maintenance/dumpTextPass.php

   1 <?php
   2 /**
   3  * Copyright (C) 2005 Brion Vibber <brion@pobox.com>
   4  * http://www.mediawiki.org/
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License as published by
   8  * the Free Software Foundation; either version 2 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * This program is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License along
  17  * with this program; if not, write to the Free Software Foundation, Inc.,
  18  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  19  * http://www.gnu.org/copyleft/gpl.html
  20  *
  21  * @file
  22  * @ingroup Maintenance
  23  */
  24
  25 $originalDir = getcwd();
  26
  27 require_once( dirname( __FILE__ ) . '/commandLine.inc' );
  28 require_once( 'backup.inc' );
  29
  30 /**
  31  * @ingroup Maintenance
  32  */
  33 class TextPassDumper extends BackupDumper {
  34         var $prefetch = null;
  35         var $input = "php://stdin";
  36         var $history = WikiExporter::FULL;
  37         var $fetchCount = 0;
  38         var $prefetchCount = 0;
  39
  40         var $failures = 0;
  41         var $maxFailures = 5;
  42         var $failedTextRetrievals = 0;
  43         var $maxConsecutiveFailedTextRetrievals = 200;
  44         var $failureTimeout = 5; // Seconds to sleep after db failure
  45
  46         var $php = "php";
  47         var $spawn = false;
  48         var $spawnProc = false;
  49         var $spawnWrite = false;
  50         var $spawnRead = false;
  51         var $spawnErr = false;
  52
  53         function dump() {
  54                 # This shouldn't happen if on console... ;)
  55                 header( 'Content-type: text/html; charset=UTF-8' );
  56
  57                 # Notice messages will foul up your XML output even if they're
  58                 # relatively harmless.
  59                 if ( ini_get( 'display_errors' ) )
  60                         ini_set( 'display_errors', 'stderr' );
  61
  62                 $this->initProgress( $this->history );
  63
  64                 $this->db = $this->backupDb();
  65
  66                 $this->egress = new ExportProgressFilter( $this->sink, $this );
  67
  68                 $input = fopen( $this->input, "rt" );
  69                 $result = $this->readDump( $input );
  70
  71                 if ( WikiError::isError( $result ) ) {
  72                         wfDie( $result->getMessage() );
  73                 }
  74
  75                 if ( $this->spawnProc ) {
  76                         $this->closeSpawn();
  77                 }
  78
  79                 $this->report( true );
  80         }
  81
  82         function processOption( $opt, $val, $param ) {
  83                 global $IP;
  84                 $url = $this->processFileOpt( $val, $param );
  85
  86                 switch( $opt ) {
  87                 case 'prefetch':
  88                         require_once "$IP/maintenance/backupPrefetch.inc";
  89                         $this->prefetch = new BaseDump( $url );
  90                         break;
  91                 case 'stub':
  92                         $this->input = $url;
  93                         break;
  94                 case 'current':
  95                         $this->history = WikiExporter::CURRENT;
  96                         break;
  97                 case 'full':
  98                         $this->history = WikiExporter::FULL;
  99                         break;
 100                 case 'spawn':
 101                         $this->spawn = true;
 102                         if ( $val ) {
 103                                 $this->php = $val;
 104                         }
 105                         break;
 106                 }
 107         }
 108
 109         function processFileOpt( $val, $param ) {
 110                 switch( $val ) {
 111                 case "file":
 112                         return $param;
 113                 case "gzip":
 114                         return "compress.zlib://$param";
 115                 case "bzip2":
 116                         return "compress.bzip2://$param";
 117                 case "7zip":
 118                         return "mediawiki.compress.7z://$param";
 119                 default:
 120                         return $val;
 121                 }
 122         }
 123
 124         /**
 125          * Overridden to include prefetch ratio if enabled.
 126          */
 127         function showReport() {
 128                 if ( !$this->prefetch ) {
 129                         return parent::showReport();
 130                 }
 131
 132                 if ( $this->reporting ) {
 133                         $delta = wfTime() - $this->startTime;
 134                         $now = wfTimestamp( TS_DB );
 135                         if ( $delta ) {
 136                                 $rate = $this->pageCount / $delta;
 137                                 $revrate = $this->revCount / $delta;
 138                                 $portion = $this->revCount / $this->maxCount;
 139                                 $eta = $this->startTime + $delta / $portion;
 140                                 $etats = wfTimestamp( TS_DB, intval( $eta ) );
 141                                 $fetchrate = 100.0 * $this->prefetchCount / $this->fetchCount;
 142                         } else {
 143                                 $rate = '-';
 144                                 $revrate = '-';
 145                                 $etats = '-';
 146                                 $fetchrate = '-';
 147                         }
 148                         $this->progress( sprintf( "%s: %s %d pages (%0.3f/sec), %d revs (%0.3f/sec), %0.1f%% prefetched, ETA %s [max %d]",
 149                                 $now, wfWikiID(), $this->pageCount, $rate, $this->revCount, $revrate, $fetchrate, $etats, $this->maxCount ) );
 150                 }
 151         }
 152
 153         function readDump( $input ) {
 154                 $this->buffer = "";
 155                 $this->openElement = false;
 156                 $this->atStart = true;
 157                 $this->state = "";
 158                 $this->lastName = "";
 159                 $this->thisPage = 0;
 160                 $this->thisRev = 0;
 161
 162                 $parser = xml_parser_create( "UTF-8" );
 163                 xml_parser_set_option( $parser, XML_OPTION_CASE_FOLDING, false );
 164
 165                 xml_set_element_handler( $parser, array( &$this, 'startElement' ), array( &$this, 'endElement' ) );
 166                 xml_set_character_data_handler( $parser, array( &$this, 'characterData' ) );
 167
 168                 $offset = 0; // for context extraction on error reporting
 169                 $bufferSize = 512 * 1024;
 170                 do {
 171                         $chunk = fread( $input, $bufferSize );
 172                         if ( !xml_parse( $parser, $chunk, feof( $input ) ) ) {
 173                                 wfDebug( "TextDumpPass::readDump encountered XML parsing error\n" );
 174                                 return new WikiXmlError( $parser, 'XML import parse failure', $chunk, $offset );
 175                         }
 176                         $offset += strlen( $chunk );
 177                 } while ( $chunk !== false && !feof( $input ) );
 178                 xml_parser_free( $parser );
 179
 180                 return true;
 181         }
 182
 183         function getText( $id ) {
 184                 $this->fetchCount++;
 185                 if ( isset( $this->prefetch ) ) {
 186                         $text = $this->prefetch->prefetch( $this->thisPage, $this->thisRev );
 187                         if ( $text === null ) {
 188                                 // Entry missing from prefetch dump
 189                         } elseif ( $text === "" ) {
 190                                 // Blank entries may indicate that the prior dump was broken.
 191                                 // To be safe, reload it.
 192                         } else {
 193                                 $dbr = wfGetDB( DB_SLAVE );
 194                                 $revID = intval($this->thisRev);
 195                                 $revLength = $dbr->selectField( 'revision', 'rev_len', array('rev_id' => $revID ) );
 196                                 // if length of rev text in file doesn't match length in db, we reload
 197                                 // this avoids carrying forward broken data from previous xml dumps
 198                                 if( strlen($text) == $revLength ) {
 199                                         $this->prefetchCount++;
 200                                         return $text;
 201                                 }
 202                         }
 203                 }
 204                 return $this->doGetText( $id );
 205         }
 206
 207         private function doGetText( $id ) {
 208
 209                 $id = intval( $id );
 210                 $this->failures = 0;
 211                 $ex = new MWException( "Graceful storage failure" );
 212                 while (true) {
 213                         if ( $this->spawn ) {
 214                                 if ($this->failures) {
 215                                         // we don't know why it failed, could be the child process
 216                                         // borked, could be db entry busted, could be db server out to lunch,
 217                                         // so cover all bases
 218                                         $this->closeSpawn();
 219                                         $this->openSpawn();
 220                                 }
 221                                 $text =  $this->getTextSpawned( $id );
 222                         } else {
 223                                 $text =  $this->getTextDbSafe( $id );
 224                         }
 225                         if ( $text === false ) {
 226                                 $this->failures++;
 227                                 if ( $this->failures > $this->maxFailures) {
 228                                         $this->progress( "Failed to retrieve revision text for text id ".
 229                                                                          "$id after $this->maxFailures tries, giving up" );
 230                                         // were there so many bad retrievals in a row we want to bail?
 231                                         // at some point we have to declare the dump irretrievably broken
 232                                         $this->failedTextRetrievals++;
 233                                         if ($this->failedTextRetrievals > $this->maxConsecutiveFailedTextRetrievals) {
 234                                                 throw $ex;
 235                                         }
 236                                         else {
 237                                                 // would be nice to return something better to the caller someday,
 238                                                 // log what we know about the failure and about the revision
 239                                                 return("");
 240                                         }
 241                                 } else {
 242                                         $this->progress( "Error $this->failures " .
 243                                                                  "of allowed $this->maxFailures retrieving revision text for text id $id! " .
 244                                                                  "Pausing $this->failureTimeout seconds before retry..." );
 245                                         sleep( $this->failureTimeout );
 246                                 }
 247                         } else {
 248                                 $this->failedTextRetrievals= 0;
 249                                 return( $text );
 250                         }
 251                 }
 252
 253         }
 254
 255         /**
 256          * Fetch a text revision from the database, retrying in case of failure.
 257          * This may survive some transitory errors by reconnecting, but
 258          * may not survive a long-term server outage.
 259          */
 260         private function getTextDbSafe( $id ) {
 261                 while ( true ) {
 262                         try {
 263                                 $text = $this->getTextDb( $id );
 264                                 $ex = new MWException( "Graceful storage failure" );
 265                         } catch ( DBQueryError $ex ) {
 266                                 $text = false;
 267                         }
 268                         return $text;
 269                 }
 270         }
 271
 272         /**
 273          * May throw a database error if, say, the server dies during query.
 274          */
 275         private function getTextDb( $id ) {
 276                 global $wgContLang;
 277                 $row = $this->db->selectRow( 'text',
 278                         array( 'old_text', 'old_flags' ),
 279                         array( 'old_id' => $id ),
 280                         'TextPassDumper::getText' );
 281                 $text = Revision::getRevisionText( $row );
 282                 if ( $text === false ) {
 283                         return false;
 284                 }
 285                 $stripped = str_replace( "\r", "", $text );
 286                 $normalized = $wgContLang->normalize( $stripped );
 287                 return $normalized;
 288         }
 289
 290         private function getTextSpawned( $id ) {
 291                 wfSuppressWarnings();
 292                 if ( !$this->spawnProc ) {
 293                         // First time?
 294                         $this->openSpawn();
 295                 }
 296                 $text = $this->getTextSpawnedOnce( $id );
 297                 wfRestoreWarnings();
 298                 return $text;
 299         }
 300
 301         function openSpawn() {
 302                 global $IP, $wgDBname;
 303
 304                 $cmd = implode( " ",
 305                         array_map( 'wfEscapeShellArg',
 306                                 array(
 307                                         $this->php,
 308                                         "$IP/maintenance/fetchText.php",
 309                                         $wgDBname ) ) );
 310                 $spec = array(
 311                         0 => array( "pipe", "r" ),
 312                         1 => array( "pipe", "w" ),
 313                         2 => array( "file", "/dev/null", "a" ) );
 314                 $pipes = array();
 315
 316                 $this->progress( "Spawning database subprocess: $cmd" );
 317                 $this->spawnProc = proc_open( $cmd, $spec, $pipes );
 318                 if ( !$this->spawnProc ) {
 319                         // shit
 320                         $this->progress( "Subprocess spawn failed." );
 321                         return false;
 322                 }
 323                 list(
 324                         $this->spawnWrite, // -> stdin
 325                         $this->spawnRead,  // <- stdout
 326                 ) = $pipes;
 327
 328                 return true;
 329         }
 330
 331         private function closeSpawn() {
 332                 wfSuppressWarnings();
 333                 if ( $this->spawnRead )
 334                         fclose( $this->spawnRead );
 335                 $this->spawnRead = false;
 336                 if ( $this->spawnWrite )
 337                         fclose( $this->spawnWrite );
 338                 $this->spawnWrite = false;
 339                 if ( $this->spawnErr )
 340                         fclose( $this->spawnErr );
 341                 $this->spawnErr = false;
 342                 if ( $this->spawnProc )
 343                         pclose( $this->spawnProc );
 344                 $this->spawnProc = false;
 345                 wfRestoreWarnings();
 346         }
 347
 348         private function getTextSpawnedOnce( $id ) {
 349                 global $wgContLang;
 350
 351                 $ok = fwrite( $this->spawnWrite, "$id\n" );
 352                 // $this->progress( ">> $id" );
 353                 if ( !$ok ) return false;
 354
 355                 $ok = fflush( $this->spawnWrite );
 356                 // $this->progress( ">> [flush]" );
 357                 if ( !$ok ) return false;
 358
 359                 // check that the text id they are sending is the one we asked for
 360                 // this avoids out of sync revision text errors we have encountered in the past
 361                 $newId = fgets( $this->spawnRead );
 362                 if ( $newId === false ) {
 363                         return false;
 364                 }
 365                 if ( $id != intval( $newId ) ) {
 366                         return false;
 367                 }
 368
 369                 $len = fgets( $this->spawnRead );
 370                 // $this->progress( "<< " . trim( $len ) );
 371                 if ( $len === false ) return false;
 372
 373                 $nbytes = intval( $len );
 374                 // actual error, not zero-length text
 375                 if ($nbytes < 0 ) return false;
 376
 377                 $text = "";
 378
 379                 // Subprocess may not send everything at once, we have to loop.
 380                 while ( $nbytes > strlen( $text ) ) {
 381                         $buffer = fread( $this->spawnRead, $nbytes - strlen( $text ) );
 382                         if ( $buffer === false ) break;
 383                         $text .= $buffer;
 384                 }
 385
 386                 $gotbytes = strlen( $text );
 387                 if ( $gotbytes != $nbytes ) {
 388                         $this->progress( "Expected $nbytes bytes from database subprocess, got $gotbytes " );
 389                         return false;
 390                 }
 391
 392                 // Do normalization in the dump thread...
 393                 $stripped = str_replace( "\r", "", $text );
 394                 $normalized = $wgContLang->normalize( $stripped );
 395                 return $normalized;
 396         }
 397
 398         function startElement( $parser, $name, $attribs ) {
 399                 $this->clearOpenElement( null );
 400                 $this->lastName = $name;
 401
 402                 if ( $name == 'revision' ) {
 403                         $this->state = $name;
 404                         $this->egress->writeOpenPage( null, $this->buffer );
 405                         $this->buffer = "";
 406                 } elseif ( $name == 'page' ) {
 407                         $this->state = $name;
 408                         if ( $this->atStart ) {
 409                                 $this->egress->writeOpenStream( $this->buffer );
 410                                 $this->buffer = "";
 411                                 $this->atStart = false;
 412                         }
 413                 }
 414
 415                 if ( $name == "text" && isset( $attribs['id'] ) ) {
 416                         $text = $this->getText( $attribs['id'] );
 417                         $this->openElement = array( $name, array( 'xml:space' => 'preserve' ) );
 418                         if ( strlen( $text ) > 0 ) {
 419                                 $this->characterData( $parser, $text );
 420                         }
 421                 } else {
 422                         $this->openElement = array( $name, $attribs );
 423                 }
 424         }
 425
 426         function endElement( $parser, $name ) {
 427                 if ( $this->openElement ) {
 428                         $this->clearOpenElement( "" );
 429                 } else {
 430                         $this->buffer .= "</$name>";
 431                 }
 432
 433                 if ( $name == 'revision' ) {
 434                         $this->egress->writeRevision( null, $this->buffer );
 435                         $this->buffer = "";
 436                         $this->thisRev = "";
 437                 } elseif ( $name == 'page' ) {
 438                         $this->egress->writeClosePage( $this->buffer );
 439                         $this->buffer = "";
 440                         $this->thisPage = "";
 441                 } elseif ( $name == 'mediawiki' ) {
 442                         $this->egress->writeCloseStream( $this->buffer );
 443                         $this->buffer = "";
 444                 }
 445         }
 446
 447         function characterData( $parser, $data ) {
 448                 $this->clearOpenElement( null );
 449                 if ( $this->lastName == "id" ) {
 450                         if ( $this->state == "revision" ) {
 451                                 $this->thisRev .= $data;
 452                         } elseif ( $this->state == "page" ) {
 453                                 $this->thisPage .= $data;
 454                         }
 455                 }
 456                 $this->buffer .= htmlspecialchars( $data );
 457         }
 458
 459         function clearOpenElement( $style ) {
 460                 if ( $this->openElement ) {
 461                         $this->buffer .= Xml::element( $this->openElement[0], $this->openElement[1], $style );
 462                         $this->openElement = false;
 463                 }
 464         }
 465 }
 466
 467
 468 $dumper = new TextPassDumper( $argv );
 469
 470 if ( true ) {
 471         $dumper->dump();
 472 } else {
 473         $dumper->progress( <<<ENDS
 474 This script postprocesses XML dumps from dumpBackup.php to add
 475 page text which was stubbed out (using --stub).
 476
 477 XML input is accepted on stdin.
 478 XML output is sent to stdout; progress reports are sent to stderr.
 479
 480 Usage: php dumpTextPass.php [<options>]
 481 Options:
 482   --stub=<type>:<file> To load a compressed stub dump instead of stdin
 483   --prefetch=<type>:<file> Use a prior dump file as a text source, to save
 484                           pressure on the database.
 485                           (Requires PHP 5.0+ and the XMLReader PECL extension)
 486   --quiet         Don't dump status reports to stderr.
 487   --report=n  Report position and speed after every n pages processed.
 488                           (Default: 100)
 489   --server=h  Force reading from MySQL server h
 490   --current       Base ETA on number of pages in database instead of all revisions
 491   --spawn         Spawn a subprocess for loading text records
 492 ENDS
 493 );
 494 }
 495
 496