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         /** @var string One of the ENGINE_* constants. */
  55         private $engine = self::ENGINE_PHP;
  56
  57         /** @var string Path to an executable to be used as the diff engine. */
  58         private $externalEngine;
  59
  60         /**
  61          * Convenience helper to use getTextDiff without an instance.
  62          * @param string $oldText
  63          * @param string $newText
  64          * @return string
  65          */
  66         public static function diff( $oldText, $newText ) {
  67                 /** @var TextSlotDiffRenderer $slotDiffRenderer */
  68                 $slotDiffRenderer = ContentHandler::getForModelID( CONTENT_MODEL_TEXT )
  69                         ->getSlotDiffRenderer( RequestContext::getMain() );
  70                 return $slotDiffRenderer->getTextDiff( $oldText, $newText );
  71         }
  72
  73         public function setStatsdDataFactory( IBufferingStatsdDataFactory $statsdDataFactory ) {
  74                 $this->statsdDataFactory = $statsdDataFactory;
  75         }
  76
  77         public function setLanguage( Language $language ) {
  78                 $this->language = $language;
  79         }
  80
  81         /**
  82          * Set which diff engine to use.
  83          * @param string $type One of the ENGINE_* constants.
  84          * @param string|null $executable Path to an external exectable, only when type is ENGINE_EXTERNAL.
  85          */
  86         public function setEngine( $type, $executable = null ) {
  87                 $engines = [ self::ENGINE_PHP, self::ENGINE_WIKIDIFF2, self::ENGINE_EXTERNAL ];
  88                 Assert::parameter( in_array( $type, $engines, true ), '$type',
  89                         'must be one of the TextSlotDiffRenderer::ENGINE_* constants' );
  90                 if ( $type === self::ENGINE_EXTERNAL ) {
  91                         Assert::parameter( is_string( $executable ) && is_executable( $executable ), '$executable',
  92                                 'must be a path to a valid executable' );
  93                 } else {
  94                         Assert::parameter( is_null( $executable ), '$executable',
  95                                 'must not be set unless $type is ENGINE_EXTERNAL' );
  96                 }
  97                 $this->engine = $type;
  98                 $this->externalEngine = $executable;
  99         }
 100
 101         /** @inheritDoc */
 102         public function getDiff( Content $oldContent = null, Content $newContent = null ) {
 103                 $this->normalizeContents( $oldContent, $newContent, TextContent::class );
 104
 105                 $oldText = $oldContent->serialize();
 106                 $newText = $newContent->serialize();
 107
 108                 return $this->getTextDiff( $oldText, $newText );
 109         }
 110
 111         /**
 112          * Diff the text representations of two content objects (or just two pieces of text in general).
 113          * @param string $oldText
 114          * @param string $newText
 115          * @return string
 116          */
 117         public function getTextDiff( $oldText, $newText ) {
 118                 Assert::parameterType( 'string', $oldText, '$oldText' );
 119                 Assert::parameterType( 'string', $newText, '$newText' );
 120
 121                 $diff = function () use ( $oldText, $newText ) {
 122                         $time = microtime( true );
 123
 124                         $result = $this->getTextDiffInternal( $oldText, $newText );
 125
 126                         $time = intval( ( microtime( true ) - $time ) * 1000 );
 127                         if ( $this->statsdDataFactory ) {
 128                                 $this->statsdDataFactory->timing( 'diff_time', $time );
 129                         }
 130
 131                         // TODO reimplement this using T142313
 132                         /*
 133                         // Log requests slower than 99th percentile
 134                         if ( $time > 100 && $this->mOldPage && $this->mNewPage ) {
 135                                 wfDebugLog( 'diff',
 136                                         "$time ms diff: {$this->mOldid} -> {$this->mNewid} {$this->mNewPage}" );
 137                         }
 138                         */
 139
 140                         return $result;
 141                 };
 142
 143                 /**
 144                  * @param Status $status
 145                  * @throws FatalError
 146                  */
 147                 $error = function ( $status ) {
 148                         throw new FatalError( $status->getWikiText() );
 149                 };
 150
 151                 // Use PoolCounter if the diff looks like it can be expensive
 152                 if ( strlen( $oldText ) + strlen( $newText ) > 20000 ) {
 153                         $work = new PoolCounterWorkViaCallback( 'diff',
 154                                 md5( $oldText ) . md5( $newText ),
 155                                 [ 'doWork' => $diff, 'error' => $error ]
 156                         );
 157                         return $work->execute();
 158                 }
 159
 160                 return $diff();
 161         }
 162
 163         /**
 164          * Diff the text representations of two content objects (or just two pieces of text in general).
 165          * This does the actual diffing, getTextDiff() wraps it with logging and resource limiting.
 166          * @param string $oldText
 167          * @param string $newText
 168          * @return string
 169          * @throws Exception
 170          */
 171         protected function getTextDiffInternal( $oldText, $newText ) {
 172                 // TODO move most of this into three parallel implementations of a text diff generator
 173                 // class, choose which one to use via dependecy injection
 174
 175                 $oldText = str_replace( "\r\n", "\n", $oldText );
 176                 $newText = str_replace( "\r\n", "\n", $newText );
 177
 178                 // Better external diff engine, the 2 may some day be dropped
 179                 // This one does the escaping and segmenting itself
 180                 if ( $this->engine === self::ENGINE_WIKIDIFF2 ) {
 181                         $wikidiff2Version = phpversion( 'wikidiff2' );
 182                         if (
 183                                 $wikidiff2Version !== false &&
 184                                 version_compare( $wikidiff2Version, '1.5.0', '>=' ) &&
 185                                 version_compare( $wikidiff2Version, '1.8.0', '<' )
 186                         ) {
 187                                 $text = wikidiff2_do_diff(
 188                                         $oldText,
 189                                         $newText,
 190                                         2,
 191                                         0
 192                                 );
 193                         } else {
 194                                 // Don't pass the 4th parameter introduced in version 1.5.0 and removed in version 1.8.0
 195                                 $text = wikidiff2_do_diff(
 196                                         $oldText,
 197                                         $newText,
 198                                         2
 199                                 );
 200                         }
 201
 202                         return $text;
 203                 } elseif ( $this->engine === self::ENGINE_EXTERNAL ) {
 204                         # Diff via the shell
 205                         $tmpDir = wfTempDir();
 206                         $tempName1 = tempnam( $tmpDir, 'diff_' );
 207                         $tempName2 = tempnam( $tmpDir, 'diff_' );
 208
 209                         $tempFile1 = fopen( $tempName1, "w" );
 210                         if ( !$tempFile1 ) {
 211                                 return false;
 212                         }
 213                         $tempFile2 = fopen( $tempName2, "w" );
 214                         if ( !$tempFile2 ) {
 215                                 return false;
 216                         }
 217                         fwrite( $tempFile1, $oldText );
 218                         fwrite( $tempFile2, $newText );
 219                         fclose( $tempFile1 );
 220                         fclose( $tempFile2 );
 221                         $cmd = [ $this->externalEngine, $tempName1, $tempName2 ];
 222                         $result = Shell::command( $cmd )
 223                                 ->execute();
 224                         $exitCode = $result->getExitCode();
 225                         if ( $exitCode !== 0 ) {
 226                                 throw new Exception( "External diff command returned code {$exitCode}. Stderr: "
 227                                                                          . wfEscapeWikiText( $result->getStderr() )
 228                                 );
 229                         }
 230                         $difftext = $result->getStdout();
 231                         unlink( $tempName1 );
 232                         unlink( $tempName2 );
 233
 234                         return $difftext;
 235                 } elseif ( $this->engine === self::ENGINE_PHP ) {
 236                         if ( $this->language ) {
 237                                 $oldText = $this->language->segmentForDiff( $oldText );
 238                                 $newText = $this->language->segmentForDiff( $newText );
 239                         }
 240                         $ota = explode( "\n", $oldText );
 241                         $nta = explode( "\n", $newText );
 242                         $diffs = new Diff( $ota, $nta );
 243                         $formatter = new TableDiffFormatter();
 244                         $difftext = $formatter->format( $diffs );
 245                         if ( $this->language ) {
 246                                 $difftext = $this->language->unsegmentForDiff( $difftext );
 247                         }
 248
 249                         return $difftext;
 250                 }
 251                 throw new LogicException( 'Invalid engine: ' . $this->engine );
 252         }
 253
 254 }