* General API for generating and formatting diffs - the differences between
* two sequences of strings.
* The original PHP version of this code was written by Geoffrey T. Dairiki
* <dairiki@dairiki.org>, and is used/adapted with his permission.
* Copyright 2004 Geoffrey T. Dairiki <dairiki@dairiki.org>
* Copyright 2004-2010 The Horde Project (http://www.horde.org/)
* See the enclosed file COPYING for license information (LGPL). If you did
* not receive this file, see http://opensource.org/licenses/lgpl-license.php.
* @author Geoffrey T. Dairiki <dairiki@dairiki.org>
* Computes diffs between sequences of strings.
* @param string $engine Name of the diffing engine to use. 'auto'
* will automatically select the best.
* @param array $params Parameters to pass to the diffing engine.
* Normally an array of two arrays, each
* containing the lines from a file.
function __construct( $engine, $params )
// Backward compatibility workaround.
if (!is_string($engine)) {
$params = array($engine, $params);
$engine = extension_loaded('xdiff') ? 'xdiff' : 'native';
$engine = basename($engine);
require_once dirname(__FILE__).'/Diff/Engine/' . $engine . '.php';
$class = 'Text_Diff_Engine_' . $engine;
$diff_engine = new $class();
$this->_edits = call_user_func_array(array($diff_engine, 'diff'), $params);
public function Text_Diff( $engine, $params ) {
self::__construct( $engine, $params );
* Returns the array of differences.
* returns the number of new (added) lines in a given diff.
* @return int The number of new lines
function countAddedLines()
foreach ($this->_edits as $edit) {
if (is_a($edit, 'Text_Diff_Op_add') ||
is_a($edit, 'Text_Diff_Op_change')) {
$count += $edit->nfinal();
* Returns the number of deleted (removed) lines in a given diff.
* @return int The number of deleted lines
function countDeletedLines()
foreach ($this->_edits as $edit) {
if (is_a($edit, 'Text_Diff_Op_delete') ||
is_a($edit, 'Text_Diff_Op_change')) {
$count += $edit->norig();
* Computes a reversed diff.
* $diff = new Text_Diff($lines1, $lines2);
* $rev = $diff->reverse();
* @return Text_Diff A Diff object representing the inverse of the
* original diff. Note that we purposely don't return a
* reference here, since this essentially is a clone()
if (version_compare(zend_version(), '2', '>')) {
foreach ($this->_edits as $edit) {
$rev->_edits[] = $edit->reverse();
* Checks for an empty diff.
* @return bool True if two sequences were identical.
foreach ($this->_edits as $edit) {
if (!is_a($edit, 'Text_Diff_Op_copy')) {
* Computes the length of the Longest Common Subsequence (LCS).
* This is mostly for diagnostic purposes.
* @return int The length of the LCS.
foreach ($this->_edits as $edit) {
if (is_a($edit, 'Text_Diff_Op_copy')) {
$lcs += count($edit->orig);
* Gets the original set of lines.
* This reconstructs the $from_lines parameter passed to the constructor.
* @return array The original sequence of strings.
foreach ($this->_edits as $edit) {
array_splice($lines, count($lines), 0, $edit->orig);
* Gets the final set of lines.
* This reconstructs the $to_lines parameter passed to the constructor.
* @return array The sequence of strings.
foreach ($this->_edits as $edit) {
array_splice($lines, count($lines), 0, $edit->final);
* Removes trailing newlines from a line of text. This is meant to be used
* @param string $line The line to trim.
* @param int $key The index of the line in the array. Not used.
static function trimNewlines(&$line, $key)
$line = str_replace(array("\n", "\r"), '', $line);
* Determines the location of the system temporary directory.
* @return string A directory name which can be used for temp files.
* Returns false if one could not be found.
static function _getTempDir()
$tmp_locations = array('/tmp', '/var/tmp', 'c:\WUTemp', 'c:\temp',
'c:\windows\temp', 'c:\winnt\temp');
/* Try PHP's upload_tmp_dir directive. */
$tmp = ini_get('upload_tmp_dir');
/* Otherwise, try to determine the TMPDIR environment variable. */
/* If we still cannot determine a value, then cycle through a list of
* preset possibilities. */
while (!strlen($tmp) && count($tmp_locations)) {
$tmp_check = array_shift($tmp_locations);
if (@is_dir($tmp_check)) {
/* If it is still empty, we have failed, so return false; otherwise
* return the directory determined. */
return strlen($tmp) ? $tmp : false;
* Checks a diff for validity.
* This is here only for debugging purposes.
function _check($from_lines, $to_lines)
if (serialize($from_lines) != serialize($this->getOriginal())) {
trigger_error("Reconstructed original doesn't match", E_USER_ERROR);
if (serialize($to_lines) != serialize($this->getFinal())) {
trigger_error("Reconstructed final doesn't match", E_USER_ERROR);
if (serialize($to_lines) != serialize($rev->getOriginal())) {
trigger_error("Reversed original doesn't match", E_USER_ERROR);
if (serialize($from_lines) != serialize($rev->getFinal())) {
trigger_error("Reversed final doesn't match", E_USER_ERROR);
foreach ($this->_edits as $edit) {
if ($edit instanceof $prevtype) {
trigger_error("Edit sequence is non-optimal", E_USER_ERROR);
$prevtype = get_class($edit);
* @author Geoffrey T. Dairiki <dairiki@dairiki.org>
class Text_MappedDiff extends Text_Diff {
* Computes a diff between sequences of strings.
* This can be used to compute things like case-insensitve diffs, or diffs
* which ignore changes in white-space.
* @param array $from_lines An array of strings.
* @param array $to_lines An array of strings.
* @param array $mapped_from_lines This array should have the same size
* number of elements as $from_lines. The
* elements in $mapped_from_lines and
* $mapped_to_lines are what is actually
* compared when computing the diff.
* @param array $mapped_to_lines This array should have the same number
* of elements as $to_lines.
function __construct($from_lines, $to_lines,
$mapped_from_lines, $mapped_to_lines)
assert(count($from_lines) == count($mapped_from_lines));
assert(count($to_lines) == count($mapped_to_lines));
parent::Text_Diff($mapped_from_lines, $mapped_to_lines);
for ($i = 0; $i < count($this->_edits); $i++) {
$orig = &$this->_edits[$i]->orig;
$orig = array_slice($from_lines, $xi, count($orig));
$final = &$this->_edits[$i]->final;
$final = array_slice($to_lines, $yi, count($final));
public function Text_MappedDiff( $from_lines, $to_lines,
$mapped_from_lines, $mapped_to_lines ) {
self::__construct( $from_lines, $to_lines,
$mapped_from_lines, $mapped_to_lines );
* @author Geoffrey T. Dairiki <dairiki@dairiki.org>
trigger_error('Abstract method', E_USER_ERROR);
return $this->orig ? count($this->orig) : 0;
return $this->final ? count($this->final) : 0;
* @author Geoffrey T. Dairiki <dairiki@dairiki.org>
class Text_Diff_Op_copy extends Text_Diff_Op {
function __construct( $orig, $final = false )
public function Text_Diff_Op_copy( $orig, $final = false ) {
self::__construct( $orig, $final );
$reverse = new Text_Diff_Op_copy($this->final, $this->orig);
* @author Geoffrey T. Dairiki <dairiki@dairiki.org>
class Text_Diff_Op_delete extends Text_Diff_Op {
function __construct( $lines )
public function Text_Diff_Op_delete( $lines ) {
self::__construct( $lines );
$reverse = new Text_Diff_Op_add($this->orig);
* @author Geoffrey T. Dairiki <dairiki@dairiki.org>
class Text_Diff_Op_add extends Text_Diff_Op {
function __construct( $lines )
public function Text_Diff_Op_add( $lines ) {
self::__construct( $lines );
$reverse = new Text_Diff_Op_delete($this->final);
* @author Geoffrey T. Dairiki <dairiki@dairiki.org>
class Text_Diff_Op_change extends Text_Diff_Op {
function __construct( $orig, $final )
public function Text_Diff_Op_change( $orig, $final ) {
self::__construct( $orig, $final );
$reverse = new Text_Diff_Op_change($this->final, $this->orig);
It is recommended that you Edit text format, this type of Fix handles quite a lot in one request
