includes/utils/CdbPHP.php

   1 <?php
   2 /**
   3  * This is a port of D.J. Bernstein's CDB to PHP. It's based on the copy that
   4  * appears in PHP 5.3. Changes are:
   5  *    * Error returns replaced with exceptions
   6  *    * Exception thrown if sizes or offsets are between 2GB and 4GB
   7  *    * Some variables renamed
   8  *
   9  * This program is free software; you can redistribute it and/or modify
  10  * it under the terms of the GNU General Public License as published by
  11  * the Free Software Foundation; either version 2 of the License, or
  12  * (at your option) any later version.
  13  *
  14  * This program is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  17  * GNU General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU General Public License along
  20  * with this program; if not, write to the Free Software Foundation, Inc.,
  21  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  22  * http://www.gnu.org/copyleft/gpl.html
  23  *
  24  * @file
  25  */
  26
  27 /**
  28  * Common functions for readers and writers
  29  */
  30 class CdbFunctions {
  31         /**
  32          * Take a modulo of a signed integer as if it were an unsigned integer.
  33          * $b must be less than 0x40000000 and greater than 0
  34          *
  35          * @param $a
  36          * @param $b
  37          *
  38          * @return int
  39          */
  40         public static function unsignedMod( $a, $b ) {
  41                 if ( $a & 0x80000000 ) {
  42                         $m = ( $a & 0x7fffffff ) % $b + 2 * ( 0x40000000 % $b );
  43
  44                         return $m % $b;
  45                 } else {
  46                         return $a % $b;
  47                 }
  48         }
  49
  50         /**
  51          * Shift a signed integer right as if it were unsigned
  52          * @param $a
  53          * @param $b
  54          * @return int
  55          */
  56         public static function unsignedShiftRight( $a, $b ) {
  57                 if ( $b == 0 ) {
  58                         return $a;
  59                 }
  60                 if ( $a & 0x80000000 ) {
  61                         return ( ( $a & 0x7fffffff ) >> $b ) | ( 0x40000000 >> ( $b - 1 ) );
  62                 } else {
  63                         return $a >> $b;
  64                 }
  65         }
  66
  67         /**
  68          * The CDB hash function.
  69          *
  70          * @param $s string
  71          *
  72          * @return
  73          */
  74         public static function hash( $s ) {
  75                 $h = 5381;
  76                 $len = strlen( $s );
  77                 for ( $i = 0; $i < $len; $i++ ) {
  78                         $h5 = ( $h << 5 ) & 0xffffffff;
  79                         // Do a 32-bit sum
  80                         // Inlined here for speed
  81                         $sum = ( $h & 0x3fffffff ) + ( $h5 & 0x3fffffff );
  82                         $h =
  83                                 (
  84                                         ( $sum & 0x40000000 ? 1 : 0 )
  85                                         + ( $h & 0x80000000 ? 2 : 0 )
  86                                         + ( $h & 0x40000000 ? 1 : 0 )
  87                                         + ( $h5 & 0x80000000 ? 2 : 0 )
  88                                         + ( $h5 & 0x40000000 ? 1 : 0 )
  89                                 ) << 30
  90                                 | ( $sum & 0x3fffffff );
  91                         $h ^= ord( $s[$i] );
  92                         $h &= 0xffffffff;
  93                 }
  94
  95                 return $h;
  96         }
  97 }
  98
  99 /**
 100  * CDB reader class
 101  */
 102 class CdbReaderPHP extends CdbReader {
 103         /** The filename */
 104         var $fileName;
 105
 106         /** The file handle */
 107         var $handle;
 108
 109         /* number of hash slots searched under this key */
 110         var $loop;
 111
 112         /* initialized if loop is nonzero */
 113         var $khash;
 114
 115         /* initialized if loop is nonzero */
 116         var $kpos;
 117
 118         /* initialized if loop is nonzero */
 119         var $hpos;
 120
 121         /* initialized if loop is nonzero */
 122         var $hslots;
 123
 124         /* initialized if findNext() returns true */
 125         var $dpos;
 126
 127         /* initialized if cdb_findnext() returns 1 */
 128         var $dlen;
 129
 130         /**
 131          * @param $fileName string
 132          * @throws MWException
 133          */
 134         function __construct( $fileName ) {
 135                 $this->fileName = $fileName;
 136                 $this->handle = fopen( $fileName, 'rb' );
 137                 if ( !$this->handle ) {
 138                         throw new MWException( 'Unable to open CDB file "' . $this->fileName . '".' );
 139                 }
 140                 $this->findStart();
 141         }
 142
 143         function close() {
 144                 if ( isset( $this->handle ) ) {
 145                         fclose( $this->handle );
 146                 }
 147                 unset( $this->handle );
 148         }
 149
 150         /**
 151          * @param $key
 152          * @return bool|string
 153          */
 154         public function get( $key ) {
 155                 // strval is required
 156                 if ( $this->find( strval( $key ) ) ) {
 157                         return $this->read( $this->dlen, $this->dpos );
 158                 } else {
 159                         return false;
 160                 }
 161         }
 162
 163         /**
 164          * @param $key
 165          * @param $pos
 166          * @return bool
 167          */
 168         protected function match( $key, $pos ) {
 169                 $buf = $this->read( strlen( $key ), $pos );
 170
 171                 return $buf === $key;
 172         }
 173
 174         protected function findStart() {
 175                 $this->loop = 0;
 176         }
 177
 178         /**
 179          * @throws MWException
 180          * @param $length
 181          * @param $pos
 182          * @return string
 183          */
 184         protected function read( $length, $pos ) {
 185                 if ( fseek( $this->handle, $pos ) == -1 ) {
 186                         // This can easily happen if the internal pointers are incorrect
 187                         throw new MWException(
 188                                 'Seek failed, file "' . $this->fileName . '" may be corrupted.' );
 189                 }
 190
 191                 if ( $length == 0 ) {
 192                         return '';
 193                 }
 194
 195                 $buf = fread( $this->handle, $length );
 196                 if ( $buf === false || strlen( $buf ) !== $length ) {
 197                         throw new MWException(
 198                                 'Read from CDB file failed, file "' . $this->fileName . '" may be corrupted.' );
 199                 }
 200
 201                 return $buf;
 202         }
 203
 204         /**
 205          * Unpack an unsigned integer and throw an exception if it needs more than 31 bits
 206          * @param $s
 207          * @throws MWException
 208          * @return mixed
 209          */
 210         protected function unpack31( $s ) {
 211                 $data = unpack( 'V', $s );
 212                 if ( $data[1] > 0x7fffffff ) {
 213                         throw new MWException(
 214                                 'Error in CDB file "' . $this->fileName . '", integer too big.' );
 215                 }
 216
 217                 return $data[1];
 218         }
 219
 220         /**
 221          * Unpack a 32-bit signed integer
 222          * @param $s
 223          * @return int
 224          */
 225         protected function unpackSigned( $s ) {
 226                 $data = unpack( 'va/vb', $s );
 227
 228                 return $data['a'] | ( $data['b'] << 16 );
 229         }
 230
 231         /**
 232          * @param $key
 233          * @return bool
 234          */
 235         protected function findNext( $key ) {
 236                 if ( !$this->loop ) {
 237                         $u = CdbFunctions::hash( $key );
 238                         $buf = $this->read( 8, ( $u << 3 ) & 2047 );
 239                         $this->hslots = $this->unpack31( substr( $buf, 4 ) );
 240                         if ( !$this->hslots ) {
 241                                 return false;
 242                         }
 243                         $this->hpos = $this->unpack31( substr( $buf, 0, 4 ) );
 244                         $this->khash = $u;
 245                         $u = CdbFunctions::unsignedShiftRight( $u, 8 );
 246                         $u = CdbFunctions::unsignedMod( $u, $this->hslots );
 247                         $u <<= 3;
 248                         $this->kpos = $this->hpos + $u;
 249                 }
 250
 251                 while ( $this->loop < $this->hslots ) {
 252                         $buf = $this->read( 8, $this->kpos );
 253                         $pos = $this->unpack31( substr( $buf, 4 ) );
 254                         if ( !$pos ) {
 255                                 return false;
 256                         }
 257                         $this->loop += 1;
 258                         $this->kpos += 8;
 259                         if ( $this->kpos == $this->hpos + ( $this->hslots << 3 ) ) {
 260                                 $this->kpos = $this->hpos;
 261                         }
 262                         $u = $this->unpackSigned( substr( $buf, 0, 4 ) );
 263                         if ( $u === $this->khash ) {
 264                                 $buf = $this->read( 8, $pos );
 265                                 $keyLen = $this->unpack31( substr( $buf, 0, 4 ) );
 266                                 if ( $keyLen == strlen( $key ) && $this->match( $key, $pos + 8 ) ) {
 267                                         // Found
 268                                         $this->dlen = $this->unpack31( substr( $buf, 4 ) );
 269                                         $this->dpos = $pos + 8 + $keyLen;
 270
 271                                         return true;
 272                                 }
 273                         }
 274                 }
 275
 276                 return false;
 277         }
 278
 279         /**
 280          * @param $key
 281          * @return bool
 282          */
 283         protected function find( $key ) {
 284                 $this->findStart();
 285
 286                 return $this->findNext( $key );
 287         }
 288 }
 289
 290 /**
 291  * CDB writer class
 292  */
 293 class CdbWriterPHP extends CdbWriter {
 294         var $handle, $realFileName, $tmpFileName;
 295
 296         var $hplist;
 297         var $numentries, $pos;
 298
 299         /**
 300          * @param $fileName string
 301          */
 302         function __construct( $fileName ) {
 303                 $this->realFileName = $fileName;
 304                 $this->tmpFileName = $fileName . '.tmp.' . mt_rand( 0, 0x7fffffff );
 305                 $this->handle = fopen( $this->tmpFileName, 'wb' );
 306                 if ( !$this->handle ) {
 307                         $this->throwException(
 308                                 'Unable to open CDB file "' . $this->tmpFileName . '" for write.' );
 309                 }
 310                 $this->hplist = array();
 311                 $this->numentries = 0;
 312                 $this->pos = 2048; // leaving space for the pointer array, 256 * 8
 313                 if ( fseek( $this->handle, $this->pos ) == -1 ) {
 314                         $this->throwException( 'fseek failed in file "' . $this->tmpFileName . '".' );
 315                 }
 316         }
 317
 318         function __destruct() {
 319                 if ( isset( $this->handle ) ) {
 320                         $this->close();
 321                 }
 322         }
 323
 324         /**
 325          * @param $key
 326          * @param $value
 327          * @return
 328          */
 329         public function set( $key, $value ) {
 330                 if ( strval( $key ) === '' ) {
 331                         // DBA cross-check hack
 332                         return;
 333                 }
 334                 $this->addbegin( strlen( $key ), strlen( $value ) );
 335                 $this->write( $key );
 336                 $this->write( $value );
 337                 $this->addend( strlen( $key ), strlen( $value ), CdbFunctions::hash( $key ) );
 338         }
 339
 340         /**
 341          * @throws MWException
 342          */
 343         public function close() {
 344                 $this->finish();
 345                 if ( isset( $this->handle ) ) {
 346                         fclose( $this->handle );
 347                 }
 348                 if ( wfIsWindows() && file_exists( $this->realFileName ) ) {
 349                         unlink( $this->realFileName );
 350                 }
 351                 if ( !rename( $this->tmpFileName, $this->realFileName ) ) {
 352                         $this->throwException( 'Unable to move the new CDB file into place.' );
 353                 }
 354                 unset( $this->handle );
 355         }
 356
 357         /**
 358          * @throws MWException
 359          * @param $buf
 360          */
 361         protected function write( $buf ) {
 362                 $len = fwrite( $this->handle, $buf );
 363                 if ( $len !== strlen( $buf ) ) {
 364                         $this->throwException( 'Error writing to CDB file "' . $this->tmpFileName . '".' );
 365                 }
 366         }
 367
 368         /**
 369          * @throws MWException
 370          * @param $len
 371          */
 372         protected function posplus( $len ) {
 373                 $newpos = $this->pos + $len;
 374                 if ( $newpos > 0x7fffffff ) {
 375                         $this->throwException(
 376                                 'A value in the CDB file "' . $this->tmpFileName . '" is too large.' );
 377                 }
 378                 $this->pos = $newpos;
 379         }
 380
 381         /**
 382          * @param $keylen
 383          * @param $datalen
 384          * @param $h
 385          */
 386         protected function addend( $keylen, $datalen, $h ) {
 387                 $this->hplist[] = array(
 388                         'h' => $h,
 389                         'p' => $this->pos
 390                 );
 391
 392                 $this->numentries++;
 393                 $this->posplus( 8 );
 394                 $this->posplus( $keylen );
 395                 $this->posplus( $datalen );
 396         }
 397
 398         /**
 399          * @throws MWException
 400          * @param $keylen
 401          * @param $datalen
 402          */
 403         protected function addbegin( $keylen, $datalen ) {
 404                 if ( $keylen > 0x7fffffff ) {
 405                         $this->throwException( 'Key length too long in file "' . $this->tmpFileName . '".' );
 406                 }
 407                 if ( $datalen > 0x7fffffff ) {
 408                         $this->throwException( 'Data length too long in file "' . $this->tmpFileName . '".' );
 409                 }
 410                 $buf = pack( 'VV', $keylen, $datalen );
 411                 $this->write( $buf );
 412         }
 413
 414         /**
 415          * @throws MWException
 416          */
 417         protected function finish() {
 418                 // Hack for DBA cross-check
 419                 $this->hplist = array_reverse( $this->hplist );
 420
 421                 // Calculate the number of items that will be in each hashtable
 422                 $counts = array_fill( 0, 256, 0 );
 423                 foreach ( $this->hplist as $item ) {
 424                         ++$counts[255 & $item['h']];
 425                 }
 426
 427                 // Fill in $starts with the *end* indexes
 428                 $starts = array();
 429                 $pos = 0;
 430                 for ( $i = 0; $i < 256; ++$i ) {
 431                         $pos += $counts[$i];
 432                         $starts[$i] = $pos;
 433                 }
 434
 435                 // Excessively clever and indulgent code to simultaneously fill $packedTables
 436                 // with the packed hashtables, and adjust the elements of $starts
 437                 // to actually point to the starts instead of the ends.
 438                 $packedTables = array_fill( 0, $this->numentries, false );
 439                 foreach ( $this->hplist as $item ) {
 440                         $packedTables[--$starts[255 & $item['h']]] = $item;
 441                 }
 442
 443                 $final = '';
 444                 for ( $i = 0; $i < 256; ++$i ) {
 445                         $count = $counts[$i];
 446
 447                         // The size of the hashtable will be double the item count.
 448                         // The rest of the slots will be empty.
 449                         $len = $count + $count;
 450                         $final .= pack( 'VV', $this->pos, $len );
 451
 452                         $hashtable = array();
 453                         for ( $u = 0; $u < $len; ++$u ) {
 454                                 $hashtable[$u] = array( 'h' => 0, 'p' => 0 );
 455                         }
 456
 457                         // Fill the hashtable, using the next empty slot if the hashed slot
 458                         // is taken.
 459                         for ( $u = 0; $u < $count; ++$u ) {
 460                                 $hp = $packedTables[$starts[$i] + $u];
 461                                 $where = CdbFunctions::unsignedMod(
 462                                         CdbFunctions::unsignedShiftRight( $hp['h'], 8 ), $len );
 463                                 while ( $hashtable[$where]['p'] ) {
 464                                         if ( ++$where == $len ) {
 465                                                 $where = 0;
 466                                         }
 467                                 }
 468                                 $hashtable[$where] = $hp;
 469                         }
 470
 471                         // Write the hashtable
 472                         for ( $u = 0; $u < $len; ++$u ) {
 473                                 $buf = pack( 'vvV',
 474                                         $hashtable[$u]['h'] & 0xffff,
 475                                         CdbFunctions::unsignedShiftRight( $hashtable[$u]['h'], 16 ),
 476                                         $hashtable[$u]['p'] );
 477                                 $this->write( $buf );
 478                                 $this->posplus( 8 );
 479                         }
 480                 }
 481
 482                 // Write the pointer array at the start of the file
 483                 rewind( $this->handle );
 484                 if ( ftell( $this->handle ) != 0 ) {
 485                         $this->throwException( 'Error rewinding to start of file "' . $this->tmpFileName . '".' );
 486                 }
 487                 $this->write( $final );
 488         }
 489
 490         /**
 491          * Clean up the temp file and throw an exception
 492          *
 493          * @param $msg string
 494          * @throws MWException
 495          */
 496         protected function throwException( $msg ) {
 497                 if ( $this->handle ) {
 498                         fclose( $this->handle );
 499                         unlink( $this->tmpFileName );
 500                 }
 501                 throw new MWException( $msg );
 502         }
 503 }