Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
0.00% |
0 / 145 |
|
0.00% |
0 / 15 |
CRAP | |
0.00% |
0 / 1 |
| WTSUtils | |
0.00% |
0 / 145 |
|
0.00% |
0 / 15 |
5700 | |
0.00% |
0 / 1 |
| isValidSep | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| hasValidTagWidths | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
6 | |||
| getAttributeKVArray | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
6 | |||
| mkTagTk | |
0.00% |
0 / 6 |
|
0.00% |
0 / 1 |
2 | |||
| mkEndTagTk | |
0.00% |
0 / 6 |
|
0.00% |
0 / 1 |
2 | |||
| getShadowInfo | |
0.00% |
0 / 24 |
|
0.00% |
0 / 1 |
42 | |||
| getAttributeShadowInfo | |
0.00% |
0 / 5 |
|
0.00% |
0 / 1 |
2 | |||
| commentWT | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| nextToDeletedBlockNodeInWT | |
0.00% |
0 / 14 |
|
0.00% |
0 / 1 |
90 | |||
| precedingSpaceSuppressesIndentPre | |
0.00% |
0 / 8 |
|
0.00% |
0 / 1 |
42 | |||
| traceNodeName | |
0.00% |
0 / 8 |
|
0.00% |
0 / 1 |
42 | |||
| origSrcValidInEditedContext | |
0.00% |
0 / 37 |
|
0.00% |
0 / 1 |
552 | |||
| dsrContainsOpenExtendedRangeAnnotationTag | |
0.00% |
0 / 11 |
|
0.00% |
0 / 1 |
56 | |||
| getAttrFromDataMw | |
0.00% |
0 / 18 |
|
0.00% |
0 / 1 |
72 | |||
| escapeNowikiTags | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| 1 | <?php |
| 2 | declare( strict_types = 1 ); |
| 3 | |
| 4 | namespace Wikimedia\Parsoid\Html2Wt; |
| 5 | |
| 6 | use Wikimedia\Assert\UnreachableException; |
| 7 | use Wikimedia\Parsoid\Core\DomSourceRange; |
| 8 | use Wikimedia\Parsoid\DOM\Element; |
| 9 | use Wikimedia\Parsoid\DOM\Node; |
| 10 | use Wikimedia\Parsoid\DOM\Text; |
| 11 | use Wikimedia\Parsoid\NodeData\DataMw; |
| 12 | use Wikimedia\Parsoid\Tokens\EndTagTk; |
| 13 | use Wikimedia\Parsoid\Tokens\KV; |
| 14 | use Wikimedia\Parsoid\Tokens\TagTk; |
| 15 | use Wikimedia\Parsoid\Utils\DiffDOMUtils; |
| 16 | use Wikimedia\Parsoid\Utils\DOMCompat; |
| 17 | use Wikimedia\Parsoid\Utils\DOMDataUtils; |
| 18 | use Wikimedia\Parsoid\Utils\DOMUtils; |
| 19 | use Wikimedia\Parsoid\Utils\PHPUtils; |
| 20 | use Wikimedia\Parsoid\Utils\WTUtils; |
| 21 | |
| 22 | class WTSUtils { |
| 23 | |
| 24 | public static function isValidSep( string $sep ): bool { |
| 25 | /* TODO (Anomie) |
| 26 | You might be able to simplify the regex a bit using a no-backtracking group: |
| 27 | '/^(?>\s*<!--.*?-->)*\s*$/s' |
| 28 | Although I'm not sure that'll actually run faster.*/ |
| 29 | return (bool)preg_match( '/^(\s|<!--([^\-]|-(?!->))*-->)*$/uD', $sep ); |
| 30 | } |
| 31 | |
| 32 | public static function hasValidTagWidths( ?DomSourceRange $dsr ): bool { |
| 33 | return $dsr !== null && $dsr->hasValidTagWidths(); |
| 34 | } |
| 35 | |
| 36 | /** |
| 37 | * Get the attributes on a node in an array of KV objects. |
| 38 | * |
| 39 | * @param Element $node |
| 40 | * @return KV[] |
| 41 | */ |
| 42 | public static function getAttributeKVArray( Element $node ): array { |
| 43 | $kvs = []; |
| 44 | foreach ( DOMUtils::attributes( $node ) as $name => $value ) { |
| 45 | $kvs[] = new KV( $name, $value ); |
| 46 | } |
| 47 | return $kvs; |
| 48 | } |
| 49 | |
| 50 | /** |
| 51 | * Create a `TagTk` corresponding to a DOM node. |
| 52 | * |
| 53 | * @param Element $node |
| 54 | * @return TagTk |
| 55 | */ |
| 56 | public static function mkTagTk( Element $node ): TagTk { |
| 57 | $attribKVs = self::getAttributeKVArray( $node ); |
| 58 | return new TagTk( |
| 59 | DOMCompat::nodeName( $node ), |
| 60 | $attribKVs, |
| 61 | DOMDataUtils::getDataParsoid( $node ) |
| 62 | ); |
| 63 | } |
| 64 | |
| 65 | /** |
| 66 | * Create a `EndTagTk` corresponding to a DOM node. |
| 67 | * |
| 68 | * @param Element $node |
| 69 | * @return EndTagTk |
| 70 | */ |
| 71 | public static function mkEndTagTk( Element $node ): EndTagTk { |
| 72 | $attribKVs = self::getAttributeKVArray( $node ); |
| 73 | return new EndTagTk( |
| 74 | DOMCompat::nodeName( $node ), |
| 75 | $attribKVs, |
| 76 | DOMDataUtils::getDataParsoid( $node ) |
| 77 | ); |
| 78 | } |
| 79 | |
| 80 | /** |
| 81 | * For new elements, attrs are always considered modified. However, For |
| 82 | * old elements, we only consider an attribute modified if we have shadow |
| 83 | * info for it and it doesn't match the current value. |
| 84 | * Returns array with data: |
| 85 | * [ |
| 86 | * value => mixed, |
| 87 | * modified => bool (If the value of the attribute changed since we parsed the wikitext), |
| 88 | * fromsrc => bool (Whether we got the value from source-based roundtripping) |
| 89 | * ] |
| 90 | * |
| 91 | * @param Element $node |
| 92 | * @param string $name |
| 93 | * @param ?string $curVal |
| 94 | * @return array |
| 95 | */ |
| 96 | public static function getShadowInfo( Element $node, string $name, ?string $curVal ): array { |
| 97 | $dp = DOMDataUtils::getDataParsoid( $node ); |
| 98 | |
| 99 | // Not the case, continue regular round-trip information. |
| 100 | if ( !isset( $dp->a ) || !array_key_exists( $name, $dp->a ) ) { |
| 101 | return [ |
| 102 | 'value' => $curVal, |
| 103 | // Mark as modified if a new element |
| 104 | 'modified' => WTUtils::isNewElt( $node ), |
| 105 | 'fromsrc' => false |
| 106 | ]; |
| 107 | } elseif ( $dp->a[$name] !== $curVal ) { |
| 108 | return [ |
| 109 | 'value' => $curVal, |
| 110 | 'modified' => true, |
| 111 | 'fromsrc' => false |
| 112 | ]; |
| 113 | } elseif ( !isset( $dp->sa ) || !array_key_exists( $name, $dp->sa ) ) { |
| 114 | return [ |
| 115 | 'value' => $curVal, |
| 116 | 'modified' => false, |
| 117 | 'fromsrc' => false |
| 118 | ]; |
| 119 | } else { |
| 120 | return [ |
| 121 | 'value' => $dp->sa[$name], |
| 122 | 'modified' => false, |
| 123 | 'fromsrc' => true |
| 124 | ]; |
| 125 | } |
| 126 | } |
| 127 | |
| 128 | /** |
| 129 | * Get shadowed information about an attribute on a node. |
| 130 | * Returns array with data: |
| 131 | * [ |
| 132 | * value => mixed, |
| 133 | * modified => bool (If the value of the attribute changed since we parsed the wikitext), |
| 134 | * fromsrc => bool (Whether we got the value from source-based roundtripping) |
| 135 | * ] |
| 136 | * |
| 137 | * @param Element $node |
| 138 | * @param string $name |
| 139 | * @return array |
| 140 | */ |
| 141 | public static function getAttributeShadowInfo( Element $node, string $name ): array { |
| 142 | return self::getShadowInfo( |
| 143 | $node, |
| 144 | $name, |
| 145 | DOMCompat::getAttribute( $node, $name ) |
| 146 | ); |
| 147 | } |
| 148 | |
| 149 | public static function commentWT( string $comment ): string { |
| 150 | return '<!--' . WTUtils::decodeComment( $comment ) . '-->'; |
| 151 | } |
| 152 | |
| 153 | /** |
| 154 | * In wikitext, did origNode occur next to a block node which has been |
| 155 | * deleted? While looking for next, we look past DOM nodes that are |
| 156 | * transparent in rendering. (See emitsSolTransparentSingleLineWT for |
| 157 | * which nodes.) |
| 158 | * |
| 159 | * @param ?Node $origNode |
| 160 | * @param bool $before |
| 161 | * @return bool |
| 162 | */ |
| 163 | public static function nextToDeletedBlockNodeInWT( |
| 164 | ?Node $origNode, bool $before |
| 165 | ): bool { |
| 166 | if ( !$origNode || DOMUtils::atTheTop( $origNode ) ) { |
| 167 | return false; |
| 168 | } |
| 169 | |
| 170 | while ( true ) { |
| 171 | // Find the nearest node that shows up in HTML (ignore nodes that show up |
| 172 | // in wikitext but don't affect sol-state or HTML rendering -- note that |
| 173 | // whitespace is being ignored, but that whitespace occurs between block nodes). |
| 174 | $node = $origNode; |
| 175 | do { |
| 176 | $node = $before ? $node->previousSibling : $node->nextSibling; |
| 177 | if ( DiffUtils::maybeDeletedNode( $node ) ) { |
| 178 | return DiffUtils::isDeletedBlockNode( $node ); |
| 179 | } |
| 180 | } while ( $node && WTUtils::emitsSolTransparentSingleLineWT( $node ) ); |
| 181 | |
| 182 | if ( $node ) { |
| 183 | return false; |
| 184 | } else { |
| 185 | // Walk up past zero-width wikitext parents |
| 186 | $node = $origNode->parentNode; |
| 187 | if ( !WTUtils::isZeroWidthWikitextElt( $node ) ) { |
| 188 | // If the parent occupies space in wikitext, |
| 189 | // clearly, we are not next to a deleted block node! |
| 190 | // We'll eventually hit BODY here and return. |
| 191 | return false; |
| 192 | } |
| 193 | $origNode = $node; |
| 194 | } |
| 195 | } |
| 196 | } |
| 197 | |
| 198 | /** |
| 199 | * Check if whitespace preceding this node would NOT trigger an indent-pre. |
| 200 | * |
| 201 | * @param Node $node |
| 202 | * @param Node $sepNode |
| 203 | * @return bool |
| 204 | */ |
| 205 | public static function precedingSpaceSuppressesIndentPre( Node $node, Node $sepNode ): bool { |
| 206 | if ( $node !== $sepNode && $node instanceof Text ) { |
| 207 | // if node is the same as sepNode, then the separator text |
| 208 | // at the beginning of it has been stripped out already, and |
| 209 | // we cannot use it to test it for indent-pre safety |
| 210 | return (bool)preg_match( '/^[ \t]*\n/', $node->nodeValue ); |
| 211 | } elseif ( DOMCompat::nodeName( $node ) === 'br' ) { |
| 212 | return true; |
| 213 | } elseif ( WTUtils::isFirstEncapsulationWrapperNode( $node ) ) { |
| 214 | DOMUtils::assertElt( $node ); |
| 215 | // Dont try any harder than this |
| 216 | return !$node->hasChildNodes() || DOMCompat::getInnerHTML( $node )[0] === "\n"; |
| 217 | } else { |
| 218 | return WTUtils::isBlockNodeWithVisibleWT( $node ); |
| 219 | } |
| 220 | } |
| 221 | |
| 222 | public static function traceNodeName( Node $node ): string { |
| 223 | switch ( $node->nodeType ) { |
| 224 | case XML_ELEMENT_NODE: |
| 225 | return ( DiffUtils::isDiffMarker( $node ) ) ? 'DIFF_MARK' : 'NODE: ' . DOMCompat::nodeName( $node ); |
| 226 | case XML_TEXT_NODE: |
| 227 | return 'TEXT: ' . PHPUtils::jsonEncode( $node->nodeValue ); |
| 228 | case XML_COMMENT_NODE: |
| 229 | return 'CMT : ' . PHPUtils::jsonEncode( self::commentWT( $node->nodeValue ) ); |
| 230 | default: |
| 231 | return DOMCompat::nodeName( $node ); |
| 232 | } |
| 233 | } |
| 234 | |
| 235 | /** |
| 236 | * In selser mode, check if an unedited node's wikitext from source wikitext |
| 237 | * is reusable as is. |
| 238 | * |
| 239 | * @param SerializerState $state |
| 240 | * @param Node $node |
| 241 | * @return bool |
| 242 | */ |
| 243 | public static function origSrcValidInEditedContext( SerializerState $state, Node $node ): bool { |
| 244 | $env = $state->getEnv(); |
| 245 | $prev = null; |
| 246 | |
| 247 | if ( WTUtils::isRedirectLink( $node ) ) { |
| 248 | return DOMUtils::atTheTop( $node->parentNode ) && !$node->previousSibling; |
| 249 | } elseif ( self::dsrContainsOpenExtendedRangeAnnotationTag( $node, $state ) ) { |
| 250 | return false; |
| 251 | } elseif ( DOMCompat::nodeName( $node ) === 'th' || DOMCompat::nodeName( $node ) === 'td' ) { |
| 252 | DOMUtils::assertElt( $node ); |
| 253 | // The wikitext representation for them is dependent |
| 254 | // on cell position (first cell is always single char). |
| 255 | |
| 256 | // If there is no previous sibling, nothing to worry about. |
| 257 | $prev = $node->previousSibling; |
| 258 | if ( !$prev ) { |
| 259 | return true; |
| 260 | } |
| 261 | |
| 262 | if ( |
| 263 | DiffUtils::hasInsertedDiffMark( $prev ) || |
| 264 | DiffUtils::hasInsertedDiffMark( $node ) |
| 265 | ) { |
| 266 | return false; |
| 267 | } |
| 268 | |
| 269 | // If previous sibling is unmodified, nothing to worry about. |
| 270 | if ( |
| 271 | !DiffUtils::isDiffMarker( $prev ) && |
| 272 | !DiffUtils::directChildrenChanged( $prev ) |
| 273 | ) { |
| 274 | return true; |
| 275 | } |
| 276 | |
| 277 | // If it didn't have a stx marker that indicated that the cell |
| 278 | // showed up on the same line via the "||" or "!!" syntax, nothing |
| 279 | // to worry about. |
| 280 | return ( DOMDataUtils::getDataParsoid( $node )->stx ?? '' ) !== 'row'; |
| 281 | } elseif ( $node instanceof Element && DOMCompat::nodeName( $node ) === 'tr' && |
| 282 | empty( DOMDataUtils::getDataParsoid( $node )->startTagSrc ) |
| 283 | ) { |
| 284 | // If this <tr> didn't have a startTagSrc, it would have been |
| 285 | // the first row of a table in original wikitext. So, it is safe |
| 286 | // to reuse the original source for the row (without a "|-") as long as |
| 287 | // it continues to be the first row of the table. If not, since we need to |
| 288 | // insert a "|-" to separate it from the newly added row (in an edit), |
| 289 | // we cannot simply reuse orig. wikitext for this <tr>. |
| 290 | return !DiffDOMUtils::previousNonSepSibling( $node ); |
| 291 | } elseif ( DOMUtils::isNestedListOrListItem( $node ) ) { |
| 292 | if ( DOMUtils::isList( $node ) ) { |
| 293 | // Lists never get bullets assigned to them. So, unless they |
| 294 | // start a fresh list ( => they have a previous sibling ), |
| 295 | // we cannot reuse source for nested lists. |
| 296 | if ( !$node->previousSibling ) { |
| 297 | return false; |
| 298 | } |
| 299 | } else { |
| 300 | // Consider this wikitext snippet and its output below: |
| 301 | // |
| 302 | // ** a |
| 303 | // *** b |
| 304 | // |
| 305 | // <ul><li-*> |
| 306 | // <ul><li-*> a <-- cannot reuse source of this <li> |
| 307 | // <ul><li-***> b</li></ul> <-- can reuse source of this <li> |
| 308 | // </li></ul> |
| 309 | // </li></ul> |
| 310 | // |
| 311 | // If we reuse the src for the inner li with the a, we'd be missing |
| 312 | // one bullet because the tag handler for lists in the serializer only |
| 313 | // emits start tag src when it hits a first child that isn't a list |
| 314 | // element. We need to walk up and get the other bullet(s). |
| 315 | // |
| 316 | // The above logic can be condensed into this observation. |
| 317 | // Reusable nested <li> nodes will always have multiple bullets. |
| 318 | // Don't reuse source from any nested list |
| 319 | $dp = DOMDataUtils::getDataParsoid( $node ); |
| 320 | if ( !isset( $dp->dsr ) || $dp->dsr->openWidth < 2 ) { |
| 321 | return false; |
| 322 | } |
| 323 | } |
| 324 | |
| 325 | // If a previous sibling was modified, we can't reuse the start dsr. |
| 326 | $prev = $node->previousSibling; |
| 327 | while ( $prev ) { |
| 328 | if ( DiffUtils::isDiffMarker( $prev ) || DiffUtils::hasInsertedDiffMark( $prev ) ) { |
| 329 | return false; |
| 330 | } |
| 331 | $prev = $prev->previousSibling; |
| 332 | } |
| 333 | |
| 334 | return true; |
| 335 | } elseif ( WTUtils::isMovedMetaTag( $node ) ) { |
| 336 | return false; |
| 337 | } else { |
| 338 | return true; |
| 339 | } |
| 340 | } |
| 341 | |
| 342 | /** |
| 343 | * We keep track in $state of all extended ranges that are currently open by a <meta> tag. |
| 344 | * This method checks whether the wikitext source pointed by the dsr of the node contains either |
| 345 | * an opening or closing tag matching that annotation (<translate> or </translate> for example.) |
| 346 | * @param Node $node |
| 347 | * @param SerializerState $state |
| 348 | * @return bool |
| 349 | */ |
| 350 | private static function dsrContainsOpenExtendedRangeAnnotationTag( Node $node, |
| 351 | SerializerState $state |
| 352 | ): bool { |
| 353 | if ( empty( $state->openAnnotations ) || !$node instanceof Element ) { |
| 354 | return false; |
| 355 | } |
| 356 | |
| 357 | $dsr = DOMDataUtils::getDataParsoid( $node )->dsr ?? null; |
| 358 | if ( !$dsr ) { |
| 359 | return false; |
| 360 | } |
| 361 | $src = $state->getOrigSrc( $dsr->innerStart(), $dsr->innerEnd() ); |
| 362 | foreach ( $state->openAnnotations as $ann => $extended ) { |
| 363 | if ( $extended ) { |
| 364 | if ( preg_match( '</?' . $ann . '.*>', $src ) ) { |
| 365 | return true; |
| 366 | } |
| 367 | } |
| 368 | } |
| 369 | return false; |
| 370 | } |
| 371 | |
| 372 | /** |
| 373 | * FIXME: This method should probably be moved to DOMDataUtils class since |
| 374 | * it is used by both html2wt and wt2html code |
| 375 | * |
| 376 | * @param DataMw $dataMw |
| 377 | * @param string $key |
| 378 | * @param bool $keep |
| 379 | * @return array|null |
| 380 | */ |
| 381 | public static function getAttrFromDataMw( |
| 382 | DataMw $dataMw, string $key, bool $keep |
| 383 | ): ?array { |
| 384 | $arr = $dataMw->attribs ?? []; |
| 385 | $i = false; |
| 386 | foreach ( $arr as $k => $a ) { |
| 387 | if ( is_string( $a[0] ) ) { |
| 388 | $txt = $a[0]; |
| 389 | } elseif ( is_object( $a[0] ) ) { |
| 390 | $txt = $a[0]->txt ?? null; |
| 391 | } else { |
| 392 | throw new UnreachableException( 'Control should never get here!' ); |
| 393 | } |
| 394 | if ( $txt === $key ) { |
| 395 | $i = $k; |
| 396 | break; |
| 397 | } |
| 398 | } |
| 399 | if ( $i === false ) { |
| 400 | return null; |
| 401 | } |
| 402 | |
| 403 | $ret = $arr[$i]; |
| 404 | if ( !$keep && !isset( $ret[1]->html ) ) { |
| 405 | array_splice( $arr, $i, 1 ); |
| 406 | $dataMw->attribs = $arr; |
| 407 | } |
| 408 | return $ret; |
| 409 | } |
| 410 | |
| 411 | /** |
| 412 | * Escape `<nowiki>` tags. |
| 413 | * |
| 414 | * @param string $text |
| 415 | * @return string |
| 416 | */ |
| 417 | public static function escapeNowikiTags( string $text ): string { |
| 418 | return preg_replace( '#<(/?nowiki\s*/?\s*)>#i', '<$1>', $text ); |
| 419 | } |
| 420 | |
| 421 | } |