1<?php 2/** 3 * A PHP diff engine for phpwiki. (Taken from phpwiki-1.3.3) 4 * 5 * Additions by Axel Boldt for MediaWiki 6 * 7 * @copyright (C) 2000, 2001 Geoffrey T. Dairiki <dairiki@dairiki.org> 8 * @license You may copy this code freely under the conditions of the GPL. 9 */ 10define('USE_ASSERTS', function_exists('assert')); 11 12class _DiffOp { 13 var $type; 14 var $orig; 15 var $closing; 16 17 /** 18 * @return _DiffOp 19 */ 20 function reverse() { 21 trigger_error("pure virtual", E_USER_ERROR); 22 } 23 24 function norig() { 25 return $this->orig ? count($this->orig) : 0; 26 } 27 28 function nclosing() { 29 return $this->closing ? count($this->closing) : 0; 30 } 31} 32 33class _DiffOp_Copy extends _DiffOp { 34 var $type = 'copy'; 35 36 function __construct($orig, $closing = false) { 37 if (!is_array($closing)) 38 $closing = $orig; 39 $this->orig = $orig; 40 $this->closing = $closing; 41 } 42 43 function reverse() { 44 return new _DiffOp_Copy($this->closing, $this->orig); 45 } 46} 47 48class _DiffOp_Delete extends _DiffOp { 49 var $type = 'delete'; 50 51 function __construct($lines) { 52 $this->orig = $lines; 53 $this->closing = false; 54 } 55 56 function reverse() { 57 return new _DiffOp_Add($this->orig); 58 } 59} 60 61class _DiffOp_Add extends _DiffOp { 62 var $type = 'add'; 63 64 function __construct($lines) { 65 $this->closing = $lines; 66 $this->orig = false; 67 } 68 69 function reverse() { 70 return new _DiffOp_Delete($this->closing); 71 } 72} 73 74class _DiffOp_Change extends _DiffOp { 75 var $type = 'change'; 76 77 function __construct($orig, $closing) { 78 $this->orig = $orig; 79 $this->closing = $closing; 80 } 81 82 function reverse() { 83 return new _DiffOp_Change($this->closing, $this->orig); 84 } 85} 86 87 88/** 89 * Class used internally by Diff to actually compute the diffs. 90 * 91 * The algorithm used here is mostly lifted from the perl module 92 * Algorithm::Diff (version 1.06) by Ned Konz, which is available at: 93 * http://www.perl.com/CPAN/authors/id/N/NE/NEDKONZ/Algorithm-Diff-1.06.zip 94 * 95 * More ideas are taken from: 96 * http://www.ics.uci.edu/~eppstein/161/960229.html 97 * 98 * Some ideas are (and a bit of code) are from from analyze.c, from GNU 99 * diffutils-2.7, which can be found at: 100 * ftp://gnudist.gnu.org/pub/gnu/diffutils/diffutils-2.7.tar.gz 101 * 102 * closingly, some ideas (subdivision by NCHUNKS > 2, and some optimizations) 103 * are my own. 104 * 105 * @author Geoffrey T. Dairiki 106 * @access private 107 */ 108class _DiffEngine { 109 110 var $xchanged = array(); 111 var $ychanged = array(); 112 var $xv = array(); 113 var $yv = array(); 114 var $xind = array(); 115 var $yind = array(); 116 var $seq; 117 var $in_seq; 118 var $lcs; 119 120 /** 121 * @param array $from_lines 122 * @param array $to_lines 123 * @return _DiffOp[] 124 */ 125 function diff($from_lines, $to_lines) { 126 $n_from = count($from_lines); 127 $n_to = count($to_lines); 128 129 $this->xchanged = $this->ychanged = array(); 130 $this->xv = $this->yv = array(); 131 $this->xind = $this->yind = array(); 132 unset($this->seq); 133 unset($this->in_seq); 134 unset($this->lcs); 135 136 // Skip leading common lines. 137 for ($skip = 0; $skip < $n_from && $skip < $n_to; $skip++) { 138 if ($from_lines[$skip] != $to_lines[$skip]) 139 break; 140 $this->xchanged[$skip] = $this->ychanged[$skip] = false; 141 } 142 // Skip trailing common lines. 143 $xi = $n_from; 144 $yi = $n_to; 145 for ($endskip = 0; --$xi > $skip && --$yi > $skip; $endskip++) { 146 if ($from_lines[$xi] != $to_lines[$yi]) 147 break; 148 $this->xchanged[$xi] = $this->ychanged[$yi] = false; 149 } 150 151 // Ignore lines which do not exist in both files. 152 for ($xi = $skip; $xi < $n_from - $endskip; $xi++) 153 $xhash[$from_lines[$xi]] = 1; 154 for ($yi = $skip; $yi < $n_to - $endskip; $yi++) { 155 $line = $to_lines[$yi]; 156 if (($this->ychanged[$yi] = empty($xhash[$line]))) 157 continue; 158 $yhash[$line] = 1; 159 $this->yv[] = $line; 160 $this->yind[] = $yi; 161 } 162 for ($xi = $skip; $xi < $n_from - $endskip; $xi++) { 163 $line = $from_lines[$xi]; 164 if (($this->xchanged[$xi] = empty($yhash[$line]))) 165 continue; 166 $this->xv[] = $line; 167 $this->xind[] = $xi; 168 } 169 170 // Find the LCS. 171 $this->_compareseq(0, count($this->xv), 0, count($this->yv)); 172 173 // Merge edits when possible 174 $this->_shift_boundaries($from_lines, $this->xchanged, $this->ychanged); 175 $this->_shift_boundaries($to_lines, $this->ychanged, $this->xchanged); 176 177 // Compute the edit operations. 178 $edits = array(); 179 $xi = $yi = 0; 180 while ($xi < $n_from || $yi < $n_to) { 181 USE_ASSERTS && assert($yi < $n_to || $this->xchanged[$xi]); 182 USE_ASSERTS && assert($xi < $n_from || $this->ychanged[$yi]); 183 184 // Skip matching "snake". 185 $copy = array(); 186 while ($xi < $n_from && $yi < $n_to && !$this->xchanged[$xi] && !$this->ychanged[$yi]) { 187 $copy[] = $from_lines[$xi++]; 188 ++$yi; 189 } 190 if ($copy) 191 $edits[] = new _DiffOp_Copy($copy); 192 193 // Find deletes & adds. 194 $delete = array(); 195 while ($xi < $n_from && $this->xchanged[$xi]) 196 $delete[] = $from_lines[$xi++]; 197 198 $add = array(); 199 while ($yi < $n_to && $this->ychanged[$yi]) 200 $add[] = $to_lines[$yi++]; 201 202 if ($delete && $add) 203 $edits[] = new _DiffOp_Change($delete, $add); 204 elseif ($delete) 205 $edits[] = new _DiffOp_Delete($delete); 206 elseif ($add) 207 $edits[] = new _DiffOp_Add($add); 208 } 209 return $edits; 210 } 211 212 213 /** 214 * Divide the Largest Common Subsequence (LCS) of the sequences 215 * [XOFF, XLIM) and [YOFF, YLIM) into NCHUNKS approximately equally 216 * sized segments. 217 * 218 * Returns (LCS, PTS). LCS is the length of the LCS. PTS is an 219 * array of NCHUNKS+1 (X, Y) indexes giving the diving points between 220 * sub sequences. The first sub-sequence is contained in [X0, X1), 221 * [Y0, Y1), the second in [X1, X2), [Y1, Y2) and so on. Note 222 * that (X0, Y0) == (XOFF, YOFF) and 223 * (X[NCHUNKS], Y[NCHUNKS]) == (XLIM, YLIM). 224 * 225 * This function assumes that the first lines of the specified portions 226 * of the two files do not match, and likewise that the last lines do not 227 * match. The caller must trim matching lines from the beginning and end 228 * of the portions it is going to specify. 229 */ 230 function _diag($xoff, $xlim, $yoff, $ylim, $nchunks) { 231 $flip = false; 232 233 if ($xlim - $xoff > $ylim - $yoff) { 234 // Things seems faster (I'm not sure I understand why) 235 // when the shortest sequence in X. 236 $flip = true; 237 list ($xoff, $xlim, $yoff, $ylim) = array($yoff, $ylim, $xoff, $xlim); 238 } 239 240 if ($flip) 241 for ($i = $ylim - 1; $i >= $yoff; $i--) 242 $ymatches[$this->xv[$i]][] = $i; 243 else 244 for ($i = $ylim - 1; $i >= $yoff; $i--) 245 $ymatches[$this->yv[$i]][] = $i; 246 247 $this->lcs = 0; 248 $this->seq[0]= $yoff - 1; 249 $this->in_seq = array(); 250 $ymids[0] = array(); 251 252 $numer = $xlim - $xoff + $nchunks - 1; 253 $x = $xoff; 254 for ($chunk = 0; $chunk < $nchunks; $chunk++) { 255 if ($chunk > 0) 256 for ($i = 0; $i <= $this->lcs; $i++) 257 $ymids[$i][$chunk-1] = $this->seq[$i]; 258 259 $x1 = $xoff + (int)(($numer + ($xlim-$xoff)*$chunk) / $nchunks); 260 for ( ; $x < $x1; $x++) { 261 $line = $flip ? $this->yv[$x] : $this->xv[$x]; 262 if (empty($ymatches[$line])) 263 continue; 264 $matches = $ymatches[$line]; 265 reset($matches); 266 while (list ($junk, $y) = each($matches)) 267 if (empty($this->in_seq[$y])) { 268 $k = $this->_lcs_pos($y); 269 USE_ASSERTS && assert($k > 0); 270 $ymids[$k] = $ymids[$k-1]; 271 break; 272 } 273 while (list ($junk, $y) = each($matches)) { 274 if ($y > $this->seq[$k-1]) { 275 USE_ASSERTS && assert($y < $this->seq[$k]); 276 // Optimization: this is a common case: 277 // next match is just replacing previous match. 278 $this->in_seq[$this->seq[$k]] = false; 279 $this->seq[$k] = $y; 280 $this->in_seq[$y] = 1; 281 } 282 else if (empty($this->in_seq[$y])) { 283 $k = $this->_lcs_pos($y); 284 USE_ASSERTS && assert($k > 0); 285 $ymids[$k] = $ymids[$k-1]; 286 } 287 } 288 } 289 } 290 291 $seps[] = $flip ? array($yoff, $xoff) : array($xoff, $yoff); 292 $ymid = $ymids[$this->lcs]; 293 for ($n = 0; $n < $nchunks - 1; $n++) { 294 $x1 = $xoff + (int)(($numer + ($xlim - $xoff) * $n) / $nchunks); 295 $y1 = $ymid[$n] + 1; 296 $seps[] = $flip ? array($y1, $x1) : array($x1, $y1); 297 } 298 $seps[] = $flip ? array($ylim, $xlim) : array($xlim, $ylim); 299 300 return array($this->lcs, $seps); 301 } 302 303 function _lcs_pos($ypos) { 304 $end = $this->lcs; 305 if ($end == 0 || $ypos > $this->seq[$end]) { 306 $this->seq[++$this->lcs] = $ypos; 307 $this->in_seq[$ypos] = 1; 308 return $this->lcs; 309 } 310 311 $beg = 1; 312 while ($beg < $end) { 313 $mid = (int)(($beg + $end) / 2); 314 if ($ypos > $this->seq[$mid]) 315 $beg = $mid + 1; 316 else 317 $end = $mid; 318 } 319 320 USE_ASSERTS && assert($ypos != $this->seq[$end]); 321 322 $this->in_seq[$this->seq[$end]] = false; 323 $this->seq[$end] = $ypos; 324 $this->in_seq[$ypos] = 1; 325 return $end; 326 } 327 328 /** 329 * Find LCS of two sequences. 330 * 331 * The results are recorded in the vectors $this->{x,y}changed[], by 332 * storing a 1 in the element for each line that is an insertion 333 * or deletion (ie. is not in the LCS). 334 * 335 * The subsequence of file 0 is [XOFF, XLIM) and likewise for file 1. 336 * 337 * Note that XLIM, YLIM are exclusive bounds. 338 * All line numbers are origin-0 and discarded lines are not counted. 339 */ 340 function _compareseq($xoff, $xlim, $yoff, $ylim) { 341 // Slide down the bottom initial diagonal. 342 while ($xoff < $xlim && $yoff < $ylim && $this->xv[$xoff] == $this->yv[$yoff]) { 343 ++$xoff; 344 ++$yoff; 345 } 346 347 // Slide up the top initial diagonal. 348 while ($xlim > $xoff && $ylim > $yoff && $this->xv[$xlim - 1] == $this->yv[$ylim - 1]) { 349 --$xlim; 350 --$ylim; 351 } 352 353 if ($xoff == $xlim || $yoff == $ylim) 354 $lcs = 0; 355 else { 356 // This is ad hoc but seems to work well. 357 //$nchunks = sqrt(min($xlim - $xoff, $ylim - $yoff) / 2.5); 358 //$nchunks = max(2,min(8,(int)$nchunks)); 359 $nchunks = min(7, $xlim - $xoff, $ylim - $yoff) + 1; 360 list ($lcs, $seps) 361 = $this->_diag($xoff,$xlim,$yoff, $ylim,$nchunks); 362 } 363 364 if ($lcs == 0) { 365 // X and Y sequences have no common subsequence: 366 // mark all changed. 367 while ($yoff < $ylim) 368 $this->ychanged[$this->yind[$yoff++]] = 1; 369 while ($xoff < $xlim) 370 $this->xchanged[$this->xind[$xoff++]] = 1; 371 } 372 else { 373 // Use the partitions to split this problem into subproblems. 374 reset($seps); 375 $pt1 = $seps[0]; 376 while ($pt2 = next($seps)) { 377 $this->_compareseq ($pt1[0], $pt2[0], $pt1[1], $pt2[1]); 378 $pt1 = $pt2; 379 } 380 } 381 } 382 383 /** 384 * Adjust inserts/deletes of identical lines to join changes 385 * as much as possible. 386 * 387 * We do something when a run of changed lines include a 388 * line at one end and has an excluded, identical line at the other. 389 * We are free to choose which identical line is included. 390 * `compareseq' usually chooses the one at the beginning, 391 * but usually it is cleaner to consider the following identical line 392 * to be the "change". 393 * 394 * This is extracted verbatim from analyze.c (GNU diffutils-2.7). 395 */ 396 function _shift_boundaries($lines, &$changed, $other_changed) { 397 $i = 0; 398 $j = 0; 399 400 USE_ASSERTS && assert('count($lines) == count($changed)'); 401 $len = count($lines); 402 $other_len = count($other_changed); 403 404 while (1) { 405 /* 406 * Scan forwards to find beginning of another run of changes. 407 * Also keep track of the corresponding point in the other file. 408 * 409 * Throughout this code, $i and $j are adjusted together so that 410 * the first $i elements of $changed and the first $j elements 411 * of $other_changed both contain the same number of zeros 412 * (unchanged lines). 413 * Furthermore, $j is always kept so that $j == $other_len or 414 * $other_changed[$j] == false. 415 */ 416 while ($j < $other_len && $other_changed[$j]) 417 $j++; 418 419 while ($i < $len && ! $changed[$i]) { 420 USE_ASSERTS && assert('$j < $other_len && ! $other_changed[$j]'); 421 $i++; 422 $j++; 423 while ($j < $other_len && $other_changed[$j]) 424 $j++; 425 } 426 427 if ($i == $len) 428 break; 429 430 $start = $i; 431 432 // Find the end of this run of changes. 433 while (++$i < $len && $changed[$i]) 434 continue; 435 436 do { 437 /* 438 * Record the length of this run of changes, so that 439 * we can later determine whether the run has grown. 440 */ 441 $runlength = $i - $start; 442 443 /* 444 * Move the changed region back, so long as the 445 * previous unchanged line matches the last changed one. 446 * This merges with previous changed regions. 447 */ 448 while ($start > 0 && $lines[$start - 1] == $lines[$i - 1]) { 449 $changed[--$start] = 1; 450 $changed[--$i] = false; 451 while ($start > 0 && $changed[$start - 1]) 452 $start--; 453 USE_ASSERTS && assert('$j > 0'); 454 while ($other_changed[--$j]) 455 continue; 456 USE_ASSERTS && assert('$j >= 0 && !$other_changed[$j]'); 457 } 458 459 /* 460 * Set CORRESPONDING to the end of the changed run, at the last 461 * point where it corresponds to a changed run in the other file. 462 * CORRESPONDING == LEN means no such point has been found. 463 */ 464 $corresponding = $j < $other_len ? $i : $len; 465 466 /* 467 * Move the changed region forward, so long as the 468 * first changed line matches the following unchanged one. 469 * This merges with following changed regions. 470 * Do this second, so that if there are no merges, 471 * the changed region is moved forward as far as possible. 472 */ 473 while ($i < $len && $lines[$start] == $lines[$i]) { 474 $changed[$start++] = false; 475 $changed[$i++] = 1; 476 while ($i < $len && $changed[$i]) 477 $i++; 478 479 USE_ASSERTS && assert('$j < $other_len && ! $other_changed[$j]'); 480 $j++; 481 if ($j < $other_len && $other_changed[$j]) { 482 $corresponding = $i; 483 while ($j < $other_len && $other_changed[$j]) 484 $j++; 485 } 486 } 487 } while ($runlength != $i - $start); 488 489 /* 490 * If possible, move the fully-merged run of changes 491 * back to a corresponding run in the other file. 492 */ 493 while ($corresponding < $i) { 494 $changed[--$start] = 1; 495 $changed[--$i] = 0; 496 USE_ASSERTS && assert('$j > 0'); 497 while ($other_changed[--$j]) 498 continue; 499 USE_ASSERTS && assert('$j >= 0 && !$other_changed[$j]'); 500 } 501 } 502 } 503} 504 505/** 506 * Class representing a 'diff' between two sequences of strings. 507 */ 508class Diff { 509 510 var $edits; 511 512 /** 513 * Constructor. 514 * Computes diff between sequences of strings. 515 * 516 * @param array $from_lines An array of strings. 517 * (Typically these are lines from a file.) 518 * @param array $to_lines An array of strings. 519 */ 520 function __construct($from_lines, $to_lines) { 521 $eng = new _DiffEngine; 522 $this->edits = $eng->diff($from_lines, $to_lines); 523 //$this->_check($from_lines, $to_lines); 524 } 525 526 /** 527 * Compute reversed Diff. 528 * 529 * SYNOPSIS: 530 * 531 * $diff = new Diff($lines1, $lines2); 532 * $rev = $diff->reverse(); 533 * 534 * @return Diff A Diff object representing the inverse of the 535 * original diff. 536 */ 537 function reverse() { 538 $rev = $this; 539 $rev->edits = array(); 540 foreach ($this->edits as $edit) { 541 $rev->edits[] = $edit->reverse(); 542 } 543 return $rev; 544 } 545 546 /** 547 * Check for empty diff. 548 * 549 * @return bool True iff two sequences were identical. 550 */ 551 function isEmpty() { 552 foreach ($this->edits as $edit) { 553 if ($edit->type != 'copy') 554 return false; 555 } 556 return true; 557 } 558 559 /** 560 * Compute the length of the Longest Common Subsequence (LCS). 561 * 562 * This is mostly for diagnostic purposed. 563 * 564 * @return int The length of the LCS. 565 */ 566 function lcs() { 567 $lcs = 0; 568 foreach ($this->edits as $edit) { 569 if ($edit->type == 'copy') 570 $lcs += count($edit->orig); 571 } 572 return $lcs; 573 } 574 575 /** 576 * Get the original set of lines. 577 * 578 * This reconstructs the $from_lines parameter passed to the 579 * constructor. 580 * 581 * @return array The original sequence of strings. 582 */ 583 function orig() { 584 $lines = array(); 585 586 foreach ($this->edits as $edit) { 587 if ($edit->orig) 588 array_splice($lines, count($lines), 0, $edit->orig); 589 } 590 return $lines; 591 } 592 593 /** 594 * Get the closing set of lines. 595 * 596 * This reconstructs the $to_lines parameter passed to the 597 * constructor. 598 * 599 * @return array The sequence of strings. 600 */ 601 function closing() { 602 $lines = array(); 603 604 foreach ($this->edits as $edit) { 605 if ($edit->closing) 606 array_splice($lines, count($lines), 0, $edit->closing); 607 } 608 return $lines; 609 } 610 611 /** 612 * Check a Diff for validity. 613 * 614 * This is here only for debugging purposes. 615 */ 616 function _check($from_lines, $to_lines) { 617 if (serialize($from_lines) != serialize($this->orig())) 618 trigger_error("Reconstructed original doesn't match", E_USER_ERROR); 619 if (serialize($to_lines) != serialize($this->closing())) 620 trigger_error("Reconstructed closing doesn't match", E_USER_ERROR); 621 622 $rev = $this->reverse(); 623 if (serialize($to_lines) != serialize($rev->orig())) 624 trigger_error("Reversed original doesn't match", E_USER_ERROR); 625 if (serialize($from_lines) != serialize($rev->closing())) 626 trigger_error("Reversed closing doesn't match", E_USER_ERROR); 627 628 $prevtype = 'none'; 629 foreach ($this->edits as $edit) { 630 if ($prevtype == $edit->type) 631 trigger_error("Edit sequence is non-optimal", E_USER_ERROR); 632 $prevtype = $edit->type; 633 } 634 635 $lcs = $this->lcs(); 636 trigger_error("Diff okay: LCS = $lcs", E_USER_NOTICE); 637 } 638} 639 640/** 641 * FIXME: bad name. 642 */ 643class MappedDiff extends Diff { 644 /** 645 * Constructor. 646 * 647 * Computes diff between sequences of strings. 648 * 649 * This can be used to compute things like 650 * case-insensitve diffs, or diffs which ignore 651 * changes in white-space. 652 * 653 * @param string[] $from_lines An array of strings. 654 * (Typically these are lines from a file.) 655 * 656 * @param string[] $to_lines An array of strings. 657 * 658 * @param string[] $mapped_from_lines This array should 659 * have the same size number of elements as $from_lines. 660 * The elements in $mapped_from_lines and 661 * $mapped_to_lines are what is actually compared 662 * when computing the diff. 663 * 664 * @param string[] $mapped_to_lines This array should 665 * have the same number of elements as $to_lines. 666 */ 667 function __construct($from_lines, $to_lines, $mapped_from_lines, $mapped_to_lines) { 668 669 assert(count($from_lines) == count($mapped_from_lines)); 670 assert(count($to_lines) == count($mapped_to_lines)); 671 672 parent::__construct($mapped_from_lines, $mapped_to_lines); 673 674 $xi = $yi = 0; 675 $ecnt = count($this->edits); 676 for ($i = 0; $i < $ecnt; $i++) { 677 $orig = &$this->edits[$i]->orig; 678 if (is_array($orig)) { 679 $orig = array_slice($from_lines, $xi, count($orig)); 680 $xi += count($orig); 681 } 682 683 $closing = &$this->edits[$i]->closing; 684 if (is_array($closing)) { 685 $closing = array_slice($to_lines, $yi, count($closing)); 686 $yi += count($closing); 687 } 688 } 689 } 690} 691 692/** 693 * A class to format Diffs 694 * 695 * This class formats the diff in classic diff format. 696 * It is intended that this class be customized via inheritance, 697 * to obtain fancier outputs. 698 */ 699class DiffFormatter { 700 /** 701 * Number of leading context "lines" to preserve. 702 * 703 * This should be left at zero for this class, but subclasses 704 * may want to set this to other values. 705 */ 706 var $leading_context_lines = 0; 707 708 /** 709 * Number of trailing context "lines" to preserve. 710 * 711 * This should be left at zero for this class, but subclasses 712 * may want to set this to other values. 713 */ 714 var $trailing_context_lines = 0; 715 716 /** 717 * Format a diff. 718 * 719 * @param Diff $diff A Diff object. 720 * @return string The formatted output. 721 */ 722 function format($diff) { 723 724 $xi = $yi = 1; 725 $x0 = $y0 = 0; 726 $block = false; 727 $context = array(); 728 729 $nlead = $this->leading_context_lines; 730 $ntrail = $this->trailing_context_lines; 731 732 $this->_start_diff(); 733 734 foreach ($diff->edits as $edit) { 735 if ($edit->type == 'copy') { 736 if (is_array($block)) { 737 if (count($edit->orig) <= $nlead + $ntrail) { 738 $block[] = $edit; 739 } 740 else{ 741 if ($ntrail) { 742 $context = array_slice($edit->orig, 0, $ntrail); 743 $block[] = new _DiffOp_Copy($context); 744 } 745 $this->_block($x0, $ntrail + $xi - $x0, $y0, $ntrail + $yi - $y0, $block); 746 $block = false; 747 } 748 } 749 $context = $edit->orig; 750 } 751 else { 752 if (! is_array($block)) { 753 $context = array_slice($context, count($context) - $nlead); 754 $x0 = $xi - count($context); 755 $y0 = $yi - count($context); 756 $block = array(); 757 if ($context) 758 $block[] = new _DiffOp_Copy($context); 759 } 760 $block[] = $edit; 761 } 762 763 if ($edit->orig) 764 $xi += count($edit->orig); 765 if ($edit->closing) 766 $yi += count($edit->closing); 767 } 768 769 if (is_array($block)) 770 $this->_block($x0, $xi - $x0, $y0, $yi - $y0, $block); 771 772 return $this->_end_diff(); 773 } 774 775 /** 776 * @param int $xbeg 777 * @param int $xlen 778 * @param int $ybeg 779 * @param int $ylen 780 * @param array $edits 781 */ 782 function _block($xbeg, $xlen, $ybeg, $ylen, &$edits) { 783 $this->_start_block($this->_block_header($xbeg, $xlen, $ybeg, $ylen)); 784 foreach ($edits as $edit) { 785 if ($edit->type == 'copy') 786 $this->_context($edit->orig); 787 elseif ($edit->type == 'add') 788 $this->_added($edit->closing); 789 elseif ($edit->type == 'delete') 790 $this->_deleted($edit->orig); 791 elseif ($edit->type == 'change') 792 $this->_changed($edit->orig, $edit->closing); 793 else 794 trigger_error("Unknown edit type", E_USER_ERROR); 795 } 796 $this->_end_block(); 797 } 798 799 function _start_diff() { 800 ob_start(); 801 } 802 803 function _end_diff() { 804 $val = ob_get_contents(); 805 ob_end_clean(); 806 return $val; 807 } 808 809 /** 810 * @param int $xbeg 811 * @param int $xlen 812 * @param int $ybeg 813 * @param int $ylen 814 * @return string 815 */ 816 function _block_header($xbeg, $xlen, $ybeg, $ylen) { 817 if ($xlen > 1) 818 $xbeg .= "," . ($xbeg + $xlen - 1); 819 if ($ylen > 1) 820 $ybeg .= "," . ($ybeg + $ylen - 1); 821 822 return $xbeg . ($xlen ? ($ylen ? 'c' : 'd') : 'a') . $ybeg; 823 } 824 825 /** 826 * @param string $header 827 */ 828 function _start_block($header) { 829 echo $header; 830 } 831 832 function _end_block() { 833 } 834 835 function _lines($lines, $prefix = ' ') { 836 foreach ($lines as $line) 837 echo "$prefix ".$this->_escape($line)."\n"; 838 } 839 840 function _context($lines) { 841 $this->_lines($lines); 842 } 843 844 function _added($lines) { 845 $this->_lines($lines, ">"); 846 } 847 function _deleted($lines) { 848 $this->_lines($lines, "<"); 849 } 850 851 function _changed($orig, $closing) { 852 $this->_deleted($orig); 853 echo "---\n"; 854 $this->_added($closing); 855 } 856 857 /** 858 * Escape string 859 * 860 * Override this method within other formatters if escaping required. 861 * Base class requires $str to be returned WITHOUT escaping. 862 * 863 * @param $str string Text string to escape 864 * @return string The escaped string. 865 */ 866 function _escape($str){ 867 return $str; 868 } 869} 870 871/** 872 * Utilityclass for styling HTML formatted diffs 873 * 874 * Depends on global var $DIFF_INLINESTYLES, if true some minimal predefined 875 * inline styles are used. Useful for HTML mails and RSS feeds 876 * 877 * @author Andreas Gohr <andi@splitbrain.org> 878 */ 879class HTMLDiff { 880 /** 881 * Holds the style names and basic CSS 882 */ 883 static public $styles = array( 884 'diff-addedline' => 'background-color: #ddffdd;', 885 'diff-deletedline' => 'background-color: #ffdddd;', 886 'diff-context' => 'background-color: #f5f5f5;', 887 'diff-mark' => 'color: #ff0000;', 888 ); 889 890 /** 891 * Return a class or style parameter 892 */ 893 static function css($classname){ 894 global $DIFF_INLINESTYLES; 895 896 if($DIFF_INLINESTYLES){ 897 if(!isset(self::$styles[$classname])) return ''; 898 return 'style="'.self::$styles[$classname].'"'; 899 }else{ 900 return 'class="'.$classname.'"'; 901 } 902 } 903} 904 905/** 906 * Additions by Axel Boldt follow, partly taken from diff.php, phpwiki-1.3.3 907 * 908 */ 909 910define('NBSP', "\xC2\xA0"); // utf-8 non-breaking space. 911 912class _HWLDF_WordAccumulator { 913 914 function __construct() { 915 $this->_lines = array(); 916 $this->_line = ''; 917 $this->_group = ''; 918 $this->_tag = ''; 919 } 920 921 function _flushGroup($new_tag) { 922 if ($this->_group !== '') { 923 if ($this->_tag == 'mark') 924 $this->_line .= '<strong '.HTMLDiff::css('diff-mark').'>'.$this->_escape($this->_group).'</strong>'; 925 elseif ($this->_tag == 'add') 926 $this->_line .= '<span '.HTMLDiff::css('diff-addedline').'>'.$this->_escape($this->_group).'</span>'; 927 elseif ($this->_tag == 'del') 928 $this->_line .= '<span '.HTMLDiff::css('diff-deletedline').'><del>'.$this->_escape($this->_group).'</del></span>'; 929 else 930 $this->_line .= $this->_escape($this->_group); 931 } 932 $this->_group = ''; 933 $this->_tag = $new_tag; 934 } 935 936 /** 937 * @param string $new_tag 938 */ 939 function _flushLine($new_tag) { 940 $this->_flushGroup($new_tag); 941 if ($this->_line != '') 942 $this->_lines[] = $this->_line; 943 $this->_line = ''; 944 } 945 946 function addWords($words, $tag = '') { 947 if ($tag != $this->_tag) 948 $this->_flushGroup($tag); 949 950 foreach ($words as $word) { 951 // new-line should only come as first char of word. 952 if ($word == '') 953 continue; 954 if ($word[0] == "\n") { 955 $this->_group .= NBSP; 956 $this->_flushLine($tag); 957 $word = substr($word, 1); 958 } 959 assert(!strstr($word, "\n")); 960 $this->_group .= $word; 961 } 962 } 963 964 function getLines() { 965 $this->_flushLine('~done'); 966 return $this->_lines; 967 } 968 969 function _escape($str){ 970 return hsc($str); 971 } 972} 973 974class WordLevelDiff extends MappedDiff { 975 976 function __construct($orig_lines, $closing_lines) { 977 list ($orig_words, $orig_stripped) = $this->_split($orig_lines); 978 list ($closing_words, $closing_stripped) = $this->_split($closing_lines); 979 980 parent::__construct($orig_words, $closing_words, $orig_stripped, $closing_stripped); 981 } 982 983 function _split($lines) { 984 if (!preg_match_all('/ ( [^\S\n]+ | [0-9_A-Za-z\x80-\xff]+ | . ) (?: (?!< \n) [^\S\n])? /xsu', 985 implode("\n", $lines), $m)) { 986 return array(array(''), array('')); 987 } 988 return array($m[0], $m[1]); 989 } 990 991 function orig() { 992 $orig = new _HWLDF_WordAccumulator; 993 994 foreach ($this->edits as $edit) { 995 if ($edit->type == 'copy') 996 $orig->addWords($edit->orig); 997 elseif ($edit->orig) 998 $orig->addWords($edit->orig, 'mark'); 999 } 1000 return $orig->getLines(); 1001 } 1002 1003 function closing() { 1004 $closing = new _HWLDF_WordAccumulator; 1005 1006 foreach ($this->edits as $edit) { 1007 if ($edit->type == 'copy') 1008 $closing->addWords($edit->closing); 1009 elseif ($edit->closing) 1010 $closing->addWords($edit->closing, 'mark'); 1011 } 1012 return $closing->getLines(); 1013 } 1014} 1015 1016class InlineWordLevelDiff extends MappedDiff { 1017 1018 function __construct($orig_lines, $closing_lines) { 1019 list ($orig_words, $orig_stripped) = $this->_split($orig_lines); 1020 list ($closing_words, $closing_stripped) = $this->_split($closing_lines); 1021 1022 parent::__construct($orig_words, $closing_words, $orig_stripped, $closing_stripped); 1023 } 1024 1025 function _split($lines) { 1026 if (!preg_match_all('/ ( [^\S\n]+ | [0-9_A-Za-z\x80-\xff]+ | . ) (?: (?!< \n) [^\S\n])? /xsu', 1027 implode("\n", $lines), $m)) { 1028 return array(array(''), array('')); 1029 } 1030 return array($m[0], $m[1]); 1031 } 1032 1033 function inline() { 1034 $orig = new _HWLDF_WordAccumulator; 1035 foreach ($this->edits as $edit) { 1036 if ($edit->type == 'copy') 1037 $orig->addWords($edit->closing); 1038 elseif ($edit->type == 'change'){ 1039 $orig->addWords($edit->orig, 'del'); 1040 $orig->addWords($edit->closing, 'add'); 1041 } elseif ($edit->type == 'delete') 1042 $orig->addWords($edit->orig, 'del'); 1043 elseif ($edit->type == 'add') 1044 $orig->addWords($edit->closing, 'add'); 1045 elseif ($edit->orig) 1046 $orig->addWords($edit->orig, 'del'); 1047 } 1048 return $orig->getLines(); 1049 } 1050} 1051 1052/** 1053 * "Unified" diff formatter. 1054 * 1055 * This class formats the diff in classic "unified diff" format. 1056 * 1057 * NOTE: output is plain text and unsafe for use in HTML without escaping. 1058 */ 1059class UnifiedDiffFormatter extends DiffFormatter { 1060 1061 function __construct($context_lines = 4) { 1062 $this->leading_context_lines = $context_lines; 1063 $this->trailing_context_lines = $context_lines; 1064 } 1065 1066 function _block_header($xbeg, $xlen, $ybeg, $ylen) { 1067 if ($xlen != 1) 1068 $xbeg .= "," . $xlen; 1069 if ($ylen != 1) 1070 $ybeg .= "," . $ylen; 1071 return "@@ -$xbeg +$ybeg @@\n"; 1072 } 1073 1074 function _added($lines) { 1075 $this->_lines($lines, "+"); 1076 } 1077 function _deleted($lines) { 1078 $this->_lines($lines, "-"); 1079 } 1080 function _changed($orig, $final) { 1081 $this->_deleted($orig); 1082 $this->_added($final); 1083 } 1084} 1085 1086/** 1087 * Wikipedia Table style diff formatter. 1088 * 1089 */ 1090class TableDiffFormatter extends DiffFormatter { 1091 var $colspan = 2; 1092 1093 function __construct() { 1094 $this->leading_context_lines = 2; 1095 $this->trailing_context_lines = 2; 1096 } 1097 1098 /** 1099 * @param Diff $diff 1100 * @return string 1101 */ 1102 function format($diff) { 1103 // Preserve whitespaces by converting some to non-breaking spaces. 1104 // Do not convert all of them to allow word-wrap. 1105 $val = parent::format($diff); 1106 $val = str_replace(' ','  ', $val); 1107 $val = preg_replace('/ (?=<)|(?<=[ >]) /', ' ', $val); 1108 return $val; 1109 } 1110 1111 function _pre($text){ 1112 $text = htmlspecialchars($text); 1113 return $text; 1114 } 1115 1116 function _block_header($xbeg, $xlen, $ybeg, $ylen) { 1117 global $lang; 1118 $l1 = $lang['line'].' '.$xbeg; 1119 $l2 = $lang['line'].' '.$ybeg; 1120 $r = '<tr><td '.HTMLDiff::css('diff-blockheader').' colspan="'.$this->colspan.'">'.$l1.":</td>\n". 1121 '<td '.HTMLDiff::css('diff-blockheader').' colspan="'.$this->colspan.'">'.$l2.":</td>\n". 1122 "</tr>\n"; 1123 return $r; 1124 } 1125 1126 function _start_block($header) { 1127 print($header); 1128 } 1129 1130 function _end_block() { 1131 } 1132 1133 function _lines($lines, $prefix=' ', $color="white") { 1134 } 1135 1136 function addedLine($line,$escaped=false) { 1137 if (!$escaped){ 1138 $line = $this->_escape($line); 1139 } 1140 return '<td '.HTMLDiff::css('diff-lineheader').'>+</td>'. 1141 '<td '.HTMLDiff::css('diff-addedline').'>' . $line.'</td>'; 1142 } 1143 1144 function deletedLine($line,$escaped=false) { 1145 if (!$escaped){ 1146 $line = $this->_escape($line); 1147 } 1148 return '<td '.HTMLDiff::css('diff-lineheader').'>-</td>'. 1149 '<td '.HTMLDiff::css('diff-deletedline').'>' . $line.'</td>'; 1150 } 1151 1152 function emptyLine() { 1153 return '<td colspan="'.$this->colspan.'"> </td>'; 1154 } 1155 1156 function contextLine($line) { 1157 return '<td '.HTMLDiff::css('diff-lineheader').'> </td>'. 1158 '<td '.HTMLDiff::css('diff-context').'>'.$this->_escape($line).'</td>'; 1159 } 1160 1161 function _added($lines) { 1162 $this->_addedLines($lines,false); 1163 } 1164 1165 function _addedLines($lines,$escaped=false){ 1166 foreach ($lines as $line) { 1167 print('<tr>' . $this->emptyLine() . $this->addedLine($line,$escaped) . "</tr>\n"); 1168 } 1169 } 1170 1171 function _deleted($lines) { 1172 foreach ($lines as $line) { 1173 print('<tr>' . $this->deletedLine($line) . $this->emptyLine() . "</tr>\n"); 1174 } 1175 } 1176 1177 function _context($lines) { 1178 foreach ($lines as $line) { 1179 print('<tr>' . $this->contextLine($line) . $this->contextLine($line) . "</tr>\n"); 1180 } 1181 } 1182 1183 function _changed($orig, $closing) { 1184 $diff = new WordLevelDiff($orig, $closing); // this escapes the diff data 1185 $del = $diff->orig(); 1186 $add = $diff->closing(); 1187 1188 while ($line = array_shift($del)) { 1189 $aline = array_shift($add); 1190 print('<tr>' . $this->deletedLine($line,true) . $this->addedLine($aline,true) . "</tr>\n"); 1191 } 1192 $this->_addedLines($add,true); # If any leftovers 1193 } 1194 1195 function _escape($str) { 1196 return hsc($str); 1197 } 1198} 1199 1200/** 1201 * Inline style diff formatter. 1202 * 1203 */ 1204class InlineDiffFormatter extends DiffFormatter { 1205 var $colspan = 2; 1206 1207 function __construct() { 1208 $this->leading_context_lines = 2; 1209 $this->trailing_context_lines = 2; 1210 } 1211 1212 /** 1213 * @param Diff $diff 1214 * @return string 1215 */ 1216 function format($diff) { 1217 // Preserve whitespaces by converting some to non-breaking spaces. 1218 // Do not convert all of them to allow word-wrap. 1219 $val = parent::format($diff); 1220 $val = str_replace(' ','  ', $val); 1221 $val = preg_replace('/ (?=<)|(?<=[ >]) /', ' ', $val); 1222 return $val; 1223 } 1224 1225 function _pre($text){ 1226 $text = htmlspecialchars($text); 1227 return $text; 1228 } 1229 1230 function _block_header($xbeg, $xlen, $ybeg, $ylen) { 1231 global $lang; 1232 if ($xlen != 1) 1233 $xbeg .= "," . $xlen; 1234 if ($ylen != 1) 1235 $ybeg .= "," . $ylen; 1236 $r = '<tr><td colspan="'.$this->colspan.'" '.HTMLDiff::css('diff-blockheader').'>@@ '.$lang['line']." -$xbeg +$ybeg @@"; 1237 $r .= ' <span '.HTMLDiff::css('diff-deletedline').'><del>'.$lang['deleted'].'</del></span>'; 1238 $r .= ' <span '.HTMLDiff::css('diff-addedline').'>'.$lang['created'].'</span>'; 1239 $r .= "</td></tr>\n"; 1240 return $r; 1241 } 1242 1243 function _start_block($header) { 1244 print($header."\n"); 1245 } 1246 1247 function _end_block() { 1248 } 1249 1250 function _lines($lines, $prefix=' ', $color="white") { 1251 } 1252 1253 function _added($lines) { 1254 foreach ($lines as $line) { 1255 print('<tr><td '.HTMLDiff::css('diff-lineheader').'> </td><td '.HTMLDiff::css('diff-addedline').'>'. $this->_escape($line) . "</td></tr>\n"); 1256 } 1257 } 1258 1259 function _deleted($lines) { 1260 foreach ($lines as $line) { 1261 print('<tr><td '.HTMLDiff::css('diff-lineheader').'> </td><td '.HTMLDiff::css('diff-deletedline').'><del>' . $this->_escape($line) . "</del></td></tr>\n"); 1262 } 1263 } 1264 1265 function _context($lines) { 1266 foreach ($lines as $line) { 1267 print('<tr><td '.HTMLDiff::css('diff-lineheader').'> </td><td '.HTMLDiff::css('diff-context').'>'. $this->_escape($line) ."</td></tr>\n"); 1268 } 1269 } 1270 1271 function _changed($orig, $closing) { 1272 $diff = new InlineWordLevelDiff($orig, $closing); // this escapes the diff data 1273 $add = $diff->inline(); 1274 1275 foreach ($add as $line) 1276 print('<tr><td '.HTMLDiff::css('diff-lineheader').'> </td><td>'.$line."</td></tr>\n"); 1277 } 1278 1279 function _escape($str) { 1280 return hsc($str); 1281 } 1282} 1283 1284 1285//Setup VIM: ex: et ts=4 : 1286