Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
94.70% |
268 / 283 |
|
81.48% |
22 / 27 |
CRAP | |
0.00% |
0 / 1 |
| CommentUtils | |
94.70% |
268 / 283 |
|
81.48% |
22 / 27 |
165.91 | |
0.00% |
0 / 1 |
| __construct | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| isBlockElement | |
100.00% |
2 / 2 |
|
100.00% |
1 / 1 |
2 | |||
| isRenderingTransparentNode | |
100.00% |
22 / 22 |
|
100.00% |
1 / 1 |
11 | |||
| isOurGeneratedNode | |
100.00% |
5 / 5 |
|
100.00% |
1 / 1 |
4 | |||
| cantHaveElementChildren | |
100.00% |
18 / 18 |
|
100.00% |
1 / 1 |
7 | |||
| isCommentSeparator | |
100.00% |
17 / 17 |
|
100.00% |
1 / 1 |
13 | |||
| isCommentContent | |
100.00% |
7 / 7 |
|
100.00% |
1 / 1 |
3 | |||
| childIndexOf | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
2 | |||
| contains | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
3 | |||
| closestElement | |
100.00% |
6 / 6 |
|
100.00% |
1 / 1 |
3 | |||
| closestElementWithSibling | |
0.00% |
0 / 7 |
|
0.00% |
0 / 1 |
72 | |||
| getTranscludedFromElement | |
100.00% |
14 / 14 |
|
100.00% |
1 / 1 |
10 | |||
| getHeadlineNode | |
66.67% |
6 / 9 |
|
0.00% |
0 / 1 |
5.93 | |||
| htmlTrim | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| getIndentLevel | |
100.00% |
9 / 9 |
|
100.00% |
1 / 1 |
6 | |||
| getCoveredSiblings | |
100.00% |
13 / 13 |
|
100.00% |
1 / 1 |
5 | |||
| getFullyCoveredSiblings | |
100.00% |
17 / 17 |
|
100.00% |
1 / 1 |
5 | |||
| unwrapParsoidSections | |
100.00% |
6 / 6 |
|
100.00% |
1 / 1 |
3 | |||
| getTitleFromUrl | |
100.00% |
22 / 22 |
|
100.00% |
1 / 1 |
6 | |||
| linearWalk | |
100.00% |
11 / 11 |
|
100.00% |
1 / 1 |
5 | |||
| linearWalkBackwards | |
100.00% |
11 / 11 |
|
100.00% |
1 / 1 |
5 | |||
| getRangeFirstNode | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
3 | |||
| getRangeLastNode | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
3 | |||
| compareRanges | |
88.00% |
22 / 25 |
|
0.00% |
0 / 1 |
23.91 | |||
| compareRangesAlmostEqualBoundaries | |
100.00% |
28 / 28 |
|
100.00% |
1 / 1 |
15 | |||
| isSingleCommentSignedBy | |
100.00% |
15 / 15 |
|
100.00% |
1 / 1 |
9 | |||
| getNewTopicsSubscriptionId | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| 1 | <?php |
| 2 | |
| 3 | namespace MediaWiki\Extension\DiscussionTools; |
| 4 | |
| 5 | use LogicException; |
| 6 | use MediaWiki\Config\Config; |
| 7 | use MediaWiki\Extension\DiscussionTools\ThreadItem\ContentCommentItem; |
| 8 | use MediaWiki\Extension\DiscussionTools\ThreadItem\ContentThreadItem; |
| 9 | use MediaWiki\MainConfigNames; |
| 10 | use MediaWiki\Title\Title; |
| 11 | use Wikimedia\Assert\Assert; |
| 12 | use Wikimedia\Parsoid\DOM\Comment; |
| 13 | use Wikimedia\Parsoid\DOM\Element; |
| 14 | use Wikimedia\Parsoid\DOM\Node; |
| 15 | use Wikimedia\Parsoid\DOM\Text; |
| 16 | use Wikimedia\Parsoid\Utils\DOMCompat; |
| 17 | use Wikimedia\Parsoid\Utils\DOMUtils; |
| 18 | |
| 19 | class CommentUtils { |
| 20 | |
| 21 | private function __construct() { |
| 22 | } |
| 23 | |
| 24 | private const BLOCK_ELEMENT_TYPES = [ |
| 25 | 'div', 'p', |
| 26 | // Tables |
| 27 | 'table', 'tbody', 'thead', 'tfoot', 'caption', 'th', 'tr', 'td', |
| 28 | // Lists |
| 29 | 'ul', 'ol', 'li', 'dl', 'dt', 'dd', |
| 30 | // HTML5 heading content |
| 31 | 'h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'hgroup', |
| 32 | // HTML5 sectioning content |
| 33 | 'article', 'aside', 'body', 'nav', 'section', 'footer', 'header', 'figure', |
| 34 | 'figcaption', 'fieldset', 'details', 'blockquote', |
| 35 | // Other |
| 36 | 'hr', 'button', 'canvas', 'center', 'col', 'colgroup', 'embed', |
| 37 | 'map', 'object', 'pre', 'progress', 'video' |
| 38 | ]; |
| 39 | |
| 40 | /** |
| 41 | * @param Node $node |
| 42 | * @return bool Node is a block element |
| 43 | */ |
| 44 | public static function isBlockElement( Node $node ): bool { |
| 45 | return $node instanceof Element && |
| 46 | in_array( strtolower( $node->tagName ), static::BLOCK_ELEMENT_TYPES, true ); |
| 47 | } |
| 48 | |
| 49 | private const SOL_TRANSPARENT_LINK_REGEX = |
| 50 | '/(?:^|\s)mw:PageProp\/(?:Category|redirect|Language)(?=$|\s)/D'; |
| 51 | |
| 52 | /** |
| 53 | * @param Node $node |
| 54 | * @return bool Node is considered a rendering-transparent node in Parsoid |
| 55 | */ |
| 56 | public static function isRenderingTransparentNode( Node $node ): bool { |
| 57 | $nextSibling = $node->nextSibling; |
| 58 | return ( |
| 59 | $node instanceof Comment || |
| 60 | ( $node instanceof Element && ( |
| 61 | strtolower( $node->tagName ) === 'meta' || |
| 62 | ( |
| 63 | strtolower( $node->tagName ) === 'link' && |
| 64 | preg_match( static::SOL_TRANSPARENT_LINK_REGEX, $node->getAttribute( 'rel' ) ?? '' ) |
| 65 | ) || |
| 66 | // Empty inline templates, e.g. tracking templates. (T269036) |
| 67 | // But not empty nodes that are just the start of a non-empty template about-group. (T290940) |
| 68 | ( |
| 69 | strtolower( $node->tagName ) === 'span' && |
| 70 | in_array( 'mw:Transclusion', explode( ' ', $node->getAttribute( 'typeof' ) ?? '' ), true ) && |
| 71 | !static::htmlTrim( DOMCompat::getInnerHTML( $node ) ) && |
| 72 | ( |
| 73 | !$nextSibling || !( $nextSibling instanceof Element ) || |
| 74 | // Maybe we should be checking all of the about-grouped nodes to see if they're empty, |
| 75 | // but that's prooobably not needed in practice, and it leads to a quadratic worst case. |
| 76 | $nextSibling->getAttribute( 'about' ) !== $node->getAttribute( 'about' ) |
| 77 | ) |
| 78 | ) |
| 79 | ) ) |
| 80 | ); |
| 81 | } |
| 82 | |
| 83 | /** |
| 84 | * @param Node $node |
| 85 | * @return bool Node was added to the page by DiscussionTools |
| 86 | */ |
| 87 | public static function isOurGeneratedNode( Node $node ): bool { |
| 88 | return $node instanceof Element && ( |
| 89 | DOMUtils::hasClass( $node, 'ext-discussiontools-init-replylink-buttons' ) || |
| 90 | $node->hasAttribute( 'data-mw-comment-start' ) || |
| 91 | $node->hasAttribute( 'data-mw-comment-end' ) |
| 92 | ); |
| 93 | } |
| 94 | |
| 95 | /** |
| 96 | * Elements which can't have element children (but some may have text content). |
| 97 | */ |
| 98 | private const NO_ELEMENT_CHILDREN_ELEMENT_TYPES = [ |
| 99 | // https://html.spec.whatwg.org/multipage/syntax.html#elements-2 |
| 100 | // Void elements |
| 101 | 'area', 'base', 'br', 'col', 'embed', 'hr', 'img', 'input', |
| 102 | 'link', 'meta', 'param', 'source', 'track', 'wbr', |
| 103 | // Raw text elements |
| 104 | 'script', 'style', |
| 105 | // Escapable raw text elements |
| 106 | 'textarea', 'title', |
| 107 | // Foreign elements |
| 108 | 'math', 'svg', |
| 109 | // Treated like text when scripting is enabled in the parser |
| 110 | // https://html.spec.whatwg.org/#the-noscript-element |
| 111 | 'noscript', |
| 112 | // Replaced elements (that aren't already included above) |
| 113 | // https://html.spec.whatwg.org/multipage/rendering.html#replaced-elements |
| 114 | // They might allow element children, but they aren't rendered on the page. |
| 115 | 'audio', 'canvas', 'iframe', 'object', 'video', |
| 116 | ]; |
| 117 | |
| 118 | /** |
| 119 | * @param Node $node |
| 120 | * @return bool If true, node can't have element children. If false, it's complicated. |
| 121 | */ |
| 122 | public static function cantHaveElementChildren( Node $node ): bool { |
| 123 | return ( |
| 124 | $node instanceof Comment || |
| 125 | ( $node instanceof Element && ( |
| 126 | in_array( strtolower( $node->tagName ), static::NO_ELEMENT_CHILDREN_ELEMENT_TYPES, true ) || |
| 127 | // Thumbnail wrappers generated by MediaTransformOutput::linkWrap (T301427), |
| 128 | // for compatibility with TimedMediaHandler. |
| 129 | // There is no better way to detect them, and we can't insert markers here, |
| 130 | // because the media DOM CSS depends on specific tag names and their order :( |
| 131 | // TODO See if we can remove this condition when wgParserEnableLegacyMediaDOM=false |
| 132 | // is enabled everywhere. |
| 133 | ( |
| 134 | in_array( strtolower( $node->tagName ), [ 'a', 'span' ], true ) && |
| 135 | $node->firstChild && |
| 136 | // We always step inside a child node so this can't be infinite, silly Phan |
| 137 | // @phan-suppress-next-line PhanInfiniteRecursion |
| 138 | static::cantHaveElementChildren( $node->firstChild ) |
| 139 | ) || |
| 140 | // Do not insert anything inside figures when using wgParserEnableLegacyMediaDOM=false, |
| 141 | // because their CSS can't handle it (T320285). |
| 142 | strtolower( $node->tagName ) === 'figure' |
| 143 | ) ) |
| 144 | ); |
| 145 | } |
| 146 | |
| 147 | /** |
| 148 | * Check whether the node is a comment separator (instead of a part of the comment). |
| 149 | */ |
| 150 | public static function isCommentSeparator( Node $node ): bool { |
| 151 | if ( !( $node instanceof Element ) ) { |
| 152 | return false; |
| 153 | } |
| 154 | |
| 155 | $tagName = strtolower( $node->tagName ); |
| 156 | if ( $tagName === 'br' || $tagName === 'hr' ) { |
| 157 | return true; |
| 158 | } |
| 159 | |
| 160 | // TemplateStyles followed by any of the others |
| 161 | if ( $node->nextSibling && |
| 162 | ( $tagName === 'link' || $tagName === 'style' ) && |
| 163 | self::isCommentSeparator( $node->nextSibling ) |
| 164 | ) { |
| 165 | return true; |
| 166 | } |
| 167 | |
| 168 | // Anything marked as not containing comments |
| 169 | // {{outdent}} templates |
| 170 | // {{tracked}} templates (T313097) |
| 171 | if ( DOMUtils::hasClass( $node, 'mw-notalk|outdent-template|mw-trackedTemplate' ) ) { |
| 172 | return true; |
| 173 | } |
| 174 | |
| 175 | // Wikitext definition list term markup (`;`) when used as a fake heading (T265964) |
| 176 | if ( $tagName === 'dl' && |
| 177 | count( $node->childNodes ) === 1 && |
| 178 | $node->firstChild instanceof Element && |
| 179 | strtolower( $node->firstChild->nodeName ) === 'dt' |
| 180 | ) { |
| 181 | return true; |
| 182 | } |
| 183 | |
| 184 | return false; |
| 185 | } |
| 186 | |
| 187 | /** |
| 188 | * Check whether the node is a comment content. It's a little vague what this means… |
| 189 | * |
| 190 | * @param Node $node Node, should be a leaf node (a node with no children) |
| 191 | * @return bool |
| 192 | */ |
| 193 | public static function isCommentContent( Node $node ): bool { |
| 194 | return ( |
| 195 | $node instanceof Text && |
| 196 | static::htmlTrim( $node->nodeValue ?? '' ) !== '' |
| 197 | ) || |
| 198 | ( |
| 199 | static::cantHaveElementChildren( $node ) |
| 200 | ); |
| 201 | } |
| 202 | |
| 203 | /** |
| 204 | * Get the index of $child in its parent |
| 205 | */ |
| 206 | public static function childIndexOf( Node $child ): int { |
| 207 | $i = 0; |
| 208 | while ( ( $child = $child->previousSibling ) ) { |
| 209 | $i++; |
| 210 | } |
| 211 | return $i; |
| 212 | } |
| 213 | |
| 214 | /** |
| 215 | * Check whether a Node contains (is an ancestor of) another Node (or is the same node) |
| 216 | */ |
| 217 | public static function contains( Node $ancestor, Node $descendant ): bool { |
| 218 | // TODO can we use Node->compareDocumentPosition() here maybe? |
| 219 | $node = $descendant; |
| 220 | while ( $node && $node !== $ancestor ) { |
| 221 | $node = $node->parentNode; |
| 222 | } |
| 223 | return $node === $ancestor; |
| 224 | } |
| 225 | |
| 226 | /** |
| 227 | * Find closest ancestor element using one of the given tag names. |
| 228 | * |
| 229 | * @param Node $node |
| 230 | * @param string[] $tagNames |
| 231 | * @return Element|null |
| 232 | */ |
| 233 | public static function closestElement( Node $node, array $tagNames ): ?Element { |
| 234 | do { |
| 235 | if ( |
| 236 | $node instanceof Element && |
| 237 | in_array( strtolower( $node->tagName ), $tagNames, true ) |
| 238 | ) { |
| 239 | return $node; |
| 240 | } |
| 241 | $node = $node->parentNode; |
| 242 | } while ( $node ); |
| 243 | return null; |
| 244 | } |
| 245 | |
| 246 | /** |
| 247 | * Find closest ancestor element that has sibling nodes |
| 248 | * |
| 249 | * @param Node $node |
| 250 | * @param string $direction Can be 'next', 'previous', or 'either' |
| 251 | * @return Element|null |
| 252 | */ |
| 253 | public static function closestElementWithSibling( Node $node, string $direction ): ?Element { |
| 254 | do { |
| 255 | if ( |
| 256 | $node instanceof Element && ( |
| 257 | ( $node->nextSibling && ( $direction === 'next' || $direction === 'either' ) ) || |
| 258 | ( $node->previousSibling && ( $direction === 'previous' || $direction === 'either' ) ) |
| 259 | ) |
| 260 | ) { |
| 261 | return $node; |
| 262 | } |
| 263 | $node = $node->parentNode; |
| 264 | } while ( $node ); |
| 265 | return null; |
| 266 | } |
| 267 | |
| 268 | /** |
| 269 | * Find the transclusion node which rendered the current node, if it exists. |
| 270 | * |
| 271 | * 1. Find the closest ancestor with an 'about' attribute |
| 272 | * 2. Find the main node of the about-group (first sibling with the same 'about' attribute) |
| 273 | * 3. If this is an mw:Transclusion node, return it; otherwise, go to step 1 |
| 274 | * |
| 275 | * @param Node $node |
| 276 | * @return Element|null Transclusion node, null if not found |
| 277 | */ |
| 278 | public static function getTranscludedFromElement( Node $node ): ?Element { |
| 279 | while ( $node ) { |
| 280 | // 1. |
| 281 | if ( |
| 282 | $node instanceof Element && |
| 283 | $node->getAttribute( 'about' ) && |
| 284 | preg_match( '/^#mwt\d+$/', $node->getAttribute( 'about' ) ?? '' ) |
| 285 | ) { |
| 286 | $about = $node->getAttribute( 'about' ); |
| 287 | |
| 288 | // 2. |
| 289 | while ( |
| 290 | ( $previousSibling = $node->previousSibling ) && |
| 291 | $previousSibling instanceof Element && |
| 292 | $previousSibling->getAttribute( 'about' ) === $about |
| 293 | ) { |
| 294 | $node = $previousSibling; |
| 295 | } |
| 296 | |
| 297 | // 3. |
| 298 | if ( |
| 299 | $node->getAttribute( 'typeof' ) && |
| 300 | in_array( 'mw:Transclusion', explode( ' ', $node->getAttribute( 'typeof' ) ?? '' ), true ) |
| 301 | ) { |
| 302 | break; |
| 303 | } |
| 304 | } |
| 305 | |
| 306 | $node = $node->parentNode; |
| 307 | } |
| 308 | return $node; |
| 309 | } |
| 310 | |
| 311 | /** |
| 312 | * Given a heading node, return the node on which the ID attribute is set. |
| 313 | * |
| 314 | * @param Element $heading Heading node (`<h1>`-`<h6>`) |
| 315 | * @return Element Headline node, normally also a `<h1>`-`<h6>` element. |
| 316 | * In integration tests and in JS, it can be a `<span class="mw-headline">` (see T363031). |
| 317 | */ |
| 318 | public static function getHeadlineNode( Element $heading ): Element { |
| 319 | // This code assumes that $wgFragmentMode is [ 'html5', 'legacy' ] or [ 'html5' ] |
| 320 | $headline = $heading; |
| 321 | |
| 322 | if ( $headline->hasAttribute( 'data-mw-comment-start' ) ) { |
| 323 | // HACK: For contaminated integration tests only (see T363031) |
| 324 | $headline = $headline->parentNode; |
| 325 | Assert::precondition( $headline !== null, 'data-mw-comment-start was attached to a heading' ); |
| 326 | } |
| 327 | |
| 328 | if ( !$headline->getAttribute( 'id' ) && !$headline->getAttribute( 'data-mw-anchor' ) ) { |
| 329 | // HACK: For outdated integration tests only (see T363031) |
| 330 | $headline = DOMCompat::querySelector( $headline, '.mw-headline' ); |
| 331 | if ( !$headline ) { |
| 332 | $headline = $heading; |
| 333 | } |
| 334 | } |
| 335 | |
| 336 | return $headline; |
| 337 | } |
| 338 | |
| 339 | /** |
| 340 | * Trim ASCII whitespace, as defined in the HTML spec. |
| 341 | */ |
| 342 | public static function htmlTrim( string $str ): string { |
| 343 | // https://infra.spec.whatwg.org/#ascii-whitespace |
| 344 | return trim( $str, "\t\n\f\r " ); |
| 345 | } |
| 346 | |
| 347 | /** |
| 348 | * Get the indent level of $node, relative to $rootNode. |
| 349 | * |
| 350 | * The indent level is the number of lists inside of which it is nested. |
| 351 | */ |
| 352 | public static function getIndentLevel( Node $node, Element $rootNode ): int { |
| 353 | $indent = 0; |
| 354 | while ( $node ) { |
| 355 | if ( $node === $rootNode ) { |
| 356 | break; |
| 357 | } |
| 358 | $tagName = $node instanceof Element ? strtolower( $node->tagName ) : null; |
| 359 | if ( $tagName === 'li' || $tagName === 'dd' ) { |
| 360 | $indent++; |
| 361 | } |
| 362 | $node = $node->parentNode; |
| 363 | } |
| 364 | return $indent; |
| 365 | } |
| 366 | |
| 367 | /** |
| 368 | * Get an array of sibling nodes that contain parts of the given range. |
| 369 | * |
| 370 | * @param ImmutableRange $range |
| 371 | * @return Node[] |
| 372 | */ |
| 373 | public static function getCoveredSiblings( ImmutableRange $range ): array { |
| 374 | $ancestor = $range->commonAncestorContainer; |
| 375 | |
| 376 | // Convert to array early because apparently NodeList acts like a linked list |
| 377 | // and accessing items by index is slow |
| 378 | $siblings = iterator_to_array( $ancestor->childNodes ); |
| 379 | $start = 0; |
| 380 | $end = count( $siblings ) - 1; |
| 381 | |
| 382 | // Find first of the siblings that contains the item |
| 383 | if ( $ancestor === $range->startContainer ) { |
| 384 | $start = $range->startOffset; |
| 385 | } else { |
| 386 | while ( !static::contains( $siblings[ $start ], $range->startContainer ) ) { |
| 387 | $start++; |
| 388 | } |
| 389 | } |
| 390 | |
| 391 | // Find last of the siblings that contains the item |
| 392 | if ( $ancestor === $range->endContainer ) { |
| 393 | $end = $range->endOffset - 1; |
| 394 | } else { |
| 395 | while ( !static::contains( $siblings[ $end ], $range->endContainer ) ) { |
| 396 | $end--; |
| 397 | } |
| 398 | } |
| 399 | |
| 400 | return array_slice( $siblings, $start, $end - $start + 1 ); |
| 401 | } |
| 402 | |
| 403 | /** |
| 404 | * Get the nodes (if any) that contain the given thread item, and nothing else. |
| 405 | * |
| 406 | * @param ContentThreadItem $item |
| 407 | * @param ?Node $excludedAncestorNode Node that shouldn't be included in the result, even if it |
| 408 | * contains the item and nothing else. This is intended to avoid traversing outside of a node |
| 409 | * which is a container for all the thread items. |
| 410 | * @return Node[]|null |
| 411 | */ |
| 412 | public static function getFullyCoveredSiblings( |
| 413 | ContentThreadItem $item, ?Node $excludedAncestorNode = null |
| 414 | ): ?array { |
| 415 | $siblings = static::getCoveredSiblings( $item->getRange() ); |
| 416 | |
| 417 | $makeRange = static function ( $siblings ) { |
| 418 | return new ImmutableRange( |
| 419 | $siblings[0]->parentNode, |
| 420 | CommentUtils::childIndexOf( $siblings[0] ), |
| 421 | end( $siblings )->parentNode, |
| 422 | CommentUtils::childIndexOf( end( $siblings ) ) + 1 |
| 423 | ); |
| 424 | }; |
| 425 | |
| 426 | $matches = static::compareRanges( $makeRange( $siblings ), $item->getRange() ) === 'equal'; |
| 427 | |
| 428 | if ( $matches ) { |
| 429 | // If these are all of the children (or the only child), go up one more level |
| 430 | while ( |
| 431 | ( $parent = $siblings[ 0 ]->parentNode ) && |
| 432 | $parent !== $excludedAncestorNode && |
| 433 | static::compareRanges( $makeRange( [ $parent ] ), $item->getRange() ) === 'equal' |
| 434 | ) { |
| 435 | $siblings = [ $parent ]; |
| 436 | } |
| 437 | return $siblings; |
| 438 | } |
| 439 | return null; |
| 440 | } |
| 441 | |
| 442 | /** |
| 443 | * Unwrap Parsoid sections |
| 444 | * |
| 445 | * @param Element $element Parent element, e.g. document body |
| 446 | */ |
| 447 | public static function unwrapParsoidSections( Element $element ): void { |
| 448 | $sections = DOMCompat::querySelectorAll( $element, 'section[data-mw-section-id]' ); |
| 449 | foreach ( $sections as $section ) { |
| 450 | $parent = $section->parentNode; |
| 451 | while ( $section->firstChild ) { |
| 452 | $parent->insertBefore( $section->firstChild, $section ); |
| 453 | } |
| 454 | $parent->removeChild( $section ); |
| 455 | } |
| 456 | } |
| 457 | |
| 458 | /** |
| 459 | * Get a MediaWiki page title from a URL |
| 460 | * |
| 461 | * @param string $url Relative URL (from a `href` attribute) |
| 462 | * @param Config $config Config settings needed to resolve the relative URL |
| 463 | * @return string|null |
| 464 | */ |
| 465 | public static function getTitleFromUrl( string $url, Config $config ): ?string { |
| 466 | // Protocol-relative URLs are handled really badly by parse_url() |
| 467 | if ( str_starts_with( $url, '//' ) ) { |
| 468 | $url = "http:$url"; |
| 469 | } |
| 470 | |
| 471 | $bits = parse_url( $url ); |
| 472 | $query = wfCgiToArray( $bits['query'] ?? '' ); |
| 473 | if ( isset( $query['title'] ) ) { |
| 474 | return $query['title']; |
| 475 | } |
| 476 | |
| 477 | // TODO: Set the correct base in the document? |
| 478 | $articlePath = $config->get( MainConfigNames::ArticlePath ); |
| 479 | if ( str_starts_with( $url, './' ) ) { |
| 480 | // Assume this is URL in the format used by Parsoid documents |
| 481 | $url = substr( $url, 2 ); |
| 482 | $path = str_replace( '$1', $url, $articlePath ); |
| 483 | } elseif ( !str_contains( $url, '://' ) ) { |
| 484 | // Assume this is URL in the format used by legacy parser documents |
| 485 | $path = $url; |
| 486 | } else { |
| 487 | // External link |
| 488 | $path = $bits['path'] ?? ''; |
| 489 | } |
| 490 | |
| 491 | $articlePathRegexp = '/^' . str_replace( |
| 492 | '\\$1', |
| 493 | '([^?]*)', |
| 494 | preg_quote( $articlePath, '/' ) |
| 495 | ) . '/'; |
| 496 | $matches = null; |
| 497 | if ( preg_match( $articlePathRegexp, $path, $matches ) ) { |
| 498 | return rawurldecode( $matches[1] ); |
| 499 | } |
| 500 | return null; |
| 501 | } |
| 502 | |
| 503 | /** |
| 504 | * Traverse the document in depth-first order, calling the callback whenever entering and leaving |
| 505 | * a node. The walk starts before the given node and ends when callback returns a truthy value, or |
| 506 | * after reaching the end of the document. |
| 507 | * |
| 508 | * You might also think about this as processing XML token stream linearly (rather than XML |
| 509 | * nodes), as if we were parsing the document. |
| 510 | * |
| 511 | * @param Node $node Node to start at |
| 512 | * @param callable $callback Function accepting two arguments: $event ('enter' or 'leave') and |
| 513 | * $node (Node) |
| 514 | * @return mixed Final return value of the callback |
| 515 | */ |
| 516 | public static function linearWalk( Node $node, callable $callback ) { |
| 517 | $result = null; |
| 518 | [ $withinNode, $beforeNode ] = [ $node->parentNode, $node ]; |
| 519 | |
| 520 | while ( $beforeNode || $withinNode ) { |
| 521 | if ( $beforeNode ) { |
| 522 | $result = $callback( 'enter', $beforeNode ); |
| 523 | [ $withinNode, $beforeNode ] = [ $beforeNode, $beforeNode->firstChild ]; |
| 524 | } else { |
| 525 | $result = $callback( 'leave', $withinNode ); |
| 526 | [ $withinNode, $beforeNode ] = [ $withinNode->parentNode, $withinNode->nextSibling ]; |
| 527 | } |
| 528 | |
| 529 | if ( $result ) { |
| 530 | return $result; |
| 531 | } |
| 532 | } |
| 533 | return $result; |
| 534 | } |
| 535 | |
| 536 | /** |
| 537 | * Like #linearWalk, but it goes backwards. |
| 538 | * |
| 539 | * @inheritDoc ::linearWalk() |
| 540 | */ |
| 541 | public static function linearWalkBackwards( Node $node, callable $callback ) { |
| 542 | $result = null; |
| 543 | [ $withinNode, $beforeNode ] = [ $node->parentNode, $node ]; |
| 544 | |
| 545 | while ( $beforeNode || $withinNode ) { |
| 546 | if ( $beforeNode ) { |
| 547 | $result = $callback( 'enter', $beforeNode ); |
| 548 | [ $withinNode, $beforeNode ] = [ $beforeNode, $beforeNode->lastChild ]; |
| 549 | } else { |
| 550 | $result = $callback( 'leave', $withinNode ); |
| 551 | [ $withinNode, $beforeNode ] = [ $withinNode->parentNode, $withinNode->previousSibling ]; |
| 552 | } |
| 553 | |
| 554 | if ( $result ) { |
| 555 | return $result; |
| 556 | } |
| 557 | } |
| 558 | return $result; |
| 559 | } |
| 560 | |
| 561 | /** |
| 562 | * @param ImmutableRange $range (must not be collapsed) |
| 563 | * @return Node |
| 564 | */ |
| 565 | public static function getRangeFirstNode( ImmutableRange $range ): Node { |
| 566 | Assert::precondition( !$range->collapsed, 'Range is not collapsed' ); |
| 567 | // PHP bug: childNodes can be null |
| 568 | return $range->startContainer->childNodes && $range->startContainer->childNodes->length ? |
| 569 | $range->startContainer->childNodes[ $range->startOffset ] : |
| 570 | $range->startContainer; |
| 571 | } |
| 572 | |
| 573 | /** |
| 574 | * @param ImmutableRange $range (must not be collapsed) |
| 575 | * @return Node |
| 576 | */ |
| 577 | public static function getRangeLastNode( ImmutableRange $range ): Node { |
| 578 | Assert::precondition( !$range->collapsed, 'Range is not collapsed' ); |
| 579 | // PHP bug: childNodes can be null |
| 580 | return $range->endContainer->childNodes && $range->endContainer->childNodes->length ? |
| 581 | $range->endContainer->childNodes[ $range->endOffset - 1 ] : |
| 582 | $range->endContainer; |
| 583 | } |
| 584 | |
| 585 | /** |
| 586 | * Check whether two ranges overlap, and how. |
| 587 | * |
| 588 | * Includes a hack to check for "almost equal" ranges (whose start/end boundaries only differ by |
| 589 | * "uninteresting" nodes that we ignore when detecting comments), and treat them as equal. |
| 590 | * |
| 591 | * Illustration of return values: |
| 592 | * [ equal ] |
| 593 | * |[ contained ]| |
| 594 | * [ | contains | ] |
| 595 | * [overlap|start] | |
| 596 | * | [overlap|end] |
| 597 | * [before] | | |
| 598 | * | | [after] |
| 599 | * |
| 600 | * @param ImmutableRange $a |
| 601 | * @param ImmutableRange $b |
| 602 | * @return string One of: |
| 603 | * - 'equal': Ranges A and B are equal |
| 604 | * - 'contains': Range A contains range B |
| 605 | * - 'contained': Range A is contained within range B |
| 606 | * - 'after': Range A is before range B |
| 607 | * - 'before': Range A is after range B |
| 608 | * - 'overlapstart': Start of range A overlaps range B |
| 609 | * - 'overlapend': End of range A overlaps range B |
| 610 | */ |
| 611 | public static function compareRanges( ImmutableRange $a, ImmutableRange $b ): string { |
| 612 | // Compare the positions of: start of A to start of B, start of A to end of B, and so on. |
| 613 | // Watch out, the constant names are the opposite of what they should be. |
| 614 | $startToStart = $a->compareBoundaryPoints( ImmutableRange::START_TO_START, $b ); |
| 615 | $startToEnd = $a->compareBoundaryPoints( ImmutableRange::END_TO_START, $b ); |
| 616 | $endToStart = $a->compareBoundaryPoints( ImmutableRange::START_TO_END, $b ); |
| 617 | $endToEnd = $a->compareBoundaryPoints( ImmutableRange::END_TO_END, $b ); |
| 618 | |
| 619 | // Handle almost equal ranges: When start or end boundary points of the two ranges are different, |
| 620 | // but only differ by "uninteresting" nodes, treat them as equal instead. |
| 621 | if ( |
| 622 | ( $startToStart < 0 && static::compareRangesAlmostEqualBoundaries( $a, $b, 'start' ) ) || |
| 623 | ( $startToStart > 0 && static::compareRangesAlmostEqualBoundaries( $b, $a, 'start' ) ) |
| 624 | ) { |
| 625 | $startToStart = 0; |
| 626 | } |
| 627 | if ( |
| 628 | ( $endToEnd < 0 && static::compareRangesAlmostEqualBoundaries( $a, $b, 'end' ) ) || |
| 629 | ( $endToEnd > 0 && static::compareRangesAlmostEqualBoundaries( $b, $a, 'end' ) ) |
| 630 | ) { |
| 631 | $endToEnd = 0; |
| 632 | } |
| 633 | |
| 634 | if ( $startToStart === 0 && $endToEnd === 0 ) { |
| 635 | return 'equal'; |
| 636 | } |
| 637 | if ( $startToStart <= 0 && $endToEnd >= 0 ) { |
| 638 | return 'contains'; |
| 639 | } |
| 640 | if ( $startToStart >= 0 && $endToEnd <= 0 ) { |
| 641 | return 'contained'; |
| 642 | } |
| 643 | if ( $startToEnd >= 0 ) { |
| 644 | return 'after'; |
| 645 | } |
| 646 | if ( $endToStart <= 0 ) { |
| 647 | return 'before'; |
| 648 | } |
| 649 | if ( $startToStart > 0 && $startToEnd < 0 && $endToEnd >= 0 ) { |
| 650 | return 'overlapstart'; |
| 651 | } |
| 652 | if ( $endToEnd < 0 && $endToStart > 0 && $startToStart <= 0 ) { |
| 653 | return 'overlapend'; |
| 654 | } |
| 655 | |
| 656 | throw new LogicException( 'Unreachable' ); |
| 657 | } |
| 658 | |
| 659 | /** |
| 660 | * Check if the given boundary points of ranges A and B are almost equal (only differing by |
| 661 | * uninteresting nodes). |
| 662 | * |
| 663 | * Boundary of A must be before the boundary of B in the tree. |
| 664 | * |
| 665 | * @param ImmutableRange $a |
| 666 | * @param ImmutableRange $b |
| 667 | * @param string $boundary 'start' or 'end' |
| 668 | * @return bool |
| 669 | */ |
| 670 | private static function compareRangesAlmostEqualBoundaries( |
| 671 | ImmutableRange $a, ImmutableRange $b, string $boundary |
| 672 | ): bool { |
| 673 | // This code is awful, but several attempts to rewrite it made it even worse. |
| 674 | // You're welcome to give it a try. |
| 675 | |
| 676 | $from = $boundary === 'end' ? static::getRangeLastNode( $a ) : static::getRangeFirstNode( $a ); |
| 677 | $to = $boundary === 'end' ? static::getRangeLastNode( $b ) : static::getRangeFirstNode( $b ); |
| 678 | |
| 679 | $skipNode = null; |
| 680 | if ( $boundary === 'end' ) { |
| 681 | $skipNode = $from; |
| 682 | } |
| 683 | |
| 684 | $foundContent = false; |
| 685 | static::linearWalk( |
| 686 | $from, |
| 687 | static function ( string $event, Node $n ) use ( |
| 688 | $from, $to, $boundary, &$skipNode, &$foundContent |
| 689 | ) { |
| 690 | if ( $n === $to && $event === ( $boundary === 'end' ? 'leave' : 'enter' ) ) { |
| 691 | return true; |
| 692 | } |
| 693 | if ( $skipNode ) { |
| 694 | if ( $n === $skipNode && $event === 'leave' ) { |
| 695 | $skipNode = null; |
| 696 | } |
| 697 | return; |
| 698 | } |
| 699 | |
| 700 | if ( $event === 'enter' ) { |
| 701 | if ( |
| 702 | CommentUtils::isCommentSeparator( $n ) || |
| 703 | CommentUtils::isRenderingTransparentNode( $n ) || |
| 704 | CommentUtils::isOurGeneratedNode( $n ) |
| 705 | ) { |
| 706 | $skipNode = $n; |
| 707 | |
| 708 | } elseif ( |
| 709 | CommentUtils::isCommentContent( $n ) |
| 710 | ) { |
| 711 | $foundContent = true; |
| 712 | return true; |
| 713 | } |
| 714 | } |
| 715 | } |
| 716 | ); |
| 717 | |
| 718 | return !$foundContent; |
| 719 | } |
| 720 | |
| 721 | /** |
| 722 | * Check whether the last item in the thread item set is a properly signed comment by |
| 723 | * the expected author (that is: there is a signature, and either there's nothing following the |
| 724 | * signature, or there's some text within the same paragraph that was detected as part of the same |
| 725 | * comment). |
| 726 | */ |
| 727 | public static function isSingleCommentSignedBy( |
| 728 | ContentThreadItemSet $itemSet, |
| 729 | string $author, |
| 730 | Element $rootNode |
| 731 | ): bool { |
| 732 | $items = $itemSet->getThreadItems(); |
| 733 | |
| 734 | if ( $items ) { |
| 735 | $lastItem = end( $items ); |
| 736 | // Check that we've detected a comment first, not just headings (T304377) |
| 737 | if ( !( $lastItem instanceof ContentCommentItem && $lastItem->getAuthor() === $author ) ) { |
| 738 | return false; |
| 739 | } |
| 740 | |
| 741 | $commentRange = $lastItem->getRange(); |
| 742 | $bodyRange = new ImmutableRange( |
| 743 | $rootNode, 0, $rootNode, count( $rootNode->childNodes ) |
| 744 | ); |
| 745 | |
| 746 | // Only check that the end of the comment range is at the end of the body range. |
| 747 | // We don't care about preceding headings, comments, or other content (T363285). |
| 748 | // This is a simplified fragment of static::compareRanges(). |
| 749 | $cmp = $commentRange->compareBoundaryPoints( ImmutableRange::END_TO_END, $bodyRange ); |
| 750 | if ( |
| 751 | $cmp === 0 || |
| 752 | ( $cmp < 0 && static::compareRangesAlmostEqualBoundaries( $commentRange, $bodyRange, 'end' ) ) || |
| 753 | ( $cmp > 0 && static::compareRangesAlmostEqualBoundaries( $bodyRange, $commentRange, 'end' ) ) |
| 754 | ) { |
| 755 | // New comment includes a signature in the proper place |
| 756 | return true; |
| 757 | } |
| 758 | } |
| 759 | |
| 760 | return false; |
| 761 | } |
| 762 | |
| 763 | /** |
| 764 | * Get the ID for a new topics subscription from a page title |
| 765 | * |
| 766 | * @param Title $title Page title |
| 767 | * @return string ID for a new topics subscription |
| 768 | */ |
| 769 | public static function getNewTopicsSubscriptionId( Title $title ) { |
| 770 | return "p-topics-{$title->getNamespace()}:{$title->getDBkey()}"; |
| 771 | } |
| 772 | } |