includes/utils/UIDGenerator.php

   1 <?php
   2 /**
   3  * This file deals with UID generation.
   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  */
  22 use Wikimedia\Assert\Assert;
  23 use MediaWiki\MediaWikiServices;
  24
  25 /**
  26  * Class for getting statistically unique IDs
  27  *
  28  * @since 1.21
  29  */
  30 class UIDGenerator {
  31         /** @var UIDGenerator */
  32         protected static $instance = null;
  33         /** @var string Local file path */
  34         protected $nodeIdFile;
  35         /** @var string Node ID in binary (32 bits) */
  36         protected $nodeId32;
  37         /** @var string Node ID in binary (48 bits) */
  38         protected $nodeId48;
  39
  40         /** @var string Local file path */
  41         protected $lockFile88;
  42         /** @var string Local file path */
  43         protected $lockFile128;
  44         /** @var string Local file path */
  45         protected $lockFileUUID;
  46
  47         /** @var array Cached file handles */
  48         protected $fileHandles = []; // cached file handles
  49
  50         const QUICK_RAND = 1; // get randomness from fast and insecure sources
  51         const QUICK_VOLATILE = 2; // use an APC like in-memory counter if available
  52
  53         protected function __construct() {
  54                 $this->nodeIdFile = wfTempDir() . '/mw-' . __CLASS__ . '-UID-nodeid';
  55                 $nodeId = '';
  56                 if ( is_file( $this->nodeIdFile ) ) {
  57                         $nodeId = file_get_contents( $this->nodeIdFile );
  58                 }
  59                 // Try to get some ID that uniquely identifies this machine (RFC 4122)...
  60                 if ( !preg_match( '/^[0-9a-f]{12}$/i', $nodeId ) ) {
  61                         Wikimedia\suppressWarnings();
  62                         if ( wfIsWindows() ) {
  63                                 // https://technet.microsoft.com/en-us/library/bb490913.aspx
  64                                 $csv = trim( wfShellExec( 'getmac /NH /FO CSV' ) );
  65                                 $line = substr( $csv, 0, strcspn( $csv, "\n" ) );
  66                                 $info = str_getcsv( $line );
  67                                 $nodeId = isset( $info[0] ) ? str_replace( '-', '', $info[0] ) : '';
  68                         } elseif ( is_executable( '/sbin/ifconfig' ) ) { // Linux/BSD/Solaris/OS X
  69                                 // See https://linux.die.net/man/8/ifconfig
  70                                 $m = [];
  71                                 preg_match( '/\s([0-9a-f]{2}(:[0-9a-f]{2}){5})\s/',
  72                                         wfShellExec( '/sbin/ifconfig -a' ), $m );
  73                                 $nodeId = isset( $m[1] ) ? str_replace( ':', '', $m[1] ) : '';
  74                         }
  75                         Wikimedia\restoreWarnings();
  76                         if ( !preg_match( '/^[0-9a-f]{12}$/i', $nodeId ) ) {
  77                                 $nodeId = MWCryptRand::generateHex( 12 );
  78                                 $nodeId[1] = dechex( hexdec( $nodeId[1] ) | 0x1 ); // set multicast bit
  79                         }
  80                         file_put_contents( $this->nodeIdFile, $nodeId ); // cache
  81                 }
  82                 $this->nodeId32 = Wikimedia\base_convert( substr( sha1( $nodeId ), 0, 8 ), 16, 2, 32 );
  83                 $this->nodeId48 = Wikimedia\base_convert( $nodeId, 16, 2, 48 );
  84                 // If different processes run as different users, they may have different temp dirs.
  85                 // This is dealt with by initializing the clock sequence number and counters randomly.
  86                 $this->lockFile88 = wfTempDir() . '/mw-' . __CLASS__ . '-UID-88';
  87                 $this->lockFile128 = wfTempDir() . '/mw-' . __CLASS__ . '-UID-128';
  88                 $this->lockFileUUID = wfTempDir() . '/mw-' . __CLASS__ . '-UUID-128';
  89         }
  90
  91         /**
  92          * @todo move to MW-specific factory class and inject temp dir
  93          * @return UIDGenerator
  94          */
  95         protected static function singleton() {
  96                 if ( self::$instance === null ) {
  97                         self::$instance = new self();
  98                 }
  99
 100                 return self::$instance;
 101         }
 102
 103         /**
 104          * Get a statistically unique 88-bit unsigned integer ID string.
 105          * The bits of the UID are prefixed with the time (down to the millisecond).
 106          *
 107          * These IDs are suitable as values for the shard key of distributed data.
 108          * If a column uses these as values, it should be declared UNIQUE to handle collisions.
 109          * New rows almost always have higher UIDs, which makes B-TREE updates on INSERT fast.
 110          * They can also be stored "DECIMAL(27) UNSIGNED" or BINARY(11) in MySQL.
 111          *
 112          * UID generation is serialized on each server (as the node ID is for the whole machine).
 113          *
 114          * @param int $base Specifies a base other than 10
 115          * @return string Number
 116          * @throws RuntimeException
 117          */
 118         public static function newTimestampedUID88( $base = 10 ) {
 119                 Assert::parameterType( 'integer', $base, '$base' );
 120                 Assert::parameter( $base <= 36, '$base', 'must be <= 36' );
 121                 Assert::parameter( $base >= 2, '$base', 'must be >= 2' );
 122
 123                 $gen = self::singleton();
 124                 $info = $gen->getTimeAndDelay( 'lockFile88', 1, 1024, 1024 );
 125                 $info['offsetCounter'] = $info['offsetCounter'] % 1024;
 126                 return Wikimedia\base_convert( $gen->getTimestampedID88( $info ), 2, $base );
 127         }
 128
 129         /**
 130          * @param array $info result of UIDGenerator::getTimeAndDelay(), or
 131          *  for sub classes, a seqencial array like (time, offsetCounter).
 132          * @return string 88 bits
 133          * @throws RuntimeException
 134          */
 135         protected function getTimestampedID88( array $info ) {
 136                 if ( isset( $info['time'] ) ) {
 137                         $time = $info['time'];
 138                         $counter = $info['offsetCounter'];
 139                 } else {
 140                         $time = $info[0];
 141                         $counter = $info[1];
 142                 }
 143                 // Take the 46 LSBs of "milliseconds since epoch"
 144                 $id_bin = $this->millisecondsSinceEpochBinary( $time );
 145                 // Add a 10 bit counter resulting in 56 bits total
 146                 $id_bin .= str_pad( decbin( $counter ), 10, '0', STR_PAD_LEFT );
 147                 // Add the 32 bit node ID resulting in 88 bits total
 148                 $id_bin .= $this->nodeId32;
 149                 // Convert to a 1-27 digit integer string
 150                 if ( strlen( $id_bin ) !== 88 ) {
 151                         throw new RuntimeException( "Detected overflow for millisecond timestamp." );
 152                 }
 153
 154                 return $id_bin;
 155         }
 156
 157         /**
 158          * Get a statistically unique 128-bit unsigned integer ID string.
 159          * The bits of the UID are prefixed with the time (down to the millisecond).
 160          *
 161          * These IDs are suitable as globally unique IDs, without any enforced uniqueness.
 162          * New rows almost always have higher UIDs, which makes B-TREE updates on INSERT fast.
 163          * They can also be stored as "DECIMAL(39) UNSIGNED" or BINARY(16) in MySQL.
 164          *
 165          * UID generation is serialized on each server (as the node ID is for the whole machine).
 166          *
 167          * @param int $base Specifies a base other than 10
 168          * @return string Number
 169          * @throws RuntimeException
 170          */
 171         public static function newTimestampedUID128( $base = 10 ) {
 172                 Assert::parameterType( 'integer', $base, '$base' );
 173                 Assert::parameter( $base <= 36, '$base', 'must be <= 36' );
 174                 Assert::parameter( $base >= 2, '$base', 'must be >= 2' );
 175
 176                 $gen = self::singleton();
 177                 $info = $gen->getTimeAndDelay( 'lockFile128', 16384, 1048576, 1048576 );
 178                 $info['offsetCounter'] = $info['offsetCounter'] % 1048576;
 179
 180                 return Wikimedia\base_convert( $gen->getTimestampedID128( $info ), 2, $base );
 181         }
 182
 183         /**
 184          * @param array $info The result of UIDGenerator::getTimeAndDelay(),
 185          *  for sub classes, a seqencial array like (time, offsetCounter, clkSeq).
 186          * @return string 128 bits
 187          * @throws RuntimeException
 188          */
 189         protected function getTimestampedID128( array $info ) {
 190                 if ( isset( $info['time'] ) ) {
 191                         $time = $info['time'];
 192                         $counter = $info['offsetCounter'];
 193                         $clkSeq = $info['clkSeq'];
 194                 } else {
 195                         $time = $info[0];
 196                         $counter = $info[1];
 197                         $clkSeq = $info[2];
 198                 }
 199                 // Take the 46 LSBs of "milliseconds since epoch"
 200                 $id_bin = $this->millisecondsSinceEpochBinary( $time );
 201                 // Add a 20 bit counter resulting in 66 bits total
 202                 $id_bin .= str_pad( decbin( $counter ), 20, '0', STR_PAD_LEFT );
 203                 // Add a 14 bit clock sequence number resulting in 80 bits total
 204                 $id_bin .= str_pad( decbin( $clkSeq ), 14, '0', STR_PAD_LEFT );
 205                 // Add the 48 bit node ID resulting in 128 bits total
 206                 $id_bin .= $this->nodeId48;
 207                 // Convert to a 1-39 digit integer string
 208                 if ( strlen( $id_bin ) !== 128 ) {
 209                         throw new RuntimeException( "Detected overflow for millisecond timestamp." );
 210                 }
 211
 212                 return $id_bin;
 213         }
 214
 215         /**
 216          * Return an RFC4122 compliant v1 UUID
 217          *
 218          * @return string
 219          * @throws RuntimeException
 220          * @since 1.27
 221          */
 222         public static function newUUIDv1() {
 223                 $gen = self::singleton();
 224                 // There can be up to 10000 intervals for the same millisecond timestamp.
 225                 // [0,4999] counter + [0,5000] offset is in [0,9999] for the offset counter.
 226                 // Add this onto the timestamp to allow making up to 5000 IDs per second.
 227                 return $gen->getUUIDv1( $gen->getTimeAndDelay( 'lockFileUUID', 16384, 5000, 5001 ) );
 228         }
 229
 230         /**
 231          * Return an RFC4122 compliant v1 UUID
 232          *
 233          * @return string 32 hex characters with no hyphens
 234          * @throws RuntimeException
 235          * @since 1.27
 236          */
 237         public static function newRawUUIDv1() {
 238                 return str_replace( '-', '', self::newUUIDv1() );
 239         }
 240
 241         /**
 242          * @param array $info Result of UIDGenerator::getTimeAndDelay()
 243          * @return string 128 bits
 244          */
 245         protected function getUUIDv1( array $info ) {
 246                 $clkSeq_bin = Wikimedia\base_convert( $info['clkSeq'], 10, 2, 14 );
 247                 $time_bin = $this->intervalsSinceGregorianBinary( $info['time'], $info['offsetCounter'] );
 248                 // Take the 32 bits of "time low"
 249                 $id_bin = substr( $time_bin, 28, 32 );
 250                 // Add 16 bits of "time mid" resulting in 48 bits total
 251                 $id_bin .= substr( $time_bin, 12, 16 );
 252                 // Add 4 bit version resulting in 52 bits total
 253                 $id_bin .= '0001';
 254                 // Add 12 bits of "time high" resulting in 64 bits total
 255                 $id_bin .= substr( $time_bin, 0, 12 );
 256                 // Add 2 bits of "variant" resulting in 66 bits total
 257                 $id_bin .= '10';
 258                 // Add 6 bits of "clock seq high" resulting in 72 bits total
 259                 $id_bin .= substr( $clkSeq_bin, 0, 6 );
 260                 // Add 8 bits of "clock seq low" resulting in 80 bits total
 261                 $id_bin .= substr( $clkSeq_bin, 6, 8 );
 262                 // Add the 48 bit node ID resulting in 128 bits total
 263                 $id_bin .= $this->nodeId48;
 264                 // Convert to a 32 char hex string with dashes
 265                 if ( strlen( $id_bin ) !== 128 ) {
 266                         throw new RuntimeException( "Detected overflow for millisecond timestamp." );
 267                 }
 268                 $hex = Wikimedia\base_convert( $id_bin, 2, 16, 32 );
 269                 return sprintf( '%s-%s-%s-%s-%s',
 270                         // "time_low" (32 bits)
 271                         substr( $hex, 0, 8 ),
 272                         // "time_mid" (16 bits)
 273                         substr( $hex, 8, 4 ),
 274                         // "time_hi_and_version" (16 bits)
 275                         substr( $hex, 12, 4 ),
 276                         // "clk_seq_hi_res" (8 bits) and "clk_seq_low" (8 bits)
 277                         substr( $hex, 16, 4 ),
 278                         // "node" (48 bits)
 279                         substr( $hex, 20, 12 )
 280                 );
 281         }
 282
 283         /**
 284          * Return an RFC4122 compliant v4 UUID
 285          *
 286          * @param int $flags Bitfield (supports UIDGenerator::QUICK_RAND)
 287          * @return string
 288          * @throws RuntimeException
 289          */
 290         public static function newUUIDv4( $flags = 0 ) {
 291                 $hex = ( $flags & self::QUICK_RAND )
 292                         ? wfRandomString( 31 )
 293                         : MWCryptRand::generateHex( 31 );
 294
 295                 return sprintf( '%s-%s-%s-%s-%s',
 296                         // "time_low" (32 bits)
 297                         substr( $hex, 0, 8 ),
 298                         // "time_mid" (16 bits)
 299                         substr( $hex, 8, 4 ),
 300                         // "time_hi_and_version" (16 bits)
 301                         '4' . substr( $hex, 12, 3 ),
 302                         // "clk_seq_hi_res" (8 bits, variant is binary 10x) and "clk_seq_low" (8 bits)
 303                         dechex( 0x8 | ( hexdec( $hex[15] ) & 0x3 ) ) . $hex[16] . substr( $hex, 17, 2 ),
 304                         // "node" (48 bits)
 305                         substr( $hex, 19, 12 )
 306                 );
 307         }
 308
 309         /**
 310          * Return an RFC4122 compliant v4 UUID
 311          *
 312          * @param int $flags Bitfield (supports UIDGenerator::QUICK_RAND)
 313          * @return string 32 hex characters with no hyphens
 314          * @throws RuntimeException
 315          */
 316         public static function newRawUUIDv4( $flags = 0 ) {
 317                 return str_replace( '-', '', self::newUUIDv4( $flags ) );
 318         }
 319
 320         /**
 321          * Return an ID that is sequential *only* for this node and bucket
 322          *
 323          * These IDs are suitable for per-host sequence numbers, e.g. for some packet protocols.
 324          * If UIDGenerator::QUICK_VOLATILE is used the counter might reset on server restart.
 325          *
 326          * @param string $bucket Arbitrary bucket name (should be ASCII)
 327          * @param int $bits Bit size (<=48) of resulting numbers before wrap-around
 328          * @param int $flags (supports UIDGenerator::QUICK_VOLATILE)
 329          * @return float Integer value as float
 330          * @since 1.23
 331          */
 332         public static function newSequentialPerNodeID( $bucket, $bits = 48, $flags = 0 ) {
 333                 return current( self::newSequentialPerNodeIDs( $bucket, $bits, 1, $flags ) );
 334         }
 335
 336         /**
 337          * Return IDs that are sequential *only* for this node and bucket
 338          *
 339          * @see UIDGenerator::newSequentialPerNodeID()
 340          * @param string $bucket Arbitrary bucket name (should be ASCII)
 341          * @param int $bits Bit size (16 to 48) of resulting numbers before wrap-around
 342          * @param int $count Number of IDs to return
 343          * @param int $flags (supports UIDGenerator::QUICK_VOLATILE)
 344          * @return array Ordered list of float integer values
 345          * @since 1.23
 346          */
 347         public static function newSequentialPerNodeIDs( $bucket, $bits, $count, $flags = 0 ) {
 348                 $gen = self::singleton();
 349                 return $gen->getSequentialPerNodeIDs( $bucket, $bits, $count, $flags );
 350         }
 351
 352         /**
 353          * Return IDs that are sequential *only* for this node and bucket
 354          *
 355          * @see UIDGenerator::newSequentialPerNodeID()
 356          * @param string $bucket Arbitrary bucket name (should be ASCII)
 357          * @param int $bits Bit size (16 to 48) of resulting numbers before wrap-around
 358          * @param int $count Number of IDs to return
 359          * @param int $flags (supports UIDGenerator::QUICK_VOLATILE)
 360          * @return array Ordered list of float integer values
 361          * @throws RuntimeException
 362          */
 363         protected function getSequentialPerNodeIDs( $bucket, $bits, $count, $flags ) {
 364                 if ( $count <= 0 ) {
 365                         return []; // nothing to do
 366                 }
 367                 if ( $bits < 16 || $bits > 48 ) {
 368                         throw new RuntimeException( "Requested bit size ($bits) is out of range." );
 369                 }
 370
 371                 $counter = null; // post-increment persistent counter value
 372
 373                 // Use APC/etc if requested, available, and not in CLI mode;
 374                 // Counter values would not survive across script instances in CLI mode.
 375                 $cache = null;
 376                 if ( ( $flags & self::QUICK_VOLATILE ) && !wfIsCLI() ) {
 377                         $cache = MediaWikiServices::getInstance()->getLocalServerObjectCache();
 378                 }
 379                 if ( $cache ) {
 380                         $counter = $cache->incrWithInit( $bucket, $cache::TTL_INDEFINITE, $count, $count );
 381                         if ( $counter === false ) {
 382                                 throw new RuntimeException( 'Unable to set value to ' . get_class( $cache ) );
 383                         }
 384                 }
 385
 386                 // Note: use of fmod() avoids "division by zero" on 32 bit machines
 387                 if ( $counter === null ) {
 388                         $path = wfTempDir() . '/mw-' . __CLASS__ . '-' . rawurlencode( $bucket ) . '-48';
 389                         // Get the UID lock file handle
 390                         if ( isset( $this->fileHandles[$path] ) ) {
 391                                 $handle = $this->fileHandles[$path];
 392                         } else {
 393                                 $handle = fopen( $path, 'cb+' );
 394                                 $this->fileHandles[$path] = $handle ?: null; // cache
 395                         }
 396                         // Acquire the UID lock file
 397                         if ( $handle === false ) {
 398                                 throw new RuntimeException( "Could not open '{$path}'." );
 399                         }
 400                         if ( !flock( $handle, LOCK_EX ) ) {
 401                                 fclose( $handle );
 402                                 throw new RuntimeException( "Could not acquire '{$path}'." );
 403                         }
 404                         // Fetch the counter value and increment it...
 405                         rewind( $handle );
 406                         $counter = floor( trim( fgets( $handle ) ) ) + $count; // fetch as float
 407                         // Write back the new counter value
 408                         ftruncate( $handle, 0 );
 409                         rewind( $handle );
 410                         fwrite( $handle, fmod( $counter, 2 ** 48 ) ); // warp-around as needed
 411                         fflush( $handle );
 412                         // Release the UID lock file
 413                         flock( $handle, LOCK_UN );
 414                 }
 415
 416                 $ids = [];
 417                 $divisor = 2 ** $bits;
 418                 $currentId = floor( $counter - $count ); // pre-increment counter value
 419                 for ( $i = 0; $i < $count; ++$i ) {
 420                         $ids[] = fmod( ++$currentId, $divisor );
 421                 }
 422
 423                 return $ids;
 424         }
 425
 426         /**
 427          * Get a (time,counter,clock sequence) where (time,counter) is higher
 428          * than any previous (time,counter) value for the given clock sequence.
 429          * This is useful for making UIDs sequential on a per-node bases.
 430          *
 431          * @param string $lockFile Name of a local lock file
 432          * @param int $clockSeqSize The number of possible clock sequence values
 433          * @param int $counterSize The number of possible counter values
 434          * @param int $offsetSize The number of possible offset values
 435          * @return array Array with the following keys:
 436          *  - array 'time': array of seconds int and milliseconds int.
 437          *  - int 'counter'.
 438          *  - int 'clkSeq'.
 439          *  - int 'offset': .
 440          *  - int 'offsetCounter'.
 441          * @throws RuntimeException
 442          */
 443         protected function getTimeAndDelay( $lockFile, $clockSeqSize, $counterSize, $offsetSize ) {
 444                 // Get the UID lock file handle
 445                 if ( isset( $this->fileHandles[$lockFile] ) ) {
 446                         $handle = $this->fileHandles[$lockFile];
 447                 } else {
 448                         $handle = fopen( $this->$lockFile, 'cb+' );
 449                         $this->fileHandles[$lockFile] = $handle ?: null; // cache
 450                 }
 451                 // Acquire the UID lock file
 452                 if ( $handle === false ) {
 453                         throw new RuntimeException( "Could not open '{$this->$lockFile}'." );
 454                 }
 455                 if ( !flock( $handle, LOCK_EX ) ) {
 456                         fclose( $handle );
 457                         throw new RuntimeException( "Could not acquire '{$this->$lockFile}'." );
 458                 }
 459
 460                 // The formatters that use this method expect a timestamp with millisecond
 461                 // precision and a counter upto a certain size. When more IDs than the counter
 462                 // size are generated during the same timestamp, an exception is thrown as we
 463                 // cannot increment further, because the formatted ID would not have enough
 464                 // bits to fit the counter.
 465                 //
 466                 // To orchestrate this between independant PHP processes on the same hosts,
 467                 // we must have a common sense of time so that we only have to maintain
 468                 // a single counter in a single lock file.
 469                 //
 470                 // Given that:
 471                 // * The system clock can be observed via time(), without milliseconds.
 472                 // * Some other clock can be observed via microtime(), which also offers
 473                 //   millisecond precision.
 474                 // * microtime() drifts in-process further and further away from the system
 475                 //   clock the longer the longer the process runs for.
 476                 //   For example, on 2018-10-03 an HHVM 3.18 JobQueue process at WMF,
 477                 //   that ran for 9 min 55 sec, drifted 7 seconds.
 478                 //   The drift is immediate for processes running while the system clock changes.
 479                 //   time() does not have this problem. See https://bugs.php.net/bug.php?id=42659.
 480                 //
 481                 // We have two choices:
 482                 //
 483                 // 1. Use microtime() with the following caveats:
 484                 //    - The last stored time may be in the future, or our current time
 485                 //    may be in the past, in which case we'll frequently enter the slow
 486                 //    timeWaitUntil() method to try and "sync" the current process with
 487                 //    the previous process. We mustn't block for long though, max 10ms?
 488                 //    - For any drift above 10ms, we pretend that the clock went backwards,
 489                 //    and treat it the same way as after an NTP sync, by incrementing clock
 490                 //    sequence instead. Given this rolls over automatically and silently
 491                 //    and is meant to be rare, this is essentially sacrifices a reasonable
 492                 //    guarantee of uniqueness.
 493                 //    - For long running processes (e.g. longer than a few seconds) the drift
 494                 //    can easily be more than 2 seconds. Because we only have a single lock
 495                 //    file and don't want to keep too many counters and deal with clearing
 496                 //    those, we fatal the user and refuse to make an ID. (T94522)
 497                 // 2. Use time() and expand the counter by 1000x and use the first digits
 498                 //    as if they are the millisecond fraction of the timestamp.
 499                 //    Known caveats or perf impact: None.
 500                 //
 501                 // We choose the latter.
 502                 $msecCounterSize = $counterSize * 1000;
 503
 504                 rewind( $handle );
 505                 // Format of lock file contents:
 506                 // "<clk seq> <sec> <msec counter> <rand offset>"
 507                 $data = explode( ' ', fgets( $handle ) );
 508
 509                 if ( count( $data ) === 4 ) {
 510                         // The UID lock file was already initialized
 511                         $clkSeq = (int)$data[0] % $clockSeqSize;
 512                         $prevSec = (int)$data[1];
 513                         // Counter for UIDs with the same timestamp,
 514                         $msecCounter = 0;
 515                         $randOffset = (int)$data[3] % $counterSize;
 516
 517                         // If the system clock moved backwards by an NTP sync,
 518                         // or if the last writer process had its clock drift ahead,
 519                         // Try to catch up if the gap is small, so that we can keep a single
 520                         // monotonic logic file.
 521                         $sec = $this->timeWaitUntil( $prevSec );
 522                         if ( $sec === false ) {
 523                                 // Gap is too big. Looks like the clock got moved back significantly.
 524                                 // Start a new clock sequence, and re-randomize the extra offset,
 525                                 // which is useful for UIDs that do not include the clock sequence number.
 526                                 $clkSeq = ( $clkSeq + 1 ) % $clockSeqSize;
 527                                 $sec = time();
 528                                 $randOffset = mt_rand( 0, $offsetSize - 1 );
 529                                 trigger_error( "Clock was set back; sequence number incremented." );
 530                         } elseif ( $sec === $prevSec ) {
 531                                 // Sanity check, only keep remainder if a previous writer wrote
 532                                 // something here that we don't accept.
 533                                 $msecCounter = (int)$data[2] % $msecCounterSize;
 534                                 // Bump the counter if the time has not changed yet
 535                                 if ( ++$msecCounter >= $msecCounterSize ) {
 536                                         // More IDs generated with the same time than counterSize can accomodate
 537                                         flock( $handle, LOCK_UN );
 538                                         throw new RuntimeException( "Counter overflow for timestamp value." );
 539                                 }
 540                         }
 541                 } else {
 542                         // Initialize UID lock file information
 543                         $clkSeq = mt_rand( 0, $clockSeqSize - 1 );
 544                         $sec = time();
 545                         $msecCounter = 0;
 546                         $randOffset = mt_rand( 0, $offsetSize - 1 );
 547                 }
 548
 549                 // Update and release the UID lock file
 550                 ftruncate( $handle, 0 );
 551                 rewind( $handle );
 552                 fwrite( $handle, "{$clkSeq} {$sec} {$msecCounter} {$randOffset}" );
 553                 fflush( $handle );
 554                 flock( $handle, LOCK_UN );
 555
 556                 // Split msecCounter back into msec and counter
 557                 $msec = (int)( $msecCounter / 1000 );
 558                 $counter = $msecCounter % 1000;
 559
 560                 return [
 561                         'time'          => [ $sec, $msec ],
 562                         'counter'       => $counter,
 563                         'clkSeq'        => $clkSeq,
 564                         'offset'        => $randOffset,
 565                         'offsetCounter' => $counter + $randOffset,
 566                 ];
 567         }
 568
 569         /**
 570          * Wait till the current timestamp reaches $time and return the current
 571          * timestamp. This returns false if it would have to wait more than 10ms.
 572          *
 573          * @param int $time Result of time()
 574          * @return int|bool Timestamp or false
 575          */
 576         protected function timeWaitUntil( $time ) {
 577                 $start = microtime( true );
 578                 do {
 579                         $ct = time();
 580                         // https://secure.php.net/manual/en/language.operators.comparison.php
 581                         if ( $ct >= $time ) {
 582                                 // current time is higher than or equal to than $time
 583                                 return $ct;
 584                         }
 585                 } while ( ( microtime( true ) - $start ) <= 0.010 ); // upto 10ms
 586
 587                 return false;
 588         }
 589
 590         /**
 591          * @param array $time Array of second and millisecond integers
 592          * @return string 46 LSBs of "milliseconds since epoch" in binary (rolls over in 4201)
 593          * @throws RuntimeException
 594          */
 595         protected function millisecondsSinceEpochBinary( array $time ) {
 596                 list( $sec, $msec ) = $time;
 597                 $ts = 1000 * $sec + $msec;
 598                 if ( $ts > 2 ** 52 ) {
 599                         throw new RuntimeException( __METHOD__ .
 600                                 ': sorry, this function doesn\'t work after the year 144680' );
 601                 }
 602
 603                 return substr( Wikimedia\base_convert( $ts, 10, 2, 46 ), -46 );
 604         }
 605
 606         /**
 607          * @param array $time Array of second and millisecond integers
 608          * @param int $delta Number of intervals to add on to the timestamp
 609          * @return string 60 bits of "100ns intervals since 15 October 1582" (rolls over in 3400)
 610          * @throws RuntimeException
 611          */
 612         protected function intervalsSinceGregorianBinary( array $time, $delta = 0 ) {
 613                 list( $sec, $msec ) = $time;
 614                 $offset = '122192928000000000';
 615                 if ( PHP_INT_SIZE >= 8 ) { // 64 bit integers
 616                         $ts = ( 1000 * $sec + $msec ) * 10000 + (int)$offset + $delta;
 617                         $id_bin = str_pad( decbin( $ts % ( 2 ** 60 ) ), 60, '0', STR_PAD_LEFT );
 618                 } elseif ( extension_loaded( 'gmp' ) ) {
 619                         $ts = gmp_add( gmp_mul( (string)$sec, '1000' ), (string)$msec ); // ms
 620                         $ts = gmp_add( gmp_mul( $ts, '10000' ), $offset ); // 100ns intervals
 621                         $ts = gmp_add( $ts, (string)$delta );
 622                         $ts = gmp_mod( $ts, gmp_pow( '2', '60' ) ); // wrap around
 623                         $id_bin = str_pad( gmp_strval( $ts, 2 ), 60, '0', STR_PAD_LEFT );
 624                 } elseif ( extension_loaded( 'bcmath' ) ) {
 625                         $ts = bcadd( bcmul( $sec, 1000 ), $msec ); // ms
 626                         $ts = bcadd( bcmul( $ts, 10000 ), $offset ); // 100ns intervals
 627                         $ts = bcadd( $ts, $delta );
 628                         $ts = bcmod( $ts, bcpow( 2, 60 ) ); // wrap around
 629                         $id_bin = Wikimedia\base_convert( $ts, 10, 2, 60 );
 630                 } else {
 631                         throw new RuntimeException( 'bcmath or gmp extension required for 32 bit machines.' );
 632                 }
 633                 return $id_bin;
 634         }
 635
 636         /**
 637          * Delete all cache files that have been created.
 638          *
 639          * This is a cleanup method primarily meant to be used from unit tests to
 640          * avoid poluting the local filesystem. If used outside of a unit test
 641          * environment it should be used with caution as it may destroy state saved
 642          * in the files.
 643          *
 644          * @see unitTestTearDown
 645          * @since 1.23
 646          * @codeCoverageIgnore
 647          */
 648         private function deleteCacheFiles() {
 649                 // T46850
 650                 foreach ( $this->fileHandles as $path => $handle ) {
 651                         if ( $handle !== null ) {
 652                                 fclose( $handle );
 653                         }
 654                         if ( is_file( $path ) ) {
 655                                 unlink( $path );
 656                         }
 657                         unset( $this->fileHandles[$path] );
 658                 }
 659                 if ( is_file( $this->nodeIdFile ) ) {
 660                         unlink( $this->nodeIdFile );
 661                 }
 662         }
 663
 664         /**
 665          * Cleanup resources when tearing down after a unit test.
 666          *
 667          * This is a cleanup method primarily meant to be used from unit tests to
 668          * avoid poluting the local filesystem. If used outside of a unit test
 669          * environment it should be used with caution as it may destroy state saved
 670          * in the files.
 671          *
 672          * @internal For use by unit tests
 673          * @see deleteCacheFiles
 674          * @since 1.23
 675          * @codeCoverageIgnore
 676          */
 677         public static function unitTestTearDown() {
 678                 // T46850
 679                 $gen = self::singleton();
 680                 $gen->deleteCacheFiles();
 681         }
 682
 683         function __destruct() {
 684                 array_map( 'fclose', array_filter( $this->fileHandles ) );
 685         }
 686 }