caxy /
php-htmldiff
Duplicate code is one of the most pungent code smells. A rule that is often used is to re-structure code once it is duplicated in three or more places.
Common duplication problems, and corresponding solutions are:
Complex classes like TableDiff often do a lot of different things. To break such a class down, we need to identify a cohesive component within that class. A common approach to find such a component is to look for fields/methods that share the same prefixes, or suffixes. You can also have a look at the cohesion graph to spot any un-connected, or weakly-connected components.
Once you have determined the fields that belong together, you can apply the Extract Class refactoring. If the component makes sense as a sub-class, Extract Subclass is also a candidate, and is often faster.
While breaking up the class, it is a good idea to analyze how other classes use TableDiff, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 13 | class TableDiff extends AbstractDiff |
||
| 14 | { |
||
| 15 | /** |
||
| 16 | * @var null|Table |
||
| 17 | */ |
||
| 18 | protected $oldTable = null; |
||
| 19 | |||
| 20 | /** |
||
| 21 | * @var null|Table |
||
| 22 | */ |
||
| 23 | protected $newTable = null; |
||
| 24 | |||
| 25 | /** |
||
| 26 | * @var null|\DOMElement |
||
| 27 | */ |
||
| 28 | protected $diffTable = null; |
||
| 29 | |||
| 30 | /** |
||
| 31 | * @var null|\DOMDocument |
||
| 32 | */ |
||
| 33 | protected $diffDom = null; |
||
| 34 | |||
| 35 | /** |
||
| 36 | * @var int |
||
| 37 | */ |
||
| 38 | protected $newRowOffsets = 0; |
||
| 39 | |||
| 40 | /** |
||
| 41 | * @var int |
||
| 42 | */ |
||
| 43 | protected $oldRowOffsets = 0; |
||
| 44 | |||
| 45 | /** |
||
| 46 | * @var array |
||
| 47 | */ |
||
| 48 | protected $cellValues = array(); |
||
| 49 | |||
| 50 | /** |
||
| 51 | * @var \HTMLPurifier |
||
| 52 | */ |
||
| 53 | protected $purifier; |
||
| 54 | |||
| 55 | /** |
||
| 56 | * TableDiff constructor. |
||
| 57 | * |
||
| 58 | * @param string $oldText |
||
| 59 | * @param string $newText |
||
| 60 | * @param string $encoding |
||
| 61 | * @param array|null $specialCaseTags |
||
| 62 | * @param bool|null $groupDiffs |
||
| 63 | */ |
||
| 64 | public function __construct($oldText, $newText, $encoding = 'UTF-8', $specialCaseTags = null, $groupDiffs = null) |
||
| 70 | |||
| 71 | /** |
||
| 72 | * @return string |
||
| 73 | */ |
||
| 74 | public function build() |
||
| 86 | |||
| 87 | protected function diffTableContent() |
||
| 123 | |||
| 124 | /** |
||
| 125 | * @param TableRow[] $oldRows |
||
| 126 | * @param TableRow[] $newRows |
||
| 127 | * @param RowMatch[] $matches |
||
| 128 | */ |
||
| 129 | protected function diffTableRowsWithMatches($oldRows, $newRows, $matches) |
||
| 201 | |||
| 202 | /** |
||
| 203 | * @param Operation $operation |
||
| 204 | * @param array $newRows |
||
| 205 | * @param array $appliedRowSpans |
||
| 206 | * @param bool $forceExpansion |
||
| 207 | */ |
||
| 208 | View Code Duplication | protected function processInsertOperation( |
|
| 219 | |||
| 220 | /** |
||
| 221 | * @param Operation $operation |
||
| 222 | * @param array $oldRows |
||
| 223 | * @param array $appliedRowSpans |
||
| 224 | * @param bool $forceExpansion |
||
| 225 | */ |
||
| 226 | View Code Duplication | protected function processDeleteOperation( |
|
| 237 | |||
| 238 | /** |
||
| 239 | * @param Operation $operation |
||
| 240 | * @param array $oldRows |
||
| 241 | * @param array $newRows |
||
| 242 | * @param array $appliedRowSpans |
||
| 243 | */ |
||
| 244 | protected function processEqualOperation(Operation $operation, $oldRows, $newRows, &$appliedRowSpans) |
||
| 261 | |||
| 262 | /** |
||
| 263 | * @param Operation $operation |
||
| 264 | * @param array $oldRows |
||
| 265 | * @param array $newRows |
||
| 266 | * @param array $appliedRowSpans |
||
| 267 | */ |
||
| 268 | protected function processReplaceOperation(Operation $operation, $oldRows, $newRows, &$appliedRowSpans) |
||
| 273 | |||
| 274 | /** |
||
| 275 | * @param array $oldMatchData |
||
| 276 | * @param array $newMatchData |
||
| 277 | * |
||
| 278 | * @return array |
||
| 279 | */ |
||
| 280 | protected function getRowMatches($oldMatchData, $newMatchData) |
||
| 293 | |||
| 294 | /** |
||
| 295 | * @param array $newMatchData |
||
| 296 | * @param int $startInOld |
||
| 297 | * @param int $endInOld |
||
| 298 | * @param int $startInNew |
||
| 299 | * @param int $endInNew |
||
| 300 | * @param array $matches |
||
| 301 | */ |
||
| 302 | protected function findRowMatches($newMatchData, $startInOld, $endInOld, $startInNew, $endInNew, &$matches) |
||
| 335 | |||
| 336 | /** |
||
| 337 | * @param array $newMatchData |
||
| 338 | * @param int $startInOld |
||
| 339 | * @param int $endInOld |
||
| 340 | * @param int $startInNew |
||
| 341 | * @param int $endInNew |
||
| 342 | * |
||
| 343 | * @return RowMatch|null |
||
| 344 | */ |
||
| 345 | protected function findRowMatch($newMatchData, $startInOld, $endInOld, $startInNew, $endInNew) |
||
| 390 | |||
| 391 | /** |
||
| 392 | * @param TableRow|null $oldRow |
||
| 393 | * @param TableRow|null $newRow |
||
| 394 | * @param array $appliedRowSpans |
||
| 395 | * @param bool $forceExpansion |
||
| 396 | * |
||
| 397 | * @return array |
||
| 398 | */ |
||
| 399 | protected function diffRows($oldRow, $newRow, array &$appliedRowSpans, $forceExpansion = false) |
||
| 400 | { |
||
| 401 | // create tr dom element |
||
| 402 | $rowToClone = $newRow ?: $oldRow; |
||
| 403 | /* @var $diffRow \DOMElement */ |
||
| 404 | $diffRow = $this->diffDom->importNode($rowToClone->getDomNode()->cloneNode(false), false); |
||
| 405 | |||
| 406 | $oldCells = $oldRow ? $oldRow->getCells() : array(); |
||
| 407 | $newCells = $newRow ? $newRow->getCells() : array(); |
||
| 408 | |||
| 409 | $position = new DiffRowPosition(); |
||
| 410 | |||
| 411 | $extraRow = null; |
||
| 412 | |||
| 413 | /* @var $expandCells \DOMElement[] */ |
||
| 414 | $expandCells = array(); |
||
| 415 | /* @var $cellsWithMultipleRows \DOMElement[] */ |
||
| 416 | $cellsWithMultipleRows = array(); |
||
| 417 | |||
| 418 | $newCellCount = count($newCells); |
||
| 419 | while ($position->getIndexInNew() < $newCellCount) { |
||
| 420 | if (!$position->areColumnsEqual()) { |
||
| 421 | $type = $position->getLesserColumnType(); |
||
| 422 | if ($type === 'new') { |
||
| 423 | $row = $newRow; |
||
| 424 | $targetRow = $extraRow; |
||
| 425 | } else { |
||
| 426 | $row = $oldRow; |
||
| 427 | $targetRow = $diffRow; |
||
| 428 | } |
||
| 429 | if ($row && $targetRow && (!$type === 'old' || isset($oldCells[$position->getIndexInOld()]))) { |
||
| 430 | $this->syncVirtualColumns($row, $position, $cellsWithMultipleRows, $targetRow, $type, true); |
||
| 431 | |||
| 432 | continue; |
||
| 433 | } |
||
| 434 | } |
||
| 435 | |||
| 436 | /* @var $newCell TableCell */ |
||
| 437 | $newCell = $newCells[$position->getIndexInNew()]; |
||
| 438 | /* @var $oldCell TableCell */ |
||
| 439 | $oldCell = isset($oldCells[$position->getIndexInOld()]) ? $oldCells[$position->getIndexInOld()] : null; |
||
| 440 | |||
| 441 | if ($oldCell && $newCell->getColspan() != $oldCell->getColspan()) { |
||
| 442 | if (null === $extraRow) { |
||
| 443 | /* @var $extraRow \DOMElement */ |
||
| 444 | $extraRow = $this->diffDom->importNode($rowToClone->getDomNode()->cloneNode(false), false); |
||
| 445 | } |
||
| 446 | |||
| 447 | if ($oldCell->getColspan() > $newCell->getColspan()) { |
||
| 448 | $this->diffCellsAndIncrementCounters( |
||
| 449 | $oldCell, |
||
| 450 | null, |
||
| 451 | $cellsWithMultipleRows, |
||
| 452 | $diffRow, |
||
| 453 | $position, |
||
| 454 | true |
||
| 455 | ); |
||
| 456 | $this->syncVirtualColumns($newRow, $position, $cellsWithMultipleRows, $extraRow, 'new', true); |
||
| 457 | } else { |
||
| 458 | $this->diffCellsAndIncrementCounters( |
||
| 459 | null, |
||
| 460 | $newCell, |
||
| 461 | $cellsWithMultipleRows, |
||
| 462 | $extraRow, |
||
| 463 | $position, |
||
| 464 | true |
||
| 465 | ); |
||
| 466 | $this->syncVirtualColumns($oldRow, $position, $cellsWithMultipleRows, $diffRow, 'old', true); |
||
| 467 | } |
||
| 468 | } else { |
||
| 469 | $diffCell = $this->diffCellsAndIncrementCounters( |
||
| 470 | $oldCell, |
||
| 471 | $newCell, |
||
| 472 | $cellsWithMultipleRows, |
||
| 473 | $diffRow, |
||
| 474 | $position |
||
| 475 | ); |
||
| 476 | $expandCells[] = $diffCell; |
||
| 477 | } |
||
| 478 | } |
||
| 479 | |||
| 480 | $oldCellCount = count($oldCells); |
||
| 481 | while ($position->getIndexInOld() < $oldCellCount) { |
||
| 482 | $diffCell = $this->diffCellsAndIncrementCounters( |
||
| 483 | $oldCells[$position->getIndexInOld()], |
||
| 484 | null, |
||
| 485 | $cellsWithMultipleRows, |
||
| 486 | $diffRow, |
||
| 487 | $position |
||
| 488 | ); |
||
| 489 | $expandCells[] = $diffCell; |
||
| 490 | } |
||
| 491 | |||
| 492 | if ($extraRow) { |
||
| 493 | foreach ($expandCells as $expandCell) { |
||
| 494 | $rowspan = $expandCell->getAttribute('rowspan') ?: 1; |
||
| 495 | $expandCell->setAttribute('rowspan', 1 + $rowspan); |
||
| 496 | } |
||
| 497 | } |
||
| 498 | |||
| 499 | if ($extraRow || $forceExpansion) { |
||
| 500 | foreach ($appliedRowSpans as $rowSpanCells) { |
||
| 501 | /* @var $rowSpanCells \DOMElement[] */ |
||
| 502 | foreach ($rowSpanCells as $extendCell) { |
||
| 503 | $rowspan = $extendCell->getAttribute('rowspan') ?: 1; |
||
| 504 | $extendCell->setAttribute('rowspan', 1 + $rowspan); |
||
| 505 | } |
||
| 506 | } |
||
| 507 | } |
||
| 508 | |||
| 509 | if (!$forceExpansion) { |
||
| 510 | array_shift($appliedRowSpans); |
||
| 511 | $appliedRowSpans = array_values($appliedRowSpans); |
||
| 512 | } |
||
| 513 | $appliedRowSpans = array_merge($appliedRowSpans, array_values($cellsWithMultipleRows)); |
||
| 514 | |||
| 515 | return array($diffRow, $extraRow); |
||
| 516 | } |
||
| 517 | |||
| 518 | /** |
||
| 519 | * @param TableCell|null $oldCell |
||
| 520 | * @param TableCell|null $newCell |
||
| 521 | * |
||
| 522 | * @return \DOMElement |
||
| 523 | */ |
||
| 524 | protected function getNewCellNode(TableCell $oldCell = null, TableCell $newCell = null) |
||
| 549 | |||
| 550 | /** |
||
| 551 | * @param TableCell|null $oldCell |
||
| 552 | * @param TableCell|null $newCell |
||
| 553 | * @param bool $usingExtraRow |
||
| 554 | * |
||
| 555 | * @return \DOMElement |
||
| 556 | */ |
||
| 557 | protected function diffCells($oldCell, $newCell, $usingExtraRow = false) |
||
| 590 | |||
| 591 | protected function buildTableDoms() |
||
| 596 | |||
| 597 | /** |
||
| 598 | * @param string $text |
||
| 599 | * |
||
| 600 | * @return \DOMDocument |
||
| 601 | */ |
||
| 602 | protected function createDocumentWithHtml($text) |
||
| 613 | |||
| 614 | /** |
||
| 615 | * @param string $text |
||
| 616 | * |
||
| 617 | * @return Table |
||
| 618 | */ |
||
| 619 | protected function parseTableStructure($text) |
||
| 631 | |||
| 632 | /** |
||
| 633 | * @param Table $table |
||
| 634 | * @param \DOMNode|null $node |
||
| 635 | */ |
||
| 636 | protected function parseTable(Table $table, \DOMNode $node = null) |
||
| 653 | |||
| 654 | /** |
||
| 655 | * @param TableRow $row |
||
| 656 | */ |
||
| 657 | protected function parseTableRow(TableRow $row) |
||
| 668 | |||
| 669 | /** |
||
| 670 | * @param \DOMNode $node |
||
| 671 | * |
||
| 672 | * @return string |
||
| 673 | */ |
||
| 674 | protected function getInnerHtml($node) |
||
| 685 | |||
| 686 | /** |
||
| 687 | * @param \DOMNode $node |
||
| 688 | * |
||
| 689 | * @return string |
||
| 690 | */ |
||
| 691 | View Code Duplication | protected function htmlFromNode($node) |
|
| 699 | |||
| 700 | /** |
||
| 701 | * @param \DOMNode $node |
||
| 702 | * @param string $html |
||
| 703 | */ |
||
| 704 | protected function setInnerHtml($node, $html) |
||
| 720 | |||
| 721 | /** |
||
| 722 | * @param Table $table |
||
| 723 | */ |
||
| 724 | protected function indexCellValues(Table $table) |
||
| 738 | |||
| 739 | /** |
||
| 740 | * @param TableRow $tableRow |
||
| 741 | * @param DiffRowPosition $position |
||
| 742 | * @param array $cellsWithMultipleRows |
||
| 743 | * @param \DOMNode $diffRow |
||
| 744 | * @param string $diffType |
||
| 745 | * @param bool $usingExtraRow |
||
| 746 | */ |
||
| 747 | protected function syncVirtualColumns( |
||
| 771 | |||
| 772 | /** |
||
| 773 | * @param null|TableCell $oldCell |
||
| 774 | * @param null|TableCell $newCell |
||
| 775 | * @param array $cellsWithMultipleRows |
||
| 776 | * @param \DOMElement $diffRow |
||
| 777 | * @param DiffRowPosition $position |
||
| 778 | * @param bool $usingExtraRow |
||
| 779 | * |
||
| 780 | * @return \DOMElement |
||
| 781 | */ |
||
| 782 | protected function diffCellsAndIncrementCounters( |
||
| 809 | |||
| 810 | /** |
||
| 811 | * @param TableRow|null $oldRow |
||
| 812 | * @param TableRow|null $newRow |
||
| 813 | * @param array $appliedRowSpans |
||
| 814 | * @param bool $forceExpansion |
||
| 815 | */ |
||
| 816 | protected function diffAndAppendRows($oldRow, $newRow, &$appliedRowSpans, $forceExpansion = false) |
||
| 831 | |||
| 832 | /** |
||
| 833 | * @param TableRow $oldRow |
||
| 834 | * @param TableRow $newRow |
||
| 835 | * @param int $oldIndex |
||
| 836 | * @param int $newIndex |
||
| 837 | * |
||
| 838 | * @return float|int |
||
| 839 | */ |
||
| 840 | protected function getMatchPercentage(TableRow $oldRow, TableRow $newRow, $oldIndex, $newIndex) |
||
| 866 | } |
||
| 867 |
Loading history...
Unless you are absolutely sure that the expression can never be null because of other conditions, we strongly recommend to add an additional type check to your code: