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