1<?php 2 3// A PHP diff engine for phpwiki. (Taken from phpwiki-1.3.3) 4// 5// Copyright (C) 2000, 2001 Geoffrey T. Dairiki <dairiki@dairiki.org> 6// You may copy this code freely under the conditions of the GPL. 7// 8 9define('USE_ASSERTS', function_exists('assert')); 10 11class _DiffOp { 12 var $type; 13 var $orig; 14 var $closing; 15 16 function reverse() { 17 trigger_error("pure virtual", E_USER_ERROR); 18 } 19 20 function norig() { 21 return $this->orig ? sizeof($this->orig) : 0; 22 } 23 24 function nclosing() { 25 return $this->closing ? sizeof($this->closing) : 0; 26 } 27} 28 29class _DiffOp_Copy extends _DiffOp { 30 var $type = 'copy'; 31 32 function _DiffOp_Copy ($orig, $closing = false) { 33 if (!is_array($closing)) 34 $closing = $orig; 35 $this->orig = $orig; 36 $this->closing = $closing; 37 } 38 39 function reverse() { 40 return new _DiffOp_Copy($this->closing, $this->orig); 41 } 42} 43 44class _DiffOp_Delete extends _DiffOp { 45 var $type = 'delete'; 46 47 function _DiffOp_Delete ($lines) { 48 $this->orig = $lines; 49 $this->closing = false; 50 } 51 52 function reverse() { 53 return new _DiffOp_Add($this->orig); 54 } 55} 56 57class _DiffOp_Add extends _DiffOp { 58 var $type = 'add'; 59 60 function _DiffOp_Add ($lines) { 61 $this->closing = $lines; 62 $this->orig = false; 63 } 64 65 function reverse() { 66 return new _DiffOp_Delete($this->closing); 67 } 68} 69 70class _DiffOp_Change extends _DiffOp { 71 var $type = 'change'; 72 73 function _DiffOp_Change ($orig, $closing) { 74 $this->orig = $orig; 75 $this->closing = $closing; 76 } 77 78 function reverse() { 79 return new _DiffOp_Change($this->closing, $this->orig); 80 } 81} 82 83 84/** 85 * Class used internally by Diff to actually compute the diffs. 86 * 87 * The algorithm used here is mostly lifted from the perl module 88 * Algorithm::Diff (version 1.06) by Ned Konz, which is available at: 89 * http://www.perl.com/CPAN/authors/id/N/NE/NEDKONZ/Algorithm-Diff-1.06.zip 90 * 91 * More ideas are taken from: 92 * http://www.ics.uci.edu/~eppstein/161/960229.html 93 * 94 * Some ideas are (and a bit of code) are from from analyze.c, from GNU 95 * diffutils-2.7, which can be found at: 96 * ftp://gnudist.gnu.org/pub/gnu/diffutils/diffutils-2.7.tar.gz 97 * 98 * closingly, some ideas (subdivision by NCHUNKS > 2, and some optimizations) 99 * are my own. 100 * 101 * @author Geoffrey T. Dairiki 102 * @access private 103 */ 104class _DiffEngine 105{ 106 function diff ($from_lines, $to_lines) { 107 $n_from = sizeof($from_lines); 108 $n_to = sizeof($to_lines); 109 110 $this->xchanged = $this->ychanged = array(); 111 $this->xv = $this->yv = array(); 112 $this->xind = $this->yind = array(); 113 unset($this->seq); 114 unset($this->in_seq); 115 unset($this->lcs); 116 117 // Skip leading common lines. 118 for ($skip = 0; $skip < $n_from && $skip < $n_to; $skip++) { 119 if ($from_lines[$skip] != $to_lines[$skip]) 120 break; 121 $this->xchanged[$skip] = $this->ychanged[$skip] = false; 122 } 123 // Skip trailing common lines. 124 $xi = $n_from; $yi = $n_to; 125 for ($endskip = 0; --$xi > $skip && --$yi > $skip; $endskip++) { 126 if ($from_lines[$xi] != $to_lines[$yi]) 127 break; 128 $this->xchanged[$xi] = $this->ychanged[$yi] = false; 129 } 130 131 // Ignore lines which do not exist in both files. 132 for ($xi = $skip; $xi < $n_from - $endskip; $xi++) 133 $xhash[$from_lines[$xi]] = 1; 134 for ($yi = $skip; $yi < $n_to - $endskip; $yi++) { 135 $line = $to_lines[$yi]; 136 if ( ($this->ychanged[$yi] = empty($xhash[$line])) ) 137 continue; 138 $yhash[$line] = 1; 139 $this->yv[] = $line; 140 $this->yind[] = $yi; 141 } 142 for ($xi = $skip; $xi < $n_from - $endskip; $xi++) { 143 $line = $from_lines[$xi]; 144 if ( ($this->xchanged[$xi] = empty($yhash[$line])) ) 145 continue; 146 $this->xv[] = $line; 147 $this->xind[] = $xi; 148 } 149 150 // Find the LCS. 151 $this->_compareseq(0, sizeof($this->xv), 0, sizeof($this->yv)); 152 153 // Merge edits when possible 154 $this->_shift_boundaries($from_lines, $this->xchanged, $this->ychanged); 155 $this->_shift_boundaries($to_lines, $this->ychanged, $this->xchanged); 156 157 // Compute the edit operations. 158 $edits = array(); 159 $xi = $yi = 0; 160 while ($xi < $n_from || $yi < $n_to) { 161 USE_ASSERTS && assert($yi < $n_to || $this->xchanged[$xi]); 162 USE_ASSERTS && assert($xi < $n_from || $this->ychanged[$yi]); 163 164 // Skip matching "snake". 165 $copy = array(); 166 while ( $xi < $n_from && $yi < $n_to 167 && !$this->xchanged[$xi] && !$this->ychanged[$yi]) { 168 $copy[] = $from_lines[$xi++]; 169 ++$yi; 170 } 171 if ($copy) 172 $edits[] = new _DiffOp_Copy($copy); 173 174 // Find deletes & adds. 175 $delete = array(); 176 while ($xi < $n_from && $this->xchanged[$xi]) 177 $delete[] = $from_lines[$xi++]; 178 179 $add = array(); 180 while ($yi < $n_to && $this->ychanged[$yi]) 181 $add[] = $to_lines[$yi++]; 182 183 if ($delete && $add) 184 $edits[] = new _DiffOp_Change($delete, $add); 185 elseif ($delete) 186 $edits[] = new _DiffOp_Delete($delete); 187 elseif ($add) 188 $edits[] = new _DiffOp_Add($add); 189 } 190 return $edits; 191 } 192 193 194 /* Divide the Largest Common Subsequence (LCS) of the sequences 195 * [XOFF, XLIM) and [YOFF, YLIM) into NCHUNKS approximately equally 196 * sized segments. 197 * 198 * Returns (LCS, PTS). LCS is the length of the LCS. PTS is an 199 * array of NCHUNKS+1 (X, Y) indexes giving the diving points between 200 * sub sequences. The first sub-sequence is contained in [X0, X1), 201 * [Y0, Y1), the second in [X1, X2), [Y1, Y2) and so on. Note 202 * that (X0, Y0) == (XOFF, YOFF) and 203 * (X[NCHUNKS], Y[NCHUNKS]) == (XLIM, YLIM). 204 * 205 * This function assumes that the first lines of the specified portions 206 * of the two files do not match, and likewise that the last lines do not 207 * match. The caller must trim matching lines from the beginning and end 208 * of the portions it is going to specify. 209 */ 210 function _diag ($xoff, $xlim, $yoff, $ylim, $nchunks) { 211 $flip = false; 212 213 if ($xlim - $xoff > $ylim - $yoff) { 214 // Things seems faster (I'm not sure I understand why) 215 // when the shortest sequence in X. 216 $flip = true; 217 list ($xoff, $xlim, $yoff, $ylim) 218 = array( $yoff, $ylim, $xoff, $xlim); 219 } 220 221 if ($flip) 222 for ($i = $ylim - 1; $i >= $yoff; $i--) 223 $ymatches[$this->xv[$i]][] = $i; 224 else 225 for ($i = $ylim - 1; $i >= $yoff; $i--) 226 $ymatches[$this->yv[$i]][] = $i; 227 228 $this->lcs = 0; 229 $this->seq[0]= $yoff - 1; 230 $this->in_seq = array(); 231 $ymids[0] = array(); 232 233 $numer = $xlim - $xoff + $nchunks - 1; 234 $x = $xoff; 235 for ($chunk = 0; $chunk < $nchunks; $chunk++) { 236 if ($chunk > 0) 237 for ($i = 0; $i <= $this->lcs; $i++) 238 $ymids[$i][$chunk-1] = $this->seq[$i]; 239 240 $x1 = $xoff + (int)(($numer + ($xlim-$xoff)*$chunk) / $nchunks); 241 for ( ; $x < $x1; $x++) { 242 $line = $flip ? $this->yv[$x] : $this->xv[$x]; 243 if (empty($ymatches[$line])) 244 continue; 245 $matches = $ymatches[$line]; 246 reset($matches); 247 while (list ($junk, $y) = each($matches)) 248 if (empty($this->in_seq[$y])) { 249 $k = $this->_lcs_pos($y); 250 USE_ASSERTS && assert($k > 0); 251 $ymids[$k] = $ymids[$k-1]; 252 break; 253 } 254 while (list ($junk, $y) = each($matches)) { 255 if ($y > $this->seq[$k-1]) { 256 USE_ASSERTS && assert($y < $this->seq[$k]); 257 // Optimization: this is a common case: 258 // next match is just replacing previous match. 259 $this->in_seq[$this->seq[$k]] = false; 260 $this->seq[$k] = $y; 261 $this->in_seq[$y] = 1; 262 } 263 else if (empty($this->in_seq[$y])) { 264 $k = $this->_lcs_pos($y); 265 USE_ASSERTS && assert($k > 0); 266 $ymids[$k] = $ymids[$k-1]; 267 } 268 } 269 } 270 } 271 272 $seps[] = $flip ? array($yoff, $xoff) : array($xoff, $yoff); 273 $ymid = $ymids[$this->lcs]; 274 for ($n = 0; $n < $nchunks - 1; $n++) { 275 $x1 = $xoff + (int)(($numer + ($xlim - $xoff) * $n) / $nchunks); 276 $y1 = $ymid[$n] + 1; 277 $seps[] = $flip ? array($y1, $x1) : array($x1, $y1); 278 } 279 $seps[] = $flip ? array($ylim, $xlim) : array($xlim, $ylim); 280 281 return array($this->lcs, $seps); 282 } 283 284 function _lcs_pos ($ypos) { 285 $end = $this->lcs; 286 if ($end == 0 || $ypos > $this->seq[$end]) { 287 $this->seq[++$this->lcs] = $ypos; 288 $this->in_seq[$ypos] = 1; 289 return $this->lcs; 290 } 291 292 $beg = 1; 293 while ($beg < $end) { 294 $mid = (int)(($beg + $end) / 2); 295 if ( $ypos > $this->seq[$mid] ) 296 $beg = $mid + 1; 297 else 298 $end = $mid; 299 } 300 301 USE_ASSERTS && assert($ypos != $this->seq[$end]); 302 303 $this->in_seq[$this->seq[$end]] = false; 304 $this->seq[$end] = $ypos; 305 $this->in_seq[$ypos] = 1; 306 return $end; 307 } 308 309 /* Find LCS of two sequences. 310 * 311 * The results are recorded in the vectors $this->{x,y}changed[], by 312 * storing a 1 in the element for each line that is an insertion 313 * or deletion (ie. is not in the LCS). 314 * 315 * The subsequence of file 0 is [XOFF, XLIM) and likewise for file 1. 316 * 317 * Note that XLIM, YLIM are exclusive bounds. 318 * All line numbers are origin-0 and discarded lines are not counted. 319 */ 320 function _compareseq ($xoff, $xlim, $yoff, $ylim) { 321 // Slide down the bottom initial diagonal. 322 while ($xoff < $xlim && $yoff < $ylim 323 && $this->xv[$xoff] == $this->yv[$yoff]) { 324 ++$xoff; 325 ++$yoff; 326 } 327 328 // Slide up the top initial diagonal. 329 while ($xlim > $xoff && $ylim > $yoff 330 && $this->xv[$xlim - 1] == $this->yv[$ylim - 1]) { 331 --$xlim; 332 --$ylim; 333 } 334 335 if ($xoff == $xlim || $yoff == $ylim) 336 $lcs = 0; 337 else { 338 // This is ad hoc but seems to work well. 339 //$nchunks = sqrt(min($xlim - $xoff, $ylim - $yoff) / 2.5); 340 //$nchunks = max(2,min(8,(int)$nchunks)); 341 $nchunks = min(7, $xlim - $xoff, $ylim - $yoff) + 1; 342 list ($lcs, $seps) 343 = $this->_diag($xoff,$xlim,$yoff, $ylim,$nchunks); 344 } 345 346 if ($lcs == 0) { 347 // X and Y sequences have no common subsequence: 348 // mark all changed. 349 while ($yoff < $ylim) 350 $this->ychanged[$this->yind[$yoff++]] = 1; 351 while ($xoff < $xlim) 352 $this->xchanged[$this->xind[$xoff++]] = 1; 353 } 354 else { 355 // Use the partitions to split this problem into subproblems. 356 reset($seps); 357 $pt1 = $seps[0]; 358 while ($pt2 = next($seps)) { 359 $this->_compareseq ($pt1[0], $pt2[0], $pt1[1], $pt2[1]); 360 $pt1 = $pt2; 361 } 362 } 363 } 364 365 /* Adjust inserts/deletes of identical lines to join changes 366 * as much as possible. 367 * 368 * We do something when a run of changed lines include a 369 * line at one end and has an excluded, identical line at the other. 370 * We are free to choose which identical line is included. 371 * `compareseq' usually chooses the one at the beginning, 372 * but usually it is cleaner to consider the following identical line 373 * to be the "change". 374 * 375 * This is extracted verbatim from analyze.c (GNU diffutils-2.7). 376 */ 377 function _shift_boundaries ($lines, &$changed, $other_changed) { 378 $i = 0; 379 $j = 0; 380 381 USE_ASSERTS && assert('sizeof($lines) == sizeof($changed)'); 382 $len = sizeof($lines); 383 $other_len = sizeof($other_changed); 384 385 while (1) { 386 /* 387 * Scan forwards to find beginning of another run of changes. 388 * Also keep track of the corresponding point in the other file. 389 * 390 * Throughout this code, $i and $j are adjusted together so that 391 * the first $i elements of $changed and the first $j elements 392 * of $other_changed both contain the same number of zeros 393 * (unchanged lines). 394 * Furthermore, $j is always kept so that $j == $other_len or 395 * $other_changed[$j] == false. 396 */ 397 while ($j < $other_len && $other_changed[$j]) 398 $j++; 399 400 while ($i < $len && ! $changed[$i]) { 401 USE_ASSERTS && assert('$j < $other_len && ! $other_changed[$j]'); 402 $i++; $j++; 403 while ($j < $other_len && $other_changed[$j]) 404 $j++; 405 } 406 407 if ($i == $len) 408 break; 409 410 $start = $i; 411 412 // Find the end of this run of changes. 413 while (++$i < $len && $changed[$i]) 414 continue; 415 416 do { 417 /* 418 * Record the length of this run of changes, so that 419 * we can later determine whether the run has grown. 420 */ 421 $runlength = $i - $start; 422 423 /* 424 * Move the changed region back, so long as the 425 * previous unchanged line matches the last changed one. 426 * This merges with previous changed regions. 427 */ 428 while ($start > 0 && $lines[$start - 1] == $lines[$i - 1]) { 429 $changed[--$start] = 1; 430 $changed[--$i] = false; 431 while ($start > 0 && $changed[$start - 1]) 432 $start--; 433 USE_ASSERTS && assert('$j > 0'); 434 while ($other_changed[--$j]) 435 continue; 436 USE_ASSERTS && assert('$j >= 0 && !$other_changed[$j]'); 437 } 438 439 /* 440 * Set CORRESPONDING to the end of the changed run, at the last 441 * point where it corresponds to a changed run in the other file. 442 * CORRESPONDING == LEN means no such point has been found. 443 */ 444 $corresponding = $j < $other_len ? $i : $len; 445 446 /* 447 * Move the changed region forward, so long as the 448 * first changed line matches the following unchanged one. 449 * This merges with following changed regions. 450 * Do this second, so that if there are no merges, 451 * the changed region is moved forward as far as possible. 452 */ 453 while ($i < $len && $lines[$start] == $lines[$i]) { 454 $changed[$start++] = false; 455 $changed[$i++] = 1; 456 while ($i < $len && $changed[$i]) 457 $i++; 458 459 USE_ASSERTS && assert('$j < $other_len && ! $other_changed[$j]'); 460 $j++; 461 if ($j < $other_len && $other_changed[$j]) { 462 $corresponding = $i; 463 while ($j < $other_len && $other_changed[$j]) 464 $j++; 465 } 466 } 467 } while ($runlength != $i - $start); 468 469 /* 470 * If possible, move the fully-merged run of changes 471 * back to a corresponding run in the other file. 472 */ 473 while ($corresponding < $i) { 474 $changed[--$start] = 1; 475 $changed[--$i] = 0; 476 USE_ASSERTS && assert('$j > 0'); 477 while ($other_changed[--$j]) 478 continue; 479 USE_ASSERTS && assert('$j >= 0 && !$other_changed[$j]'); 480 } 481 } 482 } 483} 484 485/** 486 * Class representing a 'diff' between two sequences of strings. 487 */ 488class Diff 489{ 490 var $edits; 491 492 /** 493 * Constructor. 494 * Computes diff between sequences of strings. 495 * 496 * @param $from_lines array An array of strings. 497 * (Typically these are lines from a file.) 498 * @param $to_lines array An array of strings. 499 */ 500 function Diff($from_lines, $to_lines) { 501 $eng = new _DiffEngine; 502 $this->edits = $eng->diff($from_lines, $to_lines); 503 //$this->_check($from_lines, $to_lines); 504 } 505 506 /** 507 * Compute reversed Diff. 508 * 509 * SYNOPSIS: 510 * 511 * $diff = new Diff($lines1, $lines2); 512 * $rev = $diff->reverse(); 513 * @return object A Diff object representing the inverse of the 514 * original diff. 515 */ 516 function reverse () { 517 $rev = $this; 518 $rev->edits = array(); 519 foreach ($this->edits as $edit) { 520 $rev->edits[] = $edit->reverse(); 521 } 522 return $rev; 523 } 524 525 /** 526 * Check for empty diff. 527 * 528 * @return bool True iff two sequences were identical. 529 */ 530 function isEmpty () { 531 foreach ($this->edits as $edit) { 532 if ($edit->type != 'copy') 533 return false; 534 } 535 return true; 536 } 537 538 /** 539 * Compute the length of the Longest Common Subsequence (LCS). 540 * 541 * This is mostly for diagnostic purposed. 542 * 543 * @return int The length of the LCS. 544 */ 545 function lcs () { 546 $lcs = 0; 547 foreach ($this->edits as $edit) { 548 if ($edit->type == 'copy') 549 $lcs += sizeof($edit->orig); 550 } 551 return $lcs; 552 } 553 554 /** 555 * Get the original set of lines. 556 * 557 * This reconstructs the $from_lines parameter passed to the 558 * constructor. 559 * 560 * @return array The original sequence of strings. 561 */ 562 function orig() { 563 $lines = array(); 564 565 foreach ($this->edits as $edit) { 566 if ($edit->orig) 567 array_splice($lines, sizeof($lines), 0, $edit->orig); 568 } 569 return $lines; 570 } 571 572 /** 573 * Get the closing set of lines. 574 * 575 * This reconstructs the $to_lines parameter passed to the 576 * constructor. 577 * 578 * @return array The sequence of strings. 579 */ 580 function closing() { 581 $lines = array(); 582 583 foreach ($this->edits as $edit) { 584 if ($edit->closing) 585 array_splice($lines, sizeof($lines), 0, $edit->closing); 586 } 587 return $lines; 588 } 589 590 /** 591 * Check a Diff for validity. 592 * 593 * This is here only for debugging purposes. 594 */ 595 function _check ($from_lines, $to_lines) { 596 if (serialize($from_lines) != serialize($this->orig())) 597 trigger_error("Reconstructed original doesn't match", E_USER_ERROR); 598 if (serialize($to_lines) != serialize($this->closing())) 599 trigger_error("Reconstructed closing doesn't match", E_USER_ERROR); 600 601 $rev = $this->reverse(); 602 if (serialize($to_lines) != serialize($rev->orig())) 603 trigger_error("Reversed original doesn't match", E_USER_ERROR); 604 if (serialize($from_lines) != serialize($rev->closing())) 605 trigger_error("Reversed closing doesn't match", E_USER_ERROR); 606 607 608 $prevtype = 'none'; 609 foreach ($this->edits as $edit) { 610 if ( $prevtype == $edit->type ) 611 trigger_error("Edit sequence is non-optimal", E_USER_ERROR); 612 $prevtype = $edit->type; 613 } 614 615 $lcs = $this->lcs(); 616 trigger_error("Diff okay: LCS = $lcs", E_USER_NOTICE); 617 } 618} 619 620/** 621 * FIXME: bad name. 622 */ 623class MappedDiff 624extends Diff 625{ 626 /** 627 * Constructor. 628 * 629 * Computes diff between sequences of strings. 630 * 631 * This can be used to compute things like 632 * case-insensitve diffs, or diffs which ignore 633 * changes in white-space. 634 * 635 * @param $from_lines array An array of strings. 636 * (Typically these are lines from a file.) 637 * 638 * @param $to_lines array An array of strings. 639 * 640 * @param $mapped_from_lines array This array should 641 * have the same size number of elements as $from_lines. 642 * The elements in $mapped_from_lines and 643 * $mapped_to_lines are what is actually compared 644 * when computing the diff. 645 * 646 * @param $mapped_to_lines array This array should 647 * have the same number of elements as $to_lines. 648 */ 649 function MappedDiff($from_lines, $to_lines, 650 $mapped_from_lines, $mapped_to_lines) { 651 652 assert(sizeof($from_lines) == sizeof($mapped_from_lines)); 653 assert(sizeof($to_lines) == sizeof($mapped_to_lines)); 654 655 $this->Diff($mapped_from_lines, $mapped_to_lines); 656 657 $xi = $yi = 0; 658 for ($i = 0; $i < sizeof($this->edits); $i++) { 659 $orig = &$this->edits[$i]->orig; 660 if (is_array($orig)) { 661 $orig = array_slice($from_lines, $xi, sizeof($orig)); 662 $xi += sizeof($orig); 663 } 664 665 $closing = &$this->edits[$i]->closing; 666 if (is_array($closing)) { 667 $closing = array_slice($to_lines, $yi, sizeof($closing)); 668 $yi += sizeof($closing); 669 } 670 } 671 } 672} 673 674/** 675 * A class to format Diffs 676 * 677 * This class formats the diff in classic diff format. 678 * It is intended that this class be customized via inheritance, 679 * to obtain fancier outputs. 680 */ 681class DiffFormatter 682{ 683 /** 684 * Number of leading context "lines" to preserve. 685 * 686 * This should be left at zero for this class, but subclasses 687 * may want to set this to other values. 688 */ 689 var $leading_context_lines = 0; 690 691 /** 692 * Number of trailing context "lines" to preserve. 693 * 694 * This should be left at zero for this class, but subclasses 695 * may want to set this to other values. 696 */ 697 var $trailing_context_lines = 0; 698 699 /** 700 * Format a diff. 701 * 702 * @param $diff object A Diff object. 703 * @return string The formatted output. 704 */ 705 function format($diff) { 706 707 $xi = $yi = 1; 708 $block = false; 709 $context = array(); 710 711 $nlead = $this->leading_context_lines; 712 $ntrail = $this->trailing_context_lines; 713 714 $this->_start_diff(); 715 716 foreach ($diff->edits as $edit) { 717 if ($edit->type == 'copy') { 718 if (is_array($block)) { 719 if (sizeof($edit->orig) <= $nlead + $ntrail) { 720 $block[] = $edit; 721 } 722 else{ 723 if ($ntrail) { 724 $context = array_slice($edit->orig, 0, $ntrail); 725 $block[] = new _DiffOp_Copy($context); 726 } 727 $this->_block($x0, $ntrail + $xi - $x0, 728 $y0, $ntrail + $yi - $y0, 729 $block); 730 $block = false; 731 } 732 } 733 $context = $edit->orig; 734 } 735 else { 736 if (! is_array($block)) { 737 $context = array_slice($context, sizeof($context) - $nlead); 738 $x0 = $xi - sizeof($context); 739 $y0 = $yi - sizeof($context); 740 $block = array(); 741 if ($context) 742 $block[] = new _DiffOp_Copy($context); 743 } 744 $block[] = $edit; 745 } 746 747 if ($edit->orig) 748 $xi += sizeof($edit->orig); 749 if ($edit->closing) 750 $yi += sizeof($edit->closing); 751 } 752 753 if (is_array($block)) 754 $this->_block($x0, $xi - $x0, 755 $y0, $yi - $y0, 756 $block); 757 758 return $this->_end_diff(); 759 } 760 761 function _block($xbeg, $xlen, $ybeg, $ylen, &$edits) { 762 $this->_start_block($this->_block_header($xbeg, $xlen, $ybeg, $ylen)); 763 foreach ($edits as $edit) { 764 if ($edit->type == 'copy') 765 $this->_context($edit->orig); 766 elseif ($edit->type == 'add') 767 $this->_added($edit->closing); 768 elseif ($edit->type == 'delete') 769 $this->_deleted($edit->orig); 770 elseif ($edit->type == 'change') 771 $this->_changed($edit->orig, $edit->closing); 772 else 773 trigger_error("Unknown edit type", E_USER_ERROR); 774 } 775 $this->_end_block(); 776 } 777 778 function _start_diff() { 779 ob_start(); 780 } 781 782 function _end_diff() { 783 $val = ob_get_contents(); 784 ob_end_clean(); 785 return $val; 786 } 787 788 function _block_header($xbeg, $xlen, $ybeg, $ylen) { 789 if ($xlen > 1) 790 $xbeg .= "," . ($xbeg + $xlen - 1); 791 if ($ylen > 1) 792 $ybeg .= "," . ($ybeg + $ylen - 1); 793 794 return $xbeg . ($xlen ? ($ylen ? 'c' : 'd') : 'a') . $ybeg; 795 } 796 797 function _start_block($header) { 798 echo $header; 799 } 800 801 function _end_block() { 802 } 803 804 function _lines($lines, $prefix = ' ') { 805 foreach ($lines as $line) 806 echo "$prefix $line\n"; 807 } 808 809 function _context($lines) { 810 $this->_lines($lines); 811 } 812 813 function _added($lines) { 814 $this->_lines($lines, ">"); 815 } 816 function _deleted($lines) { 817 $this->_lines($lines, "<"); 818 } 819 820 function _changed($orig, $closing) { 821 $this->_deleted($orig); 822 echo "---\n"; 823 $this->_added($closing); 824 } 825} 826 827 828/** 829 * Additions by Axel Boldt follow, partly taken from diff.php, phpwiki-1.3.3 830 * 831 */ 832 833define('NBSP', "\xA0"); // iso-8859-x non-breaking space. 834 835class _HWLDF_WordAccumulator { 836 function _HWLDF_WordAccumulator () { 837 $this->_lines = array(); 838 $this->_line = ''; 839 $this->_group = ''; 840 $this->_tag = ''; 841 } 842 843 function _flushGroup ($new_tag) { 844 if ($this->_group !== '') { 845 if ($this->_tag == 'mark') 846 $this->_line .= '<span class="diffchange">'.$this->_group.'</span>'; 847 else 848 $this->_line .= $this->_group; 849 } 850 $this->_group = ''; 851 $this->_tag = $new_tag; 852 } 853 854 function _flushLine ($new_tag) { 855 $this->_flushGroup($new_tag); 856 if ($this->_line != '') 857 $this->_lines[] = $this->_line; 858 $this->_line = ''; 859 } 860 861 function addWords ($words, $tag = '') { 862 if ($tag != $this->_tag) 863 $this->_flushGroup($tag); 864 865 foreach ($words as $word) { 866 // new-line should only come as first char of word. 867 if ($word == '') 868 continue; 869 if ($word[0] == "\n") { 870 $this->_group .= NBSP; 871 $this->_flushLine($tag); 872 $word = substr($word, 1); 873 } 874 assert(!strstr($word, "\n")); 875 $this->_group .= $word; 876 } 877 } 878 879 function getLines() { 880 $this->_flushLine('~done'); 881 return $this->_lines; 882 } 883} 884 885class WordLevelDiff extends MappedDiff 886{ 887 function WordLevelDiff ($orig_lines, $closing_lines) { 888 list ($orig_words, $orig_stripped) = $this->_split($orig_lines); 889 list ($closing_words, $closing_stripped) = $this->_split($closing_lines); 890 891 892 $this->MappedDiff($orig_words, $closing_words, 893 $orig_stripped, $closing_stripped); 894 } 895 896 function _split($lines) { 897 // FIXME: fix POSIX char class. 898# if (!preg_match_all('/ ( [^\S\n]+ | [[:alnum:]]+ | . ) (?: (?!< \n) [^\S\n])? /xs', 899 if (!preg_match_all('/ ( [^\S\n]+ | [0-9_A-Za-z\x80-\xff]+ | . ) (?: (?!< \n) [^\S\n])? /xs', 900 implode("\n", $lines), 901 $m)) { 902 return array(array(''), array('')); 903 } 904 return array($m[0], $m[1]); 905 } 906 907 function orig () { 908 $orig = new _HWLDF_WordAccumulator; 909 910 foreach ($this->edits as $edit) { 911 if ($edit->type == 'copy') 912 $orig->addWords($edit->orig); 913 elseif ($edit->orig) 914 $orig->addWords($edit->orig, 'mark'); 915 } 916 return $orig->getLines(); 917 } 918 919 function closing () { 920 $closing = new _HWLDF_WordAccumulator; 921 922 foreach ($this->edits as $edit) { 923 if ($edit->type == 'copy') 924 $closing->addWords($edit->closing); 925 elseif ($edit->closing) 926 $closing->addWords($edit->closing, 'mark'); 927 } 928 return $closing->getLines(); 929 } 930} 931 932/** 933 * "Unified" diff formatter. 934 * 935 * This class formats the diff in classic "unified diff" format. 936 */ 937class UnifiedDiffFormatter extends DiffFormatter 938{ 939 function UnifiedDiffFormatter($context_lines = 4) { 940 $this->leading_context_lines = $context_lines; 941 $this->trailing_context_lines = $context_lines; 942 } 943 944 function _block_header($xbeg, $xlen, $ybeg, $ylen) { 945 if ($xlen != 1) 946 $xbeg .= "," . $xlen; 947 if ($ylen != 1) 948 $ybeg .= "," . $ylen; 949 return "@@ -$xbeg +$ybeg @@\n"; 950 } 951 952 function _added($lines) { 953 $this->_lines($lines, "+"); 954 } 955 function _deleted($lines) { 956 $this->_lines($lines, "-"); 957 } 958 function _changed($orig, $final) { 959 $this->_deleted($orig); 960 $this->_added($final); 961 } 962} 963 964/** 965 * Wikipedia Table style diff formatter. 966 * 967 */ 968class TableDiffFormatter extends DiffFormatter 969{ 970 function TableDiffFormatter() { 971 $this->leading_context_lines = 2; 972 $this->trailing_context_lines = 2; 973 } 974 975 function _pre($text){ 976 $text = htmlspecialchars($text); 977 $text = str_replace(' ',' ',$text); 978 return $text; 979 } 980 981 function _block_header( $xbeg, $xlen, $ybeg, $ylen ) { 982 global $lang; 983 $l1 = $lang['line'].' '.$xbeg; 984 $l2 = $lang['line'].' '.$ybeg; 985 $r = '<tr><td class="diff-blockheader" colspan="2">'.$l1.":</td>\n" . 986 '<td class="diff-blockheader" colspan="2">'.$l2.":</td></tr>\n"; 987 return $r; 988 } 989 990 function _start_block( $header ) { 991 print( $header ); 992 } 993 994 function _end_block() { 995 } 996 997 function _lines( $lines, $prefix=' ', $color="white" ) { 998 } 999 1000 function addedLine( $line ) { 1001 $line = str_replace(' ',' ',$line); 1002 return '<td>+</td><td class="diff-addedline">' . 1003 $line.'</td>'; 1004 } 1005 1006 function deletedLine( $line ) { 1007 $line = str_replace(' ',' ',$line); 1008 return '<td>-</td><td class="diff-deletedline">' . 1009 $line.'</td>'; 1010 } 1011 1012 function emptyLine() { 1013 $line = str_replace(' ',' ',$line); 1014 return '<td colspan="2"> </td>'; 1015 } 1016 1017 function contextLine( $line ) { 1018 $line = str_replace(' ',' ',$line); 1019 return '<td> </td><td class="diff-context">'.$line.'</td>'; 1020 } 1021 1022 function _added($lines) { 1023 foreach ($lines as $line) { 1024 print( '<tr>' . $this->emptyLine() . 1025 $this->addedLine( $line ) . "</tr>\n" ); 1026 } 1027 } 1028 1029 function _deleted($lines) { 1030 foreach ($lines as $line) { 1031 print( '<tr>' . $this->deletedLine( $line ) . 1032 $this->emptyLine() . "</tr>\n" ); 1033 } 1034 } 1035 1036 function _context( $lines ) { 1037 foreach ($lines as $line) { 1038 print( '<tr>' . $this->contextLine( $line ) . 1039 $this->contextLine( $line ) . "</tr>\n" ); 1040 } 1041 } 1042 1043 function _changed( $orig, $closing ) { 1044 $diff = new WordLevelDiff( $orig, $closing ); 1045 $del = $diff->orig(); 1046 $add = $diff->closing(); 1047 1048 while ( $line = array_shift( $del ) ) { 1049 $aline = array_shift( $add ); 1050 print( '<tr>' . $this->deletedLine( $line ) . 1051 $this->addedLine( $aline ) . "</tr>\n" ); 1052 } 1053 $this->_added( $add ); # If any leftovers 1054 } 1055} 1056 1057?> 1058