includes/jobqueue/Job.php

   1 <?php
   2 /**
   3  * Job queue task base code.
   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  * @defgroup JobQueue JobQueue
  22  */
  23
  24 /**
  25  * Class to both describe a background job and handle jobs.
  26  * To push jobs onto queues, use JobQueueGroup::singleton()->push();
  27  *
  28  * @ingroup JobQueue
  29  */
  30 abstract class Job implements RunnableJob {
  31         /** @var string */
  32         public $command;
  33
  34         /** @var array Array of job parameters */
  35         public $params;
  36
  37         /** @var array Additional queue metadata */
  38         public $metadata = [];
  39
  40         /** @var Title */
  41         protected $title;
  42
  43         /** @var bool Expensive jobs may set this to true */
  44         protected $removeDuplicates = false;
  45
  46         /** @var string Text for error that occurred last */
  47         protected $error;
  48
  49         /** @var callable[] */
  50         protected $teardownCallbacks = [];
  51
  52         /** @var int Bitfield of JOB_* class constants */
  53         protected $executionFlags = 0;
  54
  55         /** @var int Job must not be wrapped in the usual explicit LBFactory transaction round */
  56         const JOB_NO_EXPLICIT_TRX_ROUND = 1;
  57
  58         /**
  59          * Create the appropriate object to handle a specific job
  60          *
  61          * @param string $command Job command
  62          * @param array|Title $params Job parameters
  63          * @throws InvalidArgumentException
  64          * @return Job
  65          */
  66         public static function factory( $command, $params = [] ) {
  67                 global $wgJobClasses;
  68
  69                 if ( $params instanceof Title ) {
  70                         // Backwards compatibility for old signature ($command, $title, $params)
  71                         $title = $params;
  72                         $params = func_num_args() >= 3 ? func_get_arg( 2 ) : [];
  73                 } elseif ( isset( $params['namespace'] ) && isset( $params['title'] ) ) {
  74                         // Handle job classes that take title as constructor parameter.
  75                         // If a newer classes like GenericParameterJob uses these parameters,
  76                         // then this happens in Job::__construct instead.
  77                         $title = Title::makeTitle( $params['namespace'], $params['title'] );
  78                 } else {
  79                         // Default title for job classes not implementing GenericParameterJob.
  80                         // This must be a valid title because it not directly passed to
  81                         // our Job constructor, but rather it's subclasses which may expect
  82                         // to be able to use it.
  83                         $title = Title::makeTitle( NS_SPECIAL, 'Blankpage' );
  84                 }
  85
  86                 if ( isset( $wgJobClasses[$command] ) ) {
  87                         $handler = $wgJobClasses[$command];
  88
  89                         if ( is_callable( $handler ) ) {
  90                                 $job = call_user_func( $handler, $title, $params );
  91                         } elseif ( class_exists( $handler ) ) {
  92                                 if ( is_subclass_of( $handler, GenericParameterJob::class ) ) {
  93                                         $job = new $handler( $params );
  94                                 } else {
  95                                         $job = new $handler( $title, $params );
  96                                 }
  97                         } else {
  98                                 $job = null;
  99                         }
 100
 101                         if ( $job instanceof Job ) {
 102                                 $job->command = $command;
 103
 104                                 return $job;
 105                         } else {
 106                                 throw new InvalidArgumentException(
 107                                         "Could not instantiate job '$command': bad spec!"
 108                                 );
 109                         }
 110                 }
 111
 112                 throw new InvalidArgumentException( "Invalid job command '{$command}'" );
 113         }
 114
 115         /**
 116          * @param string $command
 117          * @param array|Title|null $params
 118          */
 119         public function __construct( $command, $params = null ) {
 120                 if ( $params instanceof Title ) {
 121                         // Backwards compatibility for old signature ($command, $title, $params)
 122                         $title = $params;
 123                         $params = func_num_args() >= 3 ? func_get_arg( 2 ) : [];
 124                 } else {
 125                         // Newer jobs may choose to not have a top-level title (e.g. GenericParameterJob)
 126                         $title = null;
 127                 }
 128
 129                 if ( !is_array( $params ) ) {
 130                         throw new InvalidArgumentException( '$params must be an array' );
 131                 }
 132
 133                 if (
 134                         $title &&
 135                         !isset( $params['namespace'] ) &&
 136                         !isset( $params['title'] )
 137                 ) {
 138                         // When constructing this class for submitting to the queue,
 139                         // normalise the $title arg of old job classes as part of $params.
 140                         $params['namespace'] = $title->getNamespace();
 141                         $params['title'] = $title->getDBKey();
 142                 }
 143
 144                 $this->command = $command;
 145                 $this->params = $params + [ 'requestId' => WebRequest::getRequestId() ];
 146
 147                 if ( $this->title === null ) {
 148                         // Set this field for access via getTitle().
 149                         $this->title = ( isset( $params['namespace'] ) && isset( $params['title'] ) )
 150                                 ? Title::makeTitle( $params['namespace'], $params['title'] )
 151                                 // GenericParameterJob classes without namespace/title params
 152                                 // should not use getTitle(). Set an invalid title as placeholder.
 153                                 : Title::makeTitle( NS_SPECIAL, '' );
 154                 }
 155         }
 156
 157         /**
 158          * @param int $flag JOB_* class constant
 159          * @return bool
 160          * @since 1.31
 161          */
 162         public function hasExecutionFlag( $flag ) {
 163                 return ( $this->executionFlags & $flag ) === $flag;
 164         }
 165
 166         /**
 167          * @return string
 168          */
 169         public function getType() {
 170                 return $this->command;
 171         }
 172
 173         /**
 174          * @return Title
 175          */
 176         final public function getTitle() {
 177                 return $this->title;
 178         }
 179
 180         /**
 181          * @return array
 182          */
 183         public function getParams() {
 184                 return $this->params;
 185         }
 186
 187         /**
 188          * @param string|null $field Metadata field or null to get all the metadata
 189          * @return mixed|null Value; null if missing
 190          * @since 1.33
 191          */
 192         public function getMetadata( $field = null ) {
 193                 if ( $field === null ) {
 194                         return $this->metadata;
 195                 }
 196
 197                 return $this->metadata[$field] ?? null;
 198         }
 199
 200         /**
 201          * @param string $field Key name to set the value for
 202          * @param mixed $value The value to set the field for
 203          * @return mixed|null The prior field value; null if missing
 204          * @since 1.33
 205          */
 206         public function setMetadata( $field, $value ) {
 207                 $old = $this->getMetadata( $field );
 208                 if ( $value === null ) {
 209                         unset( $this->metadata[$field] );
 210                 } else {
 211                         $this->metadata[$field] = $value;
 212                 }
 213
 214                 return $old;
 215         }
 216
 217         /**
 218          * @return int|null UNIX timestamp to delay running this job until, otherwise null
 219          * @since 1.22
 220          */
 221         public function getReleaseTimestamp() {
 222                 return isset( $this->params['jobReleaseTimestamp'] )
 223                         ? wfTimestampOrNull( TS_UNIX, $this->params['jobReleaseTimestamp'] )
 224                         : null;
 225         }
 226
 227         /**
 228          * @return int|null UNIX timestamp of when the job was queued, or null
 229          * @since 1.26
 230          */
 231         public function getQueuedTimestamp() {
 232                 return isset( $this->metadata['timestamp'] )
 233                         ? wfTimestampOrNull( TS_UNIX, $this->metadata['timestamp'] )
 234                         : null;
 235         }
 236
 237         /**
 238          * @return string|null Id of the request that created this job. Follows
 239          *  jobs recursively, allowing to track the id of the request that started a
 240          *  job when jobs insert jobs which insert other jobs.
 241          * @since 1.27
 242          */
 243         public function getRequestId() {
 244                 return $this->params['requestId'] ?? null;
 245         }
 246
 247         /**
 248          * @return int|null UNIX timestamp of when the job was runnable, or null
 249          * @since 1.26
 250          */
 251         public function getReadyTimestamp() {
 252                 return $this->getReleaseTimestamp() ?: $this->getQueuedTimestamp();
 253         }
 254
 255         /**
 256          * Whether the queue should reject insertion of this job if a duplicate exists
 257          *
 258          * This can be used to avoid duplicated effort or combined with delayed jobs to
 259          * coalesce updates into larger batches. Claimed jobs are never treated as
 260          * duplicates of new jobs, and some queues may allow a few duplicates due to
 261          * network partitions and fail-over. Thus, additional locking is needed to
 262          * enforce mutual exclusion if this is really needed.
 263          *
 264          * @return bool
 265          */
 266         public function ignoreDuplicates() {
 267                 return $this->removeDuplicates;
 268         }
 269
 270         /**
 271          * @return bool Whether this job can be retried on failure by job runners
 272          * @since 1.21
 273          */
 274         public function allowRetries() {
 275                 return true;
 276         }
 277
 278         /**
 279          * @return int Number of actually "work items" handled in this job
 280          * @see $wgJobBackoffThrottling
 281          * @since 1.23
 282          */
 283         public function workItemCount() {
 284                 return 1;
 285         }
 286
 287         /**
 288          * Subclasses may need to override this to make duplication detection work.
 289          * The resulting map conveys everything that makes the job unique. This is
 290          * only checked if ignoreDuplicates() returns true, meaning that duplicate
 291          * jobs are supposed to be ignored.
 292          *
 293          * @return array Map of key/values
 294          * @since 1.21
 295          */
 296         public function getDeduplicationInfo() {
 297                 $info = [
 298                         'type' => $this->getType(),
 299                         'params' => $this->getParams()
 300                 ];
 301                 if ( is_array( $info['params'] ) ) {
 302                         // Identical jobs with different "root" jobs should count as duplicates
 303                         unset( $info['params']['rootJobSignature'] );
 304                         unset( $info['params']['rootJobTimestamp'] );
 305                         // Likewise for jobs with different delay times
 306                         unset( $info['params']['jobReleaseTimestamp'] );
 307                         // Identical jobs from different requests should count as duplicates
 308                         unset( $info['params']['requestId'] );
 309                         // Queues pack and hash this array, so normalize the order
 310                         ksort( $info['params'] );
 311                 }
 312
 313                 return $info;
 314         }
 315
 316         /**
 317          * Get "root job" parameters for a task
 318          *
 319          * This is used to no-op redundant jobs, including child jobs of jobs,
 320          * as long as the children inherit the root job parameters. When a job
 321          * with root job parameters and "rootJobIsSelf" set is pushed, the
 322          * deduplicateRootJob() method is automatically called on it. If the
 323          * root job is only virtual and not actually pushed (e.g. the sub-jobs
 324          * are inserted directly), then call deduplicateRootJob() directly.
 325          *
 326          * @see JobQueue::deduplicateRootJob()
 327          *
 328          * @param string $key A key that identifies the task
 329          * @return array Map of:
 330          *   - rootJobIsSelf    : true
 331          *   - rootJobSignature : hash (e.g. SHA1) that identifies the task
 332          *   - rootJobTimestamp : TS_MW timestamp of this instance of the task
 333          * @since 1.21
 334          */
 335         public static function newRootJobParams( $key ) {
 336                 return [
 337                         'rootJobIsSelf'    => true,
 338                         'rootJobSignature' => sha1( $key ),
 339                         'rootJobTimestamp' => wfTimestampNow()
 340                 ];
 341         }
 342
 343         /**
 344          * @see JobQueue::deduplicateRootJob()
 345          * @return array
 346          * @since 1.21
 347          */
 348         public function getRootJobParams() {
 349                 return [
 350                         'rootJobSignature' => $this->params['rootJobSignature'] ?? null,
 351                         'rootJobTimestamp' => $this->params['rootJobTimestamp'] ?? null
 352                 ];
 353         }
 354
 355         /**
 356          * @see JobQueue::deduplicateRootJob()
 357          * @return bool
 358          * @since 1.22
 359          */
 360         public function hasRootJobParams() {
 361                 return isset( $this->params['rootJobSignature'] )
 362                         && isset( $this->params['rootJobTimestamp'] );
 363         }
 364
 365         /**
 366          * @see JobQueue::deduplicateRootJob()
 367          * @return bool Whether this is job is a root job
 368          */
 369         public function isRootJob() {
 370                 return $this->hasRootJobParams() && !empty( $this->params['rootJobIsSelf'] );
 371         }
 372
 373         /**
 374          * @param callable $callback A function with one parameter, the success status, which will be
 375          *   false if the job failed or it succeeded but the DB changes could not be committed or
 376          *   any deferred updates threw an exception. (This parameter was added in 1.28.)
 377          * @since 1.27
 378          */
 379         protected function addTeardownCallback( $callback ) {
 380                 $this->teardownCallbacks[] = $callback;
 381         }
 382
 383         /**
 384          * Do any final cleanup after run(), deferred updates, and all DB commits happen
 385          * @param bool $status Whether the job, its deferred updates, and DB commit all succeeded
 386          * @since 1.27
 387          */
 388         public function teardown( $status ) {
 389                 foreach ( $this->teardownCallbacks as $callback ) {
 390                         call_user_func( $callback, $status );
 391                 }
 392         }
 393
 394         /**
 395          * @return string
 396          */
 397         public function toString() {
 398                 $paramString = '';
 399                 if ( $this->params ) {
 400                         foreach ( $this->params as $key => $value ) {
 401                                 if ( $paramString != '' ) {
 402                                         $paramString .= ' ';
 403                                 }
 404                                 if ( is_array( $value ) ) {
 405                                         $filteredValue = [];
 406                                         foreach ( $value as $k => $v ) {
 407                                                 $json = FormatJson::encode( $v );
 408                                                 if ( $json === false || mb_strlen( $json ) > 512 ) {
 409                                                         $filteredValue[$k] = gettype( $v ) . '(...)';
 410                                                 } else {
 411                                                         $filteredValue[$k] = $v;
 412                                                 }
 413                                         }
 414                                         if ( count( $filteredValue ) <= 10 ) {
 415                                                 $value = FormatJson::encode( $filteredValue );
 416                                         } else {
 417                                                 $value = "array(" . count( $value ) . ")";
 418                                         }
 419                                 } elseif ( is_object( $value ) && !method_exists( $value, '__toString' ) ) {
 420                                         $value = "object(" . get_class( $value ) . ")";
 421                                 }
 422
 423                                 $flatValue = (string)$value;
 424                                 if ( mb_strlen( $value ) > 1024 ) {
 425                                         $flatValue = "string(" . mb_strlen( $value ) . ")";
 426                                 }
 427
 428                                 $paramString .= "$key={$flatValue}";
 429                         }
 430                 }
 431
 432                 $metaString = '';
 433                 foreach ( $this->metadata as $key => $value ) {
 434                         if ( is_scalar( $value ) && mb_strlen( $value ) < 1024 ) {
 435                                 $metaString .= ( $metaString ? ",$key=$value" : "$key=$value" );
 436                         }
 437                 }
 438
 439                 $s = $this->command;
 440                 if ( is_object( $this->title ) ) {
 441                         $s .= " {$this->title->getPrefixedDBkey()}";
 442                 }
 443                 if ( $paramString != '' ) {
 444                         $s .= " $paramString";
 445                 }
 446                 if ( $metaString != '' ) {
 447                         $s .= " ($metaString)";
 448                 }
 449
 450                 return $s;
 451         }
 452
 453         protected function setLastError( $error ) {
 454                 $this->error = $error;
 455         }
 456
 457         public function getLastError() {
 458                 return $this->error;
 459         }
 460 }