
| 844 | /** |
| 845 | * Prints a debugging message showing the structure of the parsed delimiter. |
| 846 | * |
| 847 | * This is not meant to be used in production! |
| 848 | * |
| 849 | * @access private |
| 850 | */ |
| 851 | public function debug_print_structure(): void { |
| 852 | $c = ( ! defined( 'STDOUT' ) || posix_isatty( STDOUT ) ) |
| 853 | ? function ( $color = null ) { return $color; } // phpcs:ignore |
| 854 | : function ( $color ) { return ''; }; // phpcs:ignore |
| 855 | |
| 856 | if ( $this->is_block_type( 'core/freeform' ) ) { |
| 857 | $closer = static::CLOSER === $this->get_delimiter_type() ? '/' : ''; |
| 858 | echo "{$c( "\e[90m" )}<!-- "; // phpcs:ignore |
| 859 | echo "{$c( "\e[0;31m" )}{$closer}"; // phpcs:ignore |
| 860 | echo "{$c("\e[90m" )}wp:"; // phpcs:ignore |
| 861 | echo "{$c( "\e[0;34m" )}freeform"; // phpcs:ignore |
| 862 | echo "{$c( "\e[0;36m" )} {$c("\e[90m")}-->\n"; // phpcs:ignore |
| 863 | return; |
| 864 | } |
| 865 | |
| 866 | $namespace = substr( $this->source_text, $this->namespace_at, $this->namespace_length ); |
| 867 | $slash = 0 === $this->namespace_length ? '' : '/'; |
| 868 | $block_name = substr( $this->source_text, $this->name_at, $this->name_length ); |
| 869 | $closer = static::CLOSER === $this->type ? '/' : ''; |
| 870 | $json = substr( $this->source_text, $this->json_at, $this->json_length ); |
| 871 | |
| 872 | $opener_whitespace_at = $this->delimiter_at + 4; |
| 873 | $opener_whitespace_length = $this->namespace_at - 3 - $opener_whitespace_at - ( static::CLOSER === $this->type ? 1 : 0 ); |
| 874 | |
| 875 | $after_name_whitespace_at = $this->name_at + $this->name_length; |
| 876 | $after_name_whitespace_length = $this->json_at - $after_name_whitespace_at; |
| 877 | |
| 878 | $closing_whitespace_at = $this->json_at + $this->json_length; |
| 879 | $closing_whitespace_length = $this->delimiter_at + $this->delimiter_length - 3 - $closing_whitespace_at; |
| 880 | |
| 881 | if ( '/' === $this->source_text[ $this->delimiter_at + $this->delimiter_length - 4 ] ) { |
| 882 | $void_flag = '/'; |
| 883 | --$closing_whitespace_length; |
| 884 | } else { |
| 885 | $void_flag = ''; |
| 886 | } |
| 887 | |
| 888 | $w = function ( $whitespace ) use ( $c ) { |
| 889 | return $c( "\e[2;90m" ) . str_replace( array( ' ', "\t", "\f", "\r", "\n" ), array( '␣', '␉', '␌', '␍', '' ), $whitespace ); |
| 890 | }; |
| 891 | |
| 892 | echo "{$c( "\e[90m" )}<!--"; // phpcs:ignore |
| 893 | echo $w( substr( $this->source_text, $opener_whitespace_at, $opener_whitespace_length ) ); // phpcs:ignore |
| 894 | echo "{$c( "\e[0;31m" )}{$closer}"; // phpcs:ignore |
| 895 | echo "{$c("\e[90m" )}wp:{$c( "\e[2;34m" )}{$namespace}"; // phpcs:ignore |
| 896 | echo "{$c( "\e[2;90m" )}{$slash}"; // phpcs:ignore |
| 897 | echo "{$c( "\e[0;34m" )}{$block_name}"; // phpcs:ignore |
| 898 | echo $w( substr( $this->source_text, $after_name_whitespace_at, $after_name_whitespace_length ) ); // phpcs:ignore |
| 899 | echo "{$c("\e[0;2;32m" )}{$json}"; // phpcs:ignore |
| 900 | echo $w( substr( $this->source_text, $closing_whitespace_at, $closing_whitespace_length ) ); // phpcs:ignore |
| 901 | echo "{$c( "\e[0;36m" )}{$void_flag}{$c("\e[90m")}-->\n"; // phpcs:ignore |
| 902 | } |
| 903 | |
| 904 | // Constant declarations that would otherwise pollute the top of the class. |
| 905 | |
| 906 | /** |
| 907 | * Indicates that the block comment delimiter closes an open block. |
| 908 | */ |
| 909 | const CLOSER = 'closer'; |
| 910 | |
| 911 | /** |
| 912 | * Indicates that the parser started parsing a block comment delimiter, but |
| 913 | * the input document ended before it could finish. The document was likely truncated. |
| 914 | */ |
| 915 | const INCOMPLETE_INPUT = 'incomplete-input'; |
| 916 | |
| 917 | /** |
| 918 | * Indicates that the block comment delimiter opens a block. |
| 919 | */ |
| 920 | const OPENER = 'opener'; |
| 921 | |
| 922 | /** |
| 923 | * Indicates that the parser has not yet attempted to parse a block comment delimiter. |
| 924 | */ |
| 925 | const UNINITIALIZED = 'uninitialized'; |
| 926 | |
| 927 | /** |
| 928 | * Indicates that the block comment delimiter represents a void block |
| 929 | * with no inner content of any kind. |
| 930 | */ |
| 931 | const VOID = 'void'; |
| 932 | } |