3 * A PHP diff engine for phpwiki. (Taken from phpwiki-1.3.3)
5 * Copyright © 2000, 2001 Geoffrey T. Dairiki <dairiki@dairiki.org>
6 * You may copy this code freely under the conditions of the GPL.
8 * This program is free software; you can redistribute it and/or modify
9 * it under the terms of the GNU General Public License as published by
10 * the Free Software Foundation; either version 2 of the License, or
11 * (at your option) any later version.
13 * This program is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 * GNU General Public License for more details.
18 * You should have received a copy of the GNU General Public License along
19 * with this program; if not, write to the Free Software Foundation, Inc.,
20 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
21 * http://www.gnu.org/copyleft/gpl.html
24 * @ingroup DifferenceEngine
25 * @defgroup DifferenceEngine DifferenceEngine
31 * @ingroup DifferenceEngine
33 abstract class DiffOp
{
53 public function getType() {
60 public function getOrig() {
68 public function getClosing( $i = null ) {
70 return $this->closing
;
72 if ( array_key_exists( $i, $this->closing
) ) {
73 return $this->closing
[$i];
78 abstract public function reverse();
83 public function norig() {
84 return $this->orig ?
count( $this->orig
) : 0;
90 public function nclosing() {
91 return $this->closing ?
count( $this->closing
) : 0;
98 * @ingroup DifferenceEngine
100 class DiffOpCopy
extends DiffOp
{
101 public $type = 'copy';
103 public function __construct( $orig, $closing = false ) {
104 if ( !is_array( $closing ) ) {
108 $this->closing
= $closing;
114 public function reverse() {
115 return new DiffOpCopy( $this->closing
, $this->orig
);
122 * @ingroup DifferenceEngine
124 class DiffOpDelete
extends DiffOp
{
125 public $type = 'delete';
127 public function __construct( $lines ) {
128 $this->orig
= $lines;
129 $this->closing
= false;
135 public function reverse() {
136 return new DiffOpAdd( $this->orig
);
143 * @ingroup DifferenceEngine
145 class DiffOpAdd
extends DiffOp
{
146 public $type = 'add';
148 public function __construct( $lines ) {
149 $this->closing
= $lines;
154 * @return DiffOpDelete
156 public function reverse() {
157 return new DiffOpDelete( $this->closing
);
164 * @ingroup DifferenceEngine
166 class DiffOpChange
extends DiffOp
{
167 public $type = 'change';
169 public function __construct( $orig, $closing ) {
171 $this->closing
= $closing;
175 * @return DiffOpChange
177 public function reverse() {
178 return new DiffOpChange( $this->closing
, $this->orig
);
183 * Class used internally by Diff to actually compute the diffs.
185 * The algorithm used here is mostly lifted from the perl module
186 * Algorithm::Diff (version 1.06) by Ned Konz, which is available at:
187 * http://www.perl.com/CPAN/authors/id/N/NE/NEDKONZ/Algorithm-Diff-1.06.zip
189 * More ideas are taken from:
190 * http://www.ics.uci.edu/~eppstein/161/960229.html
192 * Some ideas (and a bit of code) are from analyze.c, from GNU
193 * diffutils-2.7, which can be found at:
194 * ftp://gnudist.gnu.org/pub/gnu/diffutils/diffutils-2.7.tar.gz
196 * closingly, some ideas (subdivision by NCHUNKS > 2, and some optimizations)
199 * Line length limits for robustness added by Tim Starling, 2005-08-31
200 * Alternative implementation added by Guy Van den Broeck, 2008-07-30
202 * @author Geoffrey T. Dairiki, Tim Starling, Guy Van den Broeck
204 * @ingroup DifferenceEngine
207 const MAX_XREF_LENGTH
= 10000;
209 protected $xchanged, $ychanged;
211 protected $xv = array(), $yv = array();
212 protected $xind = array(), $yind = array();
214 protected $seq = array(), $in_seq = array();
219 * @param string[] $from_lines
220 * @param string[] $to_lines
224 public function diff( $from_lines, $to_lines ) {
226 // Diff and store locally
227 $this->diffLocal( $from_lines, $to_lines );
229 // Merge edits when possible
230 $this->shiftBoundaries( $from_lines, $this->xchanged
, $this->ychanged
);
231 $this->shiftBoundaries( $to_lines, $this->ychanged
, $this->xchanged
);
233 // Compute the edit operations.
234 $n_from = count( $from_lines );
235 $n_to = count( $to_lines );
239 while ( $xi < $n_from ||
$yi < $n_to ) {
240 assert( '$yi < $n_to || $this->xchanged[$xi]' );
241 assert( '$xi < $n_from || $this->ychanged[$yi]' );
243 // Skip matching "snake".
245 while ( $xi < $n_from && $yi < $n_to
246 && !$this->xchanged
[$xi] && !$this->ychanged
[$yi]
248 $copy[] = $from_lines[$xi++
];
252 $edits[] = new DiffOpCopy( $copy );
255 // Find deletes & adds.
257 while ( $xi < $n_from && $this->xchanged
[$xi] ) {
258 $delete[] = $from_lines[$xi++
];
262 while ( $yi < $n_to && $this->ychanged
[$yi] ) {
263 $add[] = $to_lines[$yi++
];
266 if ( $delete && $add ) {
267 $edits[] = new DiffOpChange( $delete, $add );
268 } elseif ( $delete ) {
269 $edits[] = new DiffOpDelete( $delete );
271 $edits[] = new DiffOpAdd( $add );
279 * @param string[] $from_lines
280 * @param string[] $to_lines
282 private function diffLocal( $from_lines, $to_lines ) {
283 global $wgExternalDiffEngine;
285 if ( $wgExternalDiffEngine == 'wikidiff3' ) {
287 $wikidiff3 = new WikiDiff3();
288 $wikidiff3->diff( $from_lines, $to_lines );
289 $this->xchanged
= $wikidiff3->removed
;
290 $this->ychanged
= $wikidiff3->added
;
294 $n_from = count( $from_lines );
295 $n_to = count( $to_lines );
296 $this->xchanged
= $this->ychanged
= array();
297 $this->xv
= $this->yv
= array();
298 $this->xind
= $this->yind
= array();
299 $this->seq
= array();
300 $this->in_seq
= array();
303 // Skip leading common lines.
304 for ( $skip = 0; $skip < $n_from && $skip < $n_to; $skip++
) {
305 if ( $from_lines[$skip] !== $to_lines[$skip] ) {
308 $this->xchanged
[$skip] = $this->ychanged
[$skip] = false;
310 // Skip trailing common lines.
313 for ( $endskip = 0; --$xi > $skip && --$yi > $skip; $endskip++
) {
314 if ( $from_lines[$xi] !== $to_lines[$yi] ) {
317 $this->xchanged
[$xi] = $this->ychanged
[$yi] = false;
320 // Ignore lines which do not exist in both files.
321 for ( $xi = $skip; $xi < $n_from - $endskip; $xi++
) {
322 $xhash[$this->lineHash( $from_lines[$xi] )] = 1;
325 for ( $yi = $skip; $yi < $n_to - $endskip; $yi++
) {
326 $line = $to_lines[$yi];
327 $this->ychanged
[$yi] = empty( $xhash[$this->lineHash( $line )] );
328 if ( $this->ychanged
[$yi] ) {
331 $yhash[$this->lineHash( $line )] = 1;
335 for ( $xi = $skip; $xi < $n_from - $endskip; $xi++
) {
336 $line = $from_lines[$xi];
337 $this->xchanged
[$xi] = empty( $yhash[$this->lineHash( $line )] );
338 if ( $this->xchanged
[$xi] ) {
346 $this->compareSeq( 0, count( $this->xv
), 0, count( $this->yv
) );
351 * Returns the whole line if it's small enough, or the MD5 hash otherwise
353 * @param string $line
357 private function lineHash( $line ) {
358 if ( strlen( $line ) > self
::MAX_XREF_LENGTH
) {
366 * Divide the Largest Common Subsequence (LCS) of the sequences
367 * [XOFF, XLIM) and [YOFF, YLIM) into NCHUNKS approximately equally
370 * Returns (LCS, PTS). LCS is the length of the LCS. PTS is an
371 * array of NCHUNKS+1 (X, Y) indexes giving the diving points between
372 * sub sequences. The first sub-sequence is contained in [X0, X1),
373 * [Y0, Y1), the second in [X1, X2), [Y1, Y2) and so on. Note
374 * that (X0, Y0) == (XOFF, YOFF) and
375 * (X[NCHUNKS], Y[NCHUNKS]) == (XLIM, YLIM).
377 * This function assumes that the first lines of the specified portions
378 * of the two files do not match, and likewise that the last lines do not
379 * match. The caller must trim matching lines from the beginning and end
380 * of the portions it is going to specify.
386 * @param int $nchunks
388 * @return array List of two elements, integer and array[].
390 private function diag( $xoff, $xlim, $yoff, $ylim, $nchunks ) {
393 if ( $xlim - $xoff > $ylim - $yoff ) {
394 // Things seems faster (I'm not sure I understand why)
395 // when the shortest sequence in X.
397 list( $xoff, $xlim, $yoff, $ylim ) = array( $yoff, $ylim, $xoff, $xlim );
401 for ( $i = $ylim - 1; $i >= $yoff; $i-- ) {
402 $ymatches[$this->xv
[$i]][] = $i;
405 for ( $i = $ylim - 1; $i >= $yoff; $i-- ) {
406 $ymatches[$this->yv
[$i]][] = $i;
411 $this->seq
[0] = $yoff - 1;
412 $this->in_seq
= array();
415 $numer = $xlim - $xoff +
$nchunks - 1;
417 for ( $chunk = 0; $chunk < $nchunks; $chunk++
) {
419 for ( $i = 0; $i <= $this->lcs
; $i++
) {
420 $ymids[$i][$chunk - 1] = $this->seq
[$i];
424 $x1 = $xoff +
(int)( ( $numer +
( $xlim - $xoff ) * $chunk ) / $nchunks );
425 // @codingStandardsIgnoreStart Ignore Squiz.WhiteSpace.SemicolonSpacing.Incorrect
426 for ( ; $x < $x1; $x++
) {
427 // @codingStandardsIgnoreEnd
428 $line = $flip ?
$this->yv
[$x] : $this->xv
[$x];
429 if ( empty( $ymatches[$line] ) ) {
434 $matches = $ymatches[$line];
436 while ( list( , $y ) = each( $matches ) ) {
437 if ( empty( $this->in_seq
[$y] ) ) {
438 $k = $this->lcsPos( $y );
440 $ymids[$k] = $ymids[$k - 1];
445 while ( list( , $y ) = each( $matches ) ) {
446 if ( $y > $this->seq
[$k - 1] ) {
447 assert( '$y < $this->seq[$k]' );
448 // Optimization: this is a common case:
449 // next match is just replacing previous match.
450 $this->in_seq
[$this->seq
[$k]] = false;
452 $this->in_seq
[$y] = 1;
453 } elseif ( empty( $this->in_seq
[$y] ) ) {
454 $k = $this->lcsPos( $y );
456 $ymids[$k] = $ymids[$k - 1];
462 $seps[] = $flip ?
array( $yoff, $xoff ) : array( $xoff, $yoff );
463 $ymid = $ymids[$this->lcs
];
464 for ( $n = 0; $n < $nchunks - 1; $n++
) {
465 $x1 = $xoff +
(int)( ( $numer +
( $xlim - $xoff ) * $n ) / $nchunks );
467 $seps[] = $flip ?
array( $y1, $x1 ) : array( $x1, $y1 );
469 $seps[] = $flip ?
array( $ylim, $xlim ) : array( $xlim, $ylim );
471 return array( $this->lcs
, $seps );
479 private function lcsPos( $ypos ) {
481 if ( $end == 0 ||
$ypos > $this->seq
[$end] ) {
482 $this->seq
[++
$this->lcs
] = $ypos;
483 $this->in_seq
[$ypos] = 1;
489 while ( $beg < $end ) {
490 $mid = (int)( ( $beg +
$end ) / 2 );
491 if ( $ypos > $this->seq
[$mid] ) {
498 assert( '$ypos != $this->seq[$end]' );
500 $this->in_seq
[$this->seq
[$end]] = false;
501 $this->seq
[$end] = $ypos;
502 $this->in_seq
[$ypos] = 1;
508 * Find LCS of two sequences.
510 * The results are recorded in the vectors $this->{x,y}changed[], by
511 * storing a 1 in the element for each line that is an insertion
512 * or deletion (ie. is not in the LCS).
514 * The subsequence of file 0 is [XOFF, XLIM) and likewise for file 1.
516 * Note that XLIM, YLIM are exclusive bounds.
517 * All line numbers are origin-0 and discarded lines are not counted.
524 private function compareSeq( $xoff, $xlim, $yoff, $ylim ) {
525 // Slide down the bottom initial diagonal.
526 while ( $xoff < $xlim && $yoff < $ylim && $this->xv
[$xoff] == $this->yv
[$yoff] ) {
531 // Slide up the top initial diagonal.
532 while ( $xlim > $xoff && $ylim > $yoff
533 && $this->xv
[$xlim - 1] == $this->yv
[$ylim - 1]
539 if ( $xoff == $xlim ||
$yoff == $ylim ) {
542 // This is ad hoc but seems to work well.
543 // $nchunks = sqrt(min($xlim - $xoff, $ylim - $yoff) / 2.5);
544 // $nchunks = max(2,min(8,(int)$nchunks));
545 $nchunks = min( 7, $xlim - $xoff, $ylim - $yoff ) +
1;
546 list( $lcs, $seps ) = $this->diag( $xoff, $xlim, $yoff, $ylim, $nchunks );
550 // X and Y sequences have no common subsequence:
552 while ( $yoff < $ylim ) {
553 $this->ychanged
[$this->yind
[$yoff++
]] = 1;
555 while ( $xoff < $xlim ) {
556 $this->xchanged
[$this->xind
[$xoff++
]] = 1;
559 // Use the partitions to split this problem into subproblems.
562 while ( $pt2 = next( $seps ) ) {
563 $this->compareSeq( $pt1[0], $pt2[0], $pt1[1], $pt2[1] );
570 * Adjust inserts/deletes of identical lines to join changes
571 * as much as possible.
573 * We do something when a run of changed lines include a
574 * line at one end and has an excluded, identical line at the other.
575 * We are free to choose which identical line is included.
576 * `compareseq' usually chooses the one at the beginning,
577 * but usually it is cleaner to consider the following identical line
578 * to be the "change".
580 * This is extracted verbatim from analyze.c (GNU diffutils-2.7).
582 private function shiftBoundaries( $lines, &$changed, $other_changed ) {
586 assert( 'count($lines) == count($changed)' );
587 $len = count( $lines );
588 $other_len = count( $other_changed );
592 * Scan forwards to find beginning of another run of changes.
593 * Also keep track of the corresponding point in the other file.
595 * Throughout this code, $i and $j are adjusted together so that
596 * the first $i elements of $changed and the first $j elements
597 * of $other_changed both contain the same number of zeros
599 * Furthermore, $j is always kept so that $j == $other_len or
600 * $other_changed[$j] == false.
602 while ( $j < $other_len && $other_changed[$j] ) {
606 while ( $i < $len && !$changed[$i] ) {
607 assert( '$j < $other_len && ! $other_changed[$j]' );
610 while ( $j < $other_len && $other_changed[$j] ) {
621 // Find the end of this run of changes.
622 while ( ++
$i < $len && $changed[$i] ) {
628 * Record the length of this run of changes, so that
629 * we can later determine whether the run has grown.
631 $runlength = $i - $start;
634 * Move the changed region back, so long as the
635 * previous unchanged line matches the last changed one.
636 * This merges with previous changed regions.
638 while ( $start > 0 && $lines[$start - 1] == $lines[$i - 1] ) {
639 $changed[--$start] = 1;
640 $changed[--$i] = false;
641 while ( $start > 0 && $changed[$start - 1] ) {
645 while ( $other_changed[--$j] ) {
648 assert( '$j >= 0 && !$other_changed[$j]' );
652 * Set CORRESPONDING to the end of the changed run, at the last
653 * point where it corresponds to a changed run in the other file.
654 * CORRESPONDING == LEN means no such point has been found.
656 $corresponding = $j < $other_len ?
$i : $len;
659 * Move the changed region forward, so long as the
660 * first changed line matches the following unchanged one.
661 * This merges with following changed regions.
662 * Do this second, so that if there are no merges,
663 * the changed region is moved forward as far as possible.
665 while ( $i < $len && $lines[$start] == $lines[$i] ) {
666 $changed[$start++
] = false;
668 while ( $i < $len && $changed[$i] ) {
672 assert( '$j < $other_len && ! $other_changed[$j]' );
674 if ( $j < $other_len && $other_changed[$j] ) {
676 while ( $j < $other_len && $other_changed[$j] ) {
681 } while ( $runlength != $i - $start );
684 * If possible, move the fully-merged run of changes
685 * back to a corresponding run in the other file.
687 while ( $corresponding < $i ) {
688 $changed[--$start] = 1;
691 while ( $other_changed[--$j] ) {
694 assert( '$j >= 0 && !$other_changed[$j]' );
701 * Class representing a 'diff' between two sequences of strings.
704 * @ingroup DifferenceEngine
715 * Computes diff between sequences of strings.
717 * @param string[] $from_lines An array of strings.
718 * Typically these are lines from a file.
719 * @param string[] $to_lines An array of strings.
721 public function __construct( $from_lines, $to_lines ) {
722 $eng = new DiffEngine
;
723 $this->edits
= $eng->diff( $from_lines, $to_lines );
729 public function getEdits() {
734 * Compute reversed Diff.
738 * $diff = new Diff($lines1, $lines2);
739 * $rev = $diff->reverse();
741 * @return Object A Diff object representing the inverse of the
744 public function reverse() {
746 $rev->edits
= array();
747 /** @var DiffOp $edit */
748 foreach ( $this->edits
as $edit ) {
749 $rev->edits
[] = $edit->reverse();
756 * Check for empty diff.
758 * @return bool True if two sequences were identical.
760 public function isEmpty() {
761 foreach ( $this->edits
as $edit ) {
762 if ( $edit->type
!= 'copy' ) {
771 * Compute the length of the Longest Common Subsequence (LCS).
773 * This is mostly for diagnostic purposed.
775 * @return int The length of the LCS.
777 public function lcs() {
779 foreach ( $this->edits
as $edit ) {
780 if ( $edit->type
== 'copy' ) {
781 $lcs +
= count( $edit->orig
);
789 * Get the original set of lines.
791 * This reconstructs the $from_lines parameter passed to the
794 * @return string[] The original sequence of strings.
796 public function orig() {
799 foreach ( $this->edits
as $edit ) {
801 array_splice( $lines, count( $lines ), 0, $edit->orig
);
809 * Get the closing set of lines.
811 * This reconstructs the $to_lines parameter passed to the
814 * @return string[] The sequence of strings.
816 public function closing() {
819 foreach ( $this->edits
as $edit ) {
820 if ( $edit->closing
) {
821 array_splice( $lines, count( $lines ), 0, $edit->closing
);
830 * @todo document, bad name.
832 * @ingroup DifferenceEngine
834 class MappedDiff
extends Diff
{
838 * Computes diff between sequences of strings.
840 * This can be used to compute things like
841 * case-insensitve diffs, or diffs which ignore
842 * changes in white-space.
844 * @param string[] $from_lines An array of strings.
845 * Typically these are lines from a file.
846 * @param string[] $to_lines An array of strings.
847 * @param string[] $mapped_from_lines This array should
848 * have the same size number of elements as $from_lines.
849 * The elements in $mapped_from_lines and
850 * $mapped_to_lines are what is actually compared
851 * when computing the diff.
852 * @param string[] $mapped_to_lines This array should
853 * have the same number of elements as $to_lines.
855 public function __construct( $from_lines, $to_lines,
856 $mapped_from_lines, $mapped_to_lines ) {
858 assert( 'count( $from_lines ) == count( $mapped_from_lines )' );
859 assert( 'count( $to_lines ) == count( $mapped_to_lines )' );
861 parent
::__construct( $mapped_from_lines, $mapped_to_lines );
864 $editCount = count( $this->edits
);
865 for ( $i = 0; $i < $editCount; $i++
) {
866 $orig = &$this->edits
[$i]->orig
;
867 if ( is_array( $orig ) ) {
868 $orig = array_slice( $from_lines, $xi, count( $orig ) );
869 $xi +
= count( $orig );
872 $closing = &$this->edits
[$i]->closing
;
873 if ( is_array( $closing ) ) {
874 $closing = array_slice( $to_lines, $yi, count( $closing ) );
875 $yi +
= count( $closing );
882 * Additions by Axel Boldt follow, partly taken from diff.php, phpwiki-1.3.3
888 * @ingroup DifferenceEngine
890 class HWLDFWordAccumulator
{
891 public $insClass = ' class="diffchange diffchange-inline"';
892 public $delClass = ' class="diffchange diffchange-inline"';
894 private $lines = array();
900 * @param string $new_tag
902 private function flushGroup( $new_tag ) {
903 if ( $this->group
!== '' ) {
904 if ( $this->tag
== 'ins' ) {
905 $this->line
.= "<ins{$this->insClass}>" .
906 htmlspecialchars( $this->group
) . '</ins>';
907 } elseif ( $this->tag
== 'del' ) {
908 $this->line
.= "<del{$this->delClass}>" .
909 htmlspecialchars( $this->group
) . '</del>';
911 $this->line
.= htmlspecialchars( $this->group
);
915 $this->tag
= $new_tag;
919 * @param string $new_tag
921 private function flushLine( $new_tag ) {
922 $this->flushGroup( $new_tag );
923 if ( $this->line
!= '' ) {
924 array_push( $this->lines
, $this->line
);
926 # make empty lines visible by inserting an NBSP
927 array_push( $this->lines
, ' ' );
933 * @param string[] $words
936 public function addWords( $words, $tag = '' ) {
937 if ( $tag != $this->tag
) {
938 $this->flushGroup( $tag );
941 foreach ( $words as $word ) {
942 // new-line should only come as first char of word.
946 if ( $word[0] == "\n" ) {
947 $this->flushLine( $tag );
948 $word = substr( $word, 1 );
950 assert( '!strstr( $word, "\n" )' );
951 $this->group
.= $word;
958 public function getLines() {
959 $this->flushLine( '~done' );
968 * @ingroup DifferenceEngine
970 class WordLevelDiff
extends MappedDiff
{
971 const MAX_LINE_LENGTH
= 10000;
974 * @param string[] $orig_lines
975 * @param string[] $closing_lines
977 public function __construct( $orig_lines, $closing_lines ) {
979 list( $orig_words, $orig_stripped ) = $this->split( $orig_lines );
980 list( $closing_words, $closing_stripped ) = $this->split( $closing_lines );
982 parent
::__construct( $orig_words, $closing_words,
983 $orig_stripped, $closing_stripped );
987 * @param string[] $lines
991 private function split( $lines ) {
996 foreach ( $lines as $line ) {
997 # If the line is too long, just pretend the entire line is one big word
998 # This prevents resource exhaustion problems
1005 if ( strlen( $line ) > self
::MAX_LINE_LENGTH
) {
1007 $stripped[] = $line;
1010 if ( preg_match_all( '/ ( [^\S\n]+ | [0-9_A-Za-z\x80-\xff]+ | . ) (?: (?!< \n) [^\S\n])? /xs',
1013 foreach ( $m[0] as $word ) {
1016 foreach ( $m[1] as $stripped_word ) {
1017 $stripped[] = $stripped_word;
1023 return array( $words, $stripped );
1029 public function orig() {
1030 $orig = new HWLDFWordAccumulator
;
1032 foreach ( $this->edits
as $edit ) {
1033 if ( $edit->type
== 'copy' ) {
1034 $orig->addWords( $edit->orig
);
1035 } elseif ( $edit->orig
) {
1036 $orig->addWords( $edit->orig
, 'del' );
1039 $lines = $orig->getLines();
1047 public function closing() {
1048 $closing = new HWLDFWordAccumulator
;
1050 foreach ( $this->edits
as $edit ) {
1051 if ( $edit->type
== 'copy' ) {
1052 $closing->addWords( $edit->closing
);
1053 } elseif ( $edit->closing
) {
1054 $closing->addWords( $edit->closing
, 'ins' );
1057 $lines = $closing->getLines();