[Summary view] [Print] [Text view]

   1  <?php
   2  /**
   3  *
   4  * This file is part of the phpBB Forum Software package.
   5  *
   6  * @copyright (c) phpBB Limited <https://www.phpbb.com>
   7  * @license GNU General Public License, version 2 (GPL-2.0)
   8  *
   9  * For full copyright and license information, please see
  10  * the docs/CREDITS.txt file.
  11  *
  12  */
  13  
  14  /**
  15  * @ignore
  16  */
  17  if (!defined('IN_PHPBB'))
  18  {
  19      exit;
  20  }
  21  
  22  /**
  23  * Code from pear.php.net, Text_Diff-1.1.0 package
  24  * http://pear.php.net/package/Text_Diff/ (native engine)
  25  *
  26  * Modified by phpBB Limited to meet our coding standards
  27  * and being able to integrate into phpBB
  28  *
  29  * Class used internally by Text_Diff to actually compute the diffs. This
  30  * class is implemented using native PHP code.
  31  *
  32  * The algorithm used here is mostly lifted from the perl module
  33  * Algorithm::Diff (version 1.06) by Ned Konz, which is available at:
  34  * http://www.perl.com/CPAN/authors/id/N/NE/NEDKONZ/Algorithm-Diff-1.06.zip
  35  *
  36  * More ideas are taken from: http://www.ics.uci.edu/~eppstein/161/960229.html
  37  *
  38  * Some ideas (and a bit of code) are taken from analyze.c, of GNU
  39  * diffutils-2.7, which can be found at:
  40  * ftp://gnudist.gnu.org/pub/gnu/diffutils/diffutils-2.7.tar.gz
  41  *
  42  * Some ideas (subdivision by NCHUNKS > 2, and some optimizations) are from
  43  * Geoffrey T. Dairiki <dairiki@dairiki.org>. The original PHP version of this
  44  * code was written by him, and is used/adapted with his permission.
  45  *
  46  * Copyright 2004-2008 The Horde Project (http://www.horde.org/)
  47  *
  48  * @author  Geoffrey T. Dairiki <dairiki@dairiki.org>
  49  * @package diff
  50  *
  51  * @access private
  52  */
  53  class diff_engine
  54  {
  55      /**
  56      * If set to true we trim all lines before we compare them. This ensures that sole space/tab changes do not trigger diffs.
  57      */
  58      var $skip_whitespace_changes = true;
  59  
  60  	function diff(&$from_lines, &$to_lines, $preserve_cr = true)
  61      {
  62          // Remove empty lines...
  63          // If preserve_cr is true, we basically only change \r\n and bare \r to \n to get the same carriage returns for both files
  64          // If it is false, we try to only use \n once per line and ommit all empty lines to be able to get a proper data diff
  65  
  66          if (is_array($from_lines))
  67          {
  68              $from_lines = implode("\n", $from_lines);
  69          }
  70  
  71          if (is_array($to_lines))
  72          {
  73              $to_lines = implode("\n", $to_lines);
  74          }
  75  
  76          if ($preserve_cr)
  77          {
  78              $from_lines = explode("\n", str_replace("\r", "\n", str_replace("\r\n", "\n", $from_lines)));
  79              $to_lines = explode("\n", str_replace("\r", "\n", str_replace("\r\n", "\n", $to_lines)));
  80          }
  81          else
  82          {
  83              $from_lines = explode("\n", preg_replace('#[\n\r]+#', "\n", $from_lines));
  84              $to_lines = explode("\n", preg_replace('#[\n\r]+#', "\n", $to_lines));
  85          }
  86  
  87          $n_from = sizeof($from_lines);
  88          $n_to = sizeof($to_lines);
  89  
  90          $this->xchanged = $this->ychanged = $this->xv = $this->yv = $this->xind = $this->yind = array();
  91          unset($this->seq, $this->in_seq, $this->lcs);
  92  
  93          // Skip leading common lines.
  94          for ($skip = 0; $skip < $n_from && $skip < $n_to; $skip++)
  95          {
  96              if (trim($from_lines[$skip]) !== trim($to_lines[$skip]))
  97              {
  98                  break;
  99              }
 100              $this->xchanged[$skip] = $this->ychanged[$skip] = false;
 101          }
 102  
 103          // Skip trailing common lines.
 104          $xi = $n_from;
 105          $yi = $n_to;
 106  
 107          for ($endskip = 0; --$xi > $skip && --$yi > $skip; $endskip++)
 108          {
 109              if (trim($from_lines[$xi]) !== trim($to_lines[$yi]))
 110              {
 111                  break;
 112              }
 113              $this->xchanged[$xi] = $this->ychanged[$yi] = false;
 114          }
 115  
 116          // Ignore lines which do not exist in both files.
 117          for ($xi = $skip; $xi < $n_from - $endskip; $xi++)
 118          {
 119              if ($this->skip_whitespace_changes) $xhash[trim($from_lines[$xi])] = 1; else $xhash[$from_lines[$xi]] = 1;
 120          }
 121  
 122          for ($yi = $skip; $yi < $n_to - $endskip; $yi++)
 123          {
 124              $line = ($this->skip_whitespace_changes) ? trim($to_lines[$yi]) : $to_lines[$yi];
 125  
 126              if (($this->ychanged[$yi] = empty($xhash[$line])))
 127              {
 128                  continue;
 129              }
 130              $yhash[$line] = 1;
 131              $this->yv[] = $line;
 132              $this->yind[] = $yi;
 133          }
 134  
 135          for ($xi = $skip; $xi < $n_from - $endskip; $xi++)
 136          {
 137              $line = ($this->skip_whitespace_changes) ? trim($from_lines[$xi]) : $from_lines[$xi];
 138  
 139              if (($this->xchanged[$xi] = empty($yhash[$line])))
 140              {
 141                  continue;
 142              }
 143              $this->xv[] = $line;
 144              $this->xind[] = $xi;
 145          }
 146  
 147          // Find the LCS.
 148          $this->_compareseq(0, sizeof($this->xv), 0, sizeof($this->yv));
 149  
 150          // Merge edits when possible.
 151          if ($this->skip_whitespace_changes)
 152          {
 153              $from_lines_clean = array_map('trim', $from_lines);
 154              $to_lines_clean = array_map('trim', $to_lines);
 155  
 156              $this->_shift_boundaries($from_lines_clean, $this->xchanged, $this->ychanged);
 157              $this->_shift_boundaries($to_lines_clean, $this->ychanged, $this->xchanged);
 158  
 159              unset($from_lines_clean, $to_lines_clean);
 160          }
 161          else
 162          {
 163              $this->_shift_boundaries($from_lines, $this->xchanged, $this->ychanged);
 164              $this->_shift_boundaries($to_lines, $this->ychanged, $this->xchanged);
 165          }
 166  
 167          // Compute the edit operations.
 168          $edits = array();
 169          $xi = $yi = 0;
 170  
 171          while ($xi < $n_from || $yi < $n_to)
 172          {
 173              // Skip matching "snake".
 174              $copy = array();
 175  
 176              while ($xi < $n_from && $yi < $n_to && !$this->xchanged[$xi] && !$this->ychanged[$yi])
 177              {
 178                  $copy[] = $from_lines[$xi++];
 179                  $yi++;
 180              }
 181  
 182              if ($copy)
 183              {
 184                  $edits[] = new diff_op_copy($copy);
 185              }
 186  
 187              // Find deletes & adds.
 188              $delete = array();
 189              while ($xi < $n_from && $this->xchanged[$xi])
 190              {
 191                  $delete[] = $from_lines[$xi++];
 192              }
 193  
 194              $add = array();
 195              while ($yi < $n_to && $this->ychanged[$yi])
 196              {
 197                  $add[] = $to_lines[$yi++];
 198              }
 199  
 200              if ($delete && $add)
 201              {
 202                  $edits[] = new diff_op_change($delete, $add);
 203              }
 204              else if ($delete)
 205              {
 206                  $edits[] = new diff_op_delete($delete);
 207              }
 208              else if ($add)
 209              {
 210                  $edits[] = new diff_op_add($add);
 211              }
 212          }
 213  
 214          return $edits;
 215      }
 216  
 217      /**
 218      * Divides the Largest Common Subsequence (LCS) of the sequences (XOFF,
 219      * XLIM) and (YOFF, YLIM) into NCHUNKS approximately equally sized segments.
 220      *
 221      * Returns (LCS, PTS).  LCS is the length of the LCS. PTS is an array of
 222      * NCHUNKS+1 (X, Y) indexes giving the diving points between sub
 223      * sequences.  The first sub-sequence is contained in (X0, X1), (Y0, Y1),
 224      * the second in (X1, X2), (Y1, Y2) and so on.  Note that (X0, Y0) ==
 225      * (XOFF, YOFF) and (X[NCHUNKS], Y[NCHUNKS]) == (XLIM, YLIM).
 226      *
 227      * This function assumes that the first lines of the specified portions of
 228      * the two files do not match, and likewise that the last lines do not
 229      * match.  The caller must trim matching lines from the beginning and end
 230      * of the portions it is going to specify.
 231      */
 232  	function _diag($xoff, $xlim, $yoff, $ylim, $nchunks)
 233      {
 234          $flip = false;
 235  
 236          if ($xlim - $xoff > $ylim - $yoff)
 237          {
 238              // Things seems faster (I'm not sure I understand why) when the shortest sequence is in X.
 239              $flip = true;
 240              list($xoff, $xlim, $yoff, $ylim) = array($yoff, $ylim, $xoff, $xlim);
 241          }
 242  
 243          if ($flip)
 244          {
 245              for ($i = $ylim - 1; $i >= $yoff; $i--)
 246              {
 247                  $ymatches[$this->xv[$i]][] = $i;
 248              }
 249          }
 250          else
 251          {
 252              for ($i = $ylim - 1; $i >= $yoff; $i--)
 253              {
 254                  $ymatches[$this->yv[$i]][] = $i;
 255              }
 256          }
 257  
 258          $this->lcs = 0;
 259          $this->seq[0]= $yoff - 1;
 260          $this->in_seq = array();
 261          $ymids[0] = array();
 262  
 263          $numer = $xlim - $xoff + $nchunks - 1;
 264          $x = $xoff;
 265  
 266          for ($chunk = 0; $chunk < $nchunks; $chunk++)
 267          {
 268              if ($chunk > 0)
 269              {
 270                  for ($i = 0; $i <= $this->lcs; $i++)
 271                  {
 272                      $ymids[$i][$chunk - 1] = $this->seq[$i];
 273                  }
 274              }
 275  
 276              $x1 = $xoff + (int)(($numer + ($xlim - $xoff) * $chunk) / $nchunks);
 277  
 278              for (; $x < $x1; $x++)
 279              {
 280                  $line = $flip ? $this->yv[$x] : $this->xv[$x];
 281                  if (empty($ymatches[$line]))
 282                  {
 283                      continue;
 284                  }
 285                  $matches = $ymatches[$line];
 286  
 287                  reset($matches);
 288                  while (list(, $y) = each($matches))
 289                  {
 290                      if (empty($this->in_seq[$y]))
 291                      {
 292                          $k = $this->_lcs_pos($y);
 293                          $ymids[$k] = $ymids[$k - 1];
 294                          break;
 295                      }
 296                  }
 297  
 298                  // no reset() here
 299                  while (list(, $y) = each($matches))
 300                  {
 301                      if ($y > $this->seq[$k - 1])
 302                      {
 303                          // Optimization: this is a common case: next match is just replacing previous match.
 304                          $this->in_seq[$this->seq[$k]] = false;
 305                          $this->seq[$k] = $y;
 306                          $this->in_seq[$y] = 1;
 307                      }
 308                      else if (empty($this->in_seq[$y]))
 309                      {
 310                          $k = $this->_lcs_pos($y);
 311                          $ymids[$k] = $ymids[$k - 1];
 312                      }
 313                  }
 314              }
 315          }
 316  
 317          $seps[] = $flip ? array($yoff, $xoff) : array($xoff, $yoff);
 318          $ymid = $ymids[$this->lcs];
 319  
 320          for ($n = 0; $n < $nchunks - 1; $n++)
 321          {
 322              $x1 = $xoff + (int)(($numer + ($xlim - $xoff) * $n) / $nchunks);
 323              $y1 = $ymid[$n] + 1;
 324              $seps[] = $flip ? array($y1, $x1) : array($x1, $y1);
 325          }
 326          $seps[] = $flip ? array($ylim, $xlim) : array($xlim, $ylim);
 327  
 328          return array($this->lcs, $seps);
 329      }
 330  
 331  	function _lcs_pos($ypos)
 332      {
 333          $end = $this->lcs;
 334  
 335          if ($end == 0 || $ypos > $this->seq[$end])
 336          {
 337              $this->seq[++$this->lcs] = $ypos;
 338              $this->in_seq[$ypos] = 1;
 339              return $this->lcs;
 340          }
 341  
 342          $beg = 1;
 343          while ($beg < $end)
 344          {
 345              $mid = (int)(($beg + $end) / 2);
 346              if ($ypos > $this->seq[$mid])
 347              {
 348                  $beg = $mid + 1;
 349              }
 350              else
 351              {
 352                  $end = $mid;
 353              }
 354          }
 355  
 356          $this->in_seq[$this->seq[$end]] = false;
 357          $this->seq[$end] = $ypos;
 358          $this->in_seq[$ypos] = 1;
 359  
 360          return $end;
 361      }
 362  
 363      /**
 364      * Finds LCS of two sequences.
 365      *
 366      * The results are recorded in the vectors $this->{x,y}changed[], by
 367      * storing a 1 in the element for each line that is an insertion or
 368      * deletion (ie. is not in the LCS).
 369      *
 370      * The subsequence of file 0 is (XOFF, XLIM) and likewise for file 1.
 371      *
 372      * Note that XLIM, YLIM are exclusive bounds.  All line numbers are
 373      * origin-0 and discarded lines are not counted.
 374      */
 375  	function _compareseq($xoff, $xlim, $yoff, $ylim)
 376      {
 377          // Slide down the bottom initial diagonal.
 378          while ($xoff < $xlim && $yoff < $ylim && $this->xv[$xoff] == $this->yv[$yoff])
 379          {
 380              ++$xoff;
 381              ++$yoff;
 382          }
 383  
 384          // Slide up the top initial diagonal.
 385          while ($xlim > $xoff && $ylim > $yoff && $this->xv[$xlim - 1] == $this->yv[$ylim - 1])
 386          {
 387              --$xlim;
 388              --$ylim;
 389          }
 390  
 391          if ($xoff == $xlim || $yoff == $ylim)
 392          {
 393              $lcs = 0;
 394          }
 395          else
 396          {
 397              // This is ad hoc but seems to work well.
 398              // $nchunks = sqrt(min($xlim - $xoff, $ylim - $yoff) / 2.5);
 399              // $nchunks = max(2,min(8,(int)$nchunks));
 400              $nchunks = min(7, $xlim - $xoff, $ylim - $yoff) + 1;
 401              list($lcs, $seps) = $this->_diag($xoff, $xlim, $yoff, $ylim, $nchunks);
 402          }
 403  
 404          if ($lcs == 0)
 405          {
 406              // X and Y sequences have no common subsequence: mark all changed.
 407              while ($yoff < $ylim)
 408              {
 409                  $this->ychanged[$this->yind[$yoff++]] = 1;
 410              }
 411  
 412              while ($xoff < $xlim)
 413              {
 414                  $this->xchanged[$this->xind[$xoff++]] = 1;
 415              }
 416          }
 417          else
 418          {
 419              // Use the partitions to split this problem into subproblems.
 420              reset($seps);
 421              $pt1 = $seps[0];
 422  
 423              while ($pt2 = next($seps))
 424              {
 425                  $this->_compareseq($pt1[0], $pt2[0], $pt1[1], $pt2[1]);
 426                  $pt1 = $pt2;
 427              }
 428          }
 429      }
 430  
 431      /**
 432      * Adjusts inserts/deletes of identical lines to join changes as much as possible.
 433      *
 434      * We do something when a run of changed lines include a line at one end
 435      * and has an excluded, identical line at the other.  We are free to
 436      * choose which identical line is included. 'compareseq' usually chooses
 437      * the one at the beginning, but usually it is cleaner to consider the
 438      * following identical line to be the "change".
 439      *
 440      * This is extracted verbatim from analyze.c (GNU diffutils-2.7).
 441      */
 442  	function _shift_boundaries($lines, &$changed, $other_changed)
 443      {
 444          $i = 0;
 445          $j = 0;
 446  
 447          $len = sizeof($lines);
 448          $other_len = sizeof($other_changed);
 449  
 450          while (1)
 451          {
 452              // Scan forward to find the beginning of another run of
 453              // changes. Also keep track of the corresponding point in the other file.
 454              //
 455              // Throughout this code, $i and $j are adjusted together so that
 456              // the first $i elements of $changed and the first $j elements of
 457              // $other_changed both contain the same number of zeros (unchanged lines).
 458              //
 459              // Furthermore, $j is always kept so that $j == $other_len or $other_changed[$j] == false.
 460              while ($j < $other_len && $other_changed[$j])
 461              {
 462                  $j++;
 463              }
 464  
 465              while ($i < $len && ! $changed[$i])
 466              {
 467                  $i++;
 468                  $j++;
 469  
 470                  while ($j < $other_len && $other_changed[$j])
 471                  {
 472                      $j++;
 473                  }
 474              }
 475  
 476              if ($i == $len)
 477              {
 478                  break;
 479              }
 480  
 481              $start = $i;
 482  
 483              // Find the end of this run of changes.
 484              while (++$i < $len && $changed[$i])
 485              {
 486                  continue;
 487              }
 488  
 489              do
 490              {
 491                  // Record the length of this run of changes, so that we can later determine whether the run has grown.
 492                  $runlength = $i - $start;
 493  
 494                  // Move the changed region back, so long as the previous unchanged line matches the last changed one.
 495                  // This merges with previous changed regions.
 496                  while ($start > 0 && $lines[$start - 1] == $lines[$i - 1])
 497                  {
 498                      $changed[--$start] = 1;
 499                      $changed[--$i] = false;
 500  
 501                      while ($start > 0 && $changed[$start - 1])
 502                      {
 503                          $start--;
 504                      }
 505  
 506                      while ($other_changed[--$j])
 507                      {
 508                          continue;
 509                      }
 510                  }
 511  
 512                  // Set CORRESPONDING to the end of the changed run, at the last point where it corresponds to a changed run in the
 513                  // other file. CORRESPONDING == LEN means no such point has been found.
 514                  $corresponding = $j < $other_len ? $i : $len;
 515  
 516                  // Move the changed region forward, so long as the first changed line matches the following unchanged one.
 517                  // This merges with following changed regions.
 518                  // Do this second, so that if there are no merges, the changed region is moved forward as far as possible.
 519                  while ($i < $len && $lines[$start] == $lines[$i])
 520                  {
 521                      $changed[$start++] = false;
 522                      $changed[$i++] = 1;
 523  
 524                      while ($i < $len && $changed[$i])
 525                      {
 526                          $i++;
 527                      }
 528  
 529                      $j++;
 530                      if ($j < $other_len && $other_changed[$j])
 531                      {
 532                          $corresponding = $i;
 533                          while ($j < $other_len && $other_changed[$j])
 534                          {
 535                              $j++;
 536                          }
 537                      }
 538                  }
 539              }
 540              while ($runlength != $i - $start);
 541  
 542              // If possible, move the fully-merged run of changes back to a corresponding run in the other file.
 543              while ($corresponding < $i)
 544              {
 545                  $changed[--$start] = 1;
 546                  $changed[--$i] = 0;
 547  
 548                  while ($other_changed[--$j])
 549                  {
 550                      continue;
 551                  }
 552              }
 553          }
 554      }
 555  }