includes/diff/TextSlotDiffRenderer.php

   1 <?php
   2 /**
   3  * Renders a slot diff by doing a text diff on the native representation.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  * @ingroup DifferenceEngine
  22  */
  23
  24 use MediaWiki\Shell\Shell;
  25 use Wikimedia\Assert\Assert;
  26
  27 /**
  28  * Renders a slot diff by doing a text diff on the native representation.
  29  *
  30  * If you want to use this without content objects (to call getTextDiff() on some
  31  * non-content-related texts), obtain an instance with
  32  *     ContentHandler::getForModelID( CONTENT_MODEL_TEXT )
  33  *         ->getSlotDiffRenderer( RequestContext::getMain() )
  34  *
  35  * @ingroup DifferenceEngine
  36  */
  37 class TextSlotDiffRenderer extends SlotDiffRenderer {
  38
  39         /** Use the PHP diff implementation (DiffEngine). */
  40         const ENGINE_PHP = 'php';
  41
  42         /** Use the wikidiff2 PHP module. */
  43         const ENGINE_WIKIDIFF2 = 'wikidiff2';
  44
  45         /** Use an external executable. */
  46         const ENGINE_EXTERNAL = 'external';
  47
  48         /** @var IBufferingStatsdDataFactory|null */
  49         private $statsdDataFactory;
  50
  51         /** @var Language|null The language this content is in. */
  52         private $language;
  53
  54         /**
  55          * Number of paragraph moves the algorithm should attempt to detect.
  56          * Only used with the wikidiff2 engine.
  57          * @var int
  58          * @see $wgWikiDiff2MovedParagraphDetectionCutoff
  59          */
  60         private $wikiDiff2MovedParagraphDetectionCutoff = 0;
  61
  62         /** @var string One of the ENGINE_* constants. */
  63         private $engine = self::ENGINE_PHP;
  64
  65         /** @var string Path to an executable to be used as the diff engine. */
  66         private $externalEngine;
  67
  68         /**
  69          * Convenience helper to use getTextDiff without an instance.
  70          * @param string $oldText
  71          * @param string $newText
  72          * @return string
  73          */
  74         public static function diff( $oldText, $newText ) {
  75                 /** @var $slotDiffRenderer TextSlotDiffRenderer */
  76                 $slotDiffRenderer = ContentHandler::getForModelID( CONTENT_MODEL_TEXT )
  77                         ->getSlotDiffRenderer( RequestContext::getMain() );
  78                 return $slotDiffRenderer->getTextDiff( $oldText, $newText );
  79         }
  80
  81         public function setStatsdDataFactory( IBufferingStatsdDataFactory $statsdDataFactory ) {
  82                 $this->statsdDataFactory = $statsdDataFactory;
  83         }
  84
  85         public function setLanguage( Language $language ) {
  86                 $this->language = $language;
  87         }
  88         /**
  89          * @param int $cutoff
  90          * @see $wgWikiDiff2MovedParagraphDetectionCutoff
  91          */
  92         public function setWikiDiff2MovedParagraphDetectionCutoff( $cutoff ) {
  93                 Assert::parameterType( 'integer', $cutoff, '$cutoff' );
  94                 $this->wikiDiff2MovedParagraphDetectionCutoff = $cutoff;
  95         }
  96
  97         /**
  98          * Set which diff engine to use.
  99          * @param string $type One of the ENGINE_* constants.
 100          * @param string|null $executable Path to an external exectable, only when type is ENGINE_EXTERNAL.
 101          */
 102         public function setEngine( $type, $executable = null ) {
 103                 $engines = [ self::ENGINE_PHP, self::ENGINE_WIKIDIFF2, self::ENGINE_EXTERNAL ];
 104                 Assert::parameter( in_array( $type, $engines, true ), '$type',
 105                         'must be one of the TextSlotDiffRenderer::ENGINE_* constants' );
 106                 if ( $type === self::ENGINE_EXTERNAL ) {
 107                         Assert::parameter( is_string( $executable ) && is_executable( $executable ), '$executable',
 108                                 'must be a path to a valid executable' );
 109                 } else {
 110                         Assert::parameter( is_null( $executable ), '$executable',
 111                                 'must not be set unless $type is ENGINE_EXTERNAL' );
 112                 }
 113                 $this->engine = $type;
 114                 $this->externalEngine = $executable;
 115         }
 116
 117         /** @inheritDoc */
 118         public function getDiff( Content $oldContent = null, Content $newContent = null ) {
 119                 $this->normalizeContents( $oldContent, $newContent, TextContent::class );
 120
 121                 $oldText = $oldContent->serialize();
 122                 $newText = $newContent->serialize();
 123
 124                 return $this->getTextDiff( $oldText, $newText );
 125         }
 126
 127         /**
 128          * Diff the text representations of two content objects (or just two pieces of text in general).
 129          * @param string $oldText
 130          * @param string $newText
 131          * @return string
 132          */
 133         public function getTextDiff( $oldText, $newText ) {
 134                 Assert::parameterType( 'string', $oldText, '$oldText' );
 135                 Assert::parameterType( 'string', $newText, '$newText' );
 136
 137                 $diff = function () use ( $oldText, $newText ) {
 138                         $time = microtime( true );
 139
 140                         $result = $this->getTextDiffInternal( $oldText, $newText );
 141
 142                         $time = intval( ( microtime( true ) - $time ) * 1000 );
 143                         if ( $this->statsdDataFactory ) {
 144                                 $this->statsdDataFactory->timing( 'diff_time', $time );
 145                         }
 146
 147                         // TODO reimplement this using T142313
 148                         /*
 149                         // Log requests slower than 99th percentile
 150                         if ( $time > 100 && $this->mOldPage && $this->mNewPage ) {
 151                                 wfDebugLog( 'diff',
 152                                         "$time ms diff: {$this->mOldid} -> {$this->mNewid} {$this->mNewPage}" );
 153                         }
 154                         */
 155
 156                         return $result;
 157                 };
 158
 159                 /**
 160                  * @param Status $status
 161                  * @throws FatalError
 162                  */
 163                 $error = function ( $status ) {
 164                         throw new FatalError( $status->getWikiText() );
 165                 };
 166
 167                 // Use PoolCounter if the diff looks like it can be expensive
 168                 if ( strlen( $oldText ) + strlen( $newText ) > 20000 ) {
 169                         $work = new PoolCounterWorkViaCallback( 'diff',
 170                                 md5( $oldText ) . md5( $newText ),
 171                                 [ 'doWork' => $diff, 'error' => $error ]
 172                         );
 173                         return $work->execute();
 174                 }
 175
 176                 return $diff();
 177         }
 178
 179         /**
 180          * Diff the text representations of two content objects (or just two pieces of text in general).
 181          * This does the actual diffing, getTextDiff() wraps it with logging and resource limiting.
 182          * @param string $oldText
 183          * @param string $newText
 184          * @return string
 185          * @throws Exception
 186          */
 187         protected function getTextDiffInternal( $oldText, $newText ) {
 188                 // TODO move most of this into three parallel implementations of a text diff generator
 189                 // class, choose which one to use via dependecy injection
 190
 191                 $oldText = str_replace( "\r\n", "\n", $oldText );
 192                 $newText = str_replace( "\r\n", "\n", $newText );
 193
 194                 // Better external diff engine, the 2 may some day be dropped
 195                 // This one does the escaping and segmenting itself
 196                 if ( $this->engine === self::ENGINE_WIKIDIFF2 ) {
 197                         $wikidiff2Version = phpversion( 'wikidiff2' );
 198                         if (
 199                                 $wikidiff2Version !== false &&
 200                                 version_compare( $wikidiff2Version, '1.5.0', '>=' ) &&
 201                                 version_compare( $wikidiff2Version, '1.8.0', '<' )
 202                         ) {
 203                                 $text = wikidiff2_do_diff(
 204                                         $oldText,
 205                                         $newText,
 206                                         2,
 207                                         $this->wikiDiff2MovedParagraphDetectionCutoff
 208                                 );
 209                         } else {
 210                                 // Don't pass the 4th parameter introduced in version 1.5.0 and removed in version 1.8.0
 211                                 $text = wikidiff2_do_diff(
 212                                         $oldText,
 213                                         $newText,
 214                                         2
 215                                 );
 216
 217                                 // Log a warning in case the configuration value is set to not silently ignore it
 218                                 if ( $this->wikiDiff2MovedParagraphDetectionCutoff > 0 ) {
 219                                         wfLogWarning( '$wgWikiDiff2MovedParagraphDetectionCutoff is set but has no
 220                                                 effect since the used version of WikiDiff2 does not support it.' );
 221                                 }
 222                         }
 223
 224                         return $text;
 225                 } elseif ( $this->engine === self::ENGINE_EXTERNAL ) {
 226                         # Diff via the shell
 227                         $tmpDir = wfTempDir();
 228                         $tempName1 = tempnam( $tmpDir, 'diff_' );
 229                         $tempName2 = tempnam( $tmpDir, 'diff_' );
 230
 231                         $tempFile1 = fopen( $tempName1, "w" );
 232                         if ( !$tempFile1 ) {
 233                                 return false;
 234                         }
 235                         $tempFile2 = fopen( $tempName2, "w" );
 236                         if ( !$tempFile2 ) {
 237                                 return false;
 238                         }
 239                         fwrite( $tempFile1, $oldText );
 240                         fwrite( $tempFile2, $newText );
 241                         fclose( $tempFile1 );
 242                         fclose( $tempFile2 );
 243                         $cmd = [ $this->externalEngine, $tempName1, $tempName2 ];
 244                         $result = Shell::command( $cmd )
 245                                 ->execute();
 246                         $exitCode = $result->getExitCode();
 247                         if ( $exitCode !== 0 ) {
 248                                 throw new Exception( "External diff command returned code {$exitCode}. Stderr: "
 249                                                                          . wfEscapeWikiText( $result->getStderr() )
 250                                 );
 251                         }
 252                         $difftext = $result->getStdout();
 253                         unlink( $tempName1 );
 254                         unlink( $tempName2 );
 255
 256                         return $difftext;
 257                 } elseif ( $this->engine === self::ENGINE_PHP ) {
 258                         if ( $this->language ) {
 259                                 $oldText = $this->language->segmentForDiff( $oldText );
 260                                 $newText = $this->language->segmentForDiff( $newText );
 261                         }
 262                         $ota = explode( "\n", $oldText );
 263                         $nta = explode( "\n", $newText );
 264                         $diffs = new Diff( $ota, $nta );
 265                         $formatter = new TableDiffFormatter();
 266                         $difftext = $formatter->format( $diffs );
 267                         if ( $this->language ) {
 268                                 $difftext = $this->language->unsegmentForDiff( $difftext );
 269                         }
 270
 271                         return $difftext;
 272                 }
 273                 throw new LogicException( 'Invalid engine: ' . $this->engine );
 274         }
 275
 276 }