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  * The queue aspects of this class are now deprecated.
  27  * Using the class to push jobs onto queues is deprecated (use JobSpecification).
  28  *
  29  * @ingroup JobQueue
  30  */
  31 abstract class Job implements IJobSpecification {
  32         /** @var string */
  33         public $command;
  34
  35         /** @var array Array of job parameters */
  36         public $params;
  37
  38         /** @var array Additional queue metadata */
  39         public $metadata = [];
  40
  41         /** @var Title */
  42         protected $title;
  43
  44         /** @var bool Expensive jobs may set this to true */
  45         protected $removeDuplicates;
  46
  47         /** @var string Text for error that occurred last */
  48         protected $error;
  49
  50         /** @var callable[] */
  51         protected $teardownCallbacks = [];
  52
  53         /**
  54          * Run the job
  55          * @return bool Success
  56          */
  57         abstract public function run();
  58
  59         /**
  60          * Create the appropriate object to handle a specific job
  61          *
  62          * @param string $command Job command
  63          * @param Title $title Associated title
  64          * @param array $params Job parameters
  65          * @throws MWException
  66          * @return Job
  67          */
  68         public static function factory( $command, Title $title, $params = [] ) {
  69                 global $wgJobClasses;
  70
  71                 if ( isset( $wgJobClasses[$command] ) ) {
  72                         $class = $wgJobClasses[$command];
  73
  74                         $job = new $class( $title, $params );
  75                         $job->command = $command;
  76
  77                         return $job;
  78                 }
  79
  80                 throw new InvalidArgumentException( "Invalid job command '{$command}'" );
  81         }
  82
  83         /**
  84          * @param string $command
  85          * @param Title $title
  86          * @param array|bool $params Can not be === true
  87          */
  88         public function __construct( $command, $title, $params = false ) {
  89                 $this->command = $command;
  90                 $this->title = $title;
  91                 $this->params = is_array( $params ) ? $params : []; // sanity
  92
  93                 // expensive jobs may set this to true
  94                 $this->removeDuplicates = false;
  95
  96                 if ( !isset( $this->params['requestId'] ) ) {
  97                         $this->params['requestId'] = WebRequest::getRequestId();
  98                 }
  99         }
 100
 101         /**
 102          * Batch-insert a group of jobs into the queue.
 103          * This will be wrapped in a transaction with a forced commit.
 104          *
 105          * This may add duplicate at insert time, but they will be
 106          * removed later on, when the first one is popped.
 107          *
 108          * @param Job[] $jobs Array of Job objects
 109          * @return bool
 110          * @deprecated since 1.21
 111          */
 112         public static function batchInsert( $jobs ) {
 113                 wfDeprecated( __METHOD__, '1.21' );
 114                 JobQueueGroup::singleton()->push( $jobs );
 115                 return true;
 116         }
 117
 118         /**
 119          * @return string
 120          */
 121         public function getType() {
 122                 return $this->command;
 123         }
 124
 125         /**
 126          * @return Title
 127          */
 128         public function getTitle() {
 129                 return $this->title;
 130         }
 131
 132         /**
 133          * @return array
 134          */
 135         public function getParams() {
 136                 return $this->params;
 137         }
 138
 139         /**
 140          * @return int|null UNIX timestamp to delay running this job until, otherwise null
 141          * @since 1.22
 142          */
 143         public function getReleaseTimestamp() {
 144                 return isset( $this->params['jobReleaseTimestamp'] )
 145                         ? wfTimestampOrNull( TS_UNIX, $this->params['jobReleaseTimestamp'] )
 146                         : null;
 147         }
 148
 149         /**
 150          * @return int|null UNIX timestamp of when the job was queued, or null
 151          * @since 1.26
 152          */
 153         public function getQueuedTimestamp() {
 154                 return isset( $this->metadata['timestamp'] )
 155                         ? wfTimestampOrNull( TS_UNIX, $this->metadata['timestamp'] )
 156                         : null;
 157         }
 158
 159         /**
 160          * @return string|null Id of the request that created this job. Follows
 161          *  jobs recursively, allowing to track the id of the request that started a
 162          *  job when jobs insert jobs which insert other jobs.
 163          * @since 1.27
 164          */
 165         public function getRequestId() {
 166                 return isset( $this->params['requestId'] )
 167                         ? $this->params['requestId']
 168                         : null;
 169         }
 170
 171         /**
 172          * @return int|null UNIX timestamp of when the job was runnable, or null
 173          * @since 1.26
 174          */
 175         public function getReadyTimestamp() {
 176                 return $this->getReleaseTimestamp() ?: $this->getQueuedTimestamp();
 177         }
 178
 179         /**
 180          * Whether the queue should reject insertion of this job if a duplicate exists
 181          *
 182          * This can be used to avoid duplicated effort or combined with delayed jobs to
 183          * coalesce updates into larger batches. Claimed jobs are never treated as
 184          * duplicates of new jobs, and some queues may allow a few duplicates due to
 185          * network partitions and fail-over. Thus, additional locking is needed to
 186          * enforce mutual exclusion if this is really needed.
 187          *
 188          * @return bool
 189          */
 190         public function ignoreDuplicates() {
 191                 return $this->removeDuplicates;
 192         }
 193
 194         /**
 195          * @return bool Whether this job can be retried on failure by job runners
 196          * @since 1.21
 197          */
 198         public function allowRetries() {
 199                 return true;
 200         }
 201
 202         /**
 203          * @return int Number of actually "work items" handled in this job
 204          * @see $wgJobBackoffThrottling
 205          * @since 1.23
 206          */
 207         public function workItemCount() {
 208                 return 1;
 209         }
 210
 211         /**
 212          * Subclasses may need to override this to make duplication detection work.
 213          * The resulting map conveys everything that makes the job unique. This is
 214          * only checked if ignoreDuplicates() returns true, meaning that duplicate
 215          * jobs are supposed to be ignored.
 216          *
 217          * @return array Map of key/values
 218          * @since 1.21
 219          */
 220         public function getDeduplicationInfo() {
 221                 $info = [
 222                         'type' => $this->getType(),
 223                         'namespace' => $this->getTitle()->getNamespace(),
 224                         'title' => $this->getTitle()->getDBkey(),
 225                         'params' => $this->getParams()
 226                 ];
 227                 if ( is_array( $info['params'] ) ) {
 228                         // Identical jobs with different "root" jobs should count as duplicates
 229                         unset( $info['params']['rootJobSignature'] );
 230                         unset( $info['params']['rootJobTimestamp'] );
 231                         // Likewise for jobs with different delay times
 232                         unset( $info['params']['jobReleaseTimestamp'] );
 233                         // Identical jobs from different requests should count as duplicates
 234                         unset( $info['params']['requestId'] );
 235                         // Queues pack and hash this array, so normalize the order
 236                         ksort( $info['params'] );
 237                 }
 238
 239                 return $info;
 240         }
 241
 242         /**
 243          * Get "root job" parameters for a task
 244          *
 245          * This is used to no-op redundant jobs, including child jobs of jobs,
 246          * as long as the children inherit the root job parameters. When a job
 247          * with root job parameters and "rootJobIsSelf" set is pushed, the
 248          * deduplicateRootJob() method is automatically called on it. If the
 249          * root job is only virtual and not actually pushed (e.g. the sub-jobs
 250          * are inserted directly), then call deduplicateRootJob() directly.
 251          *
 252          * @see JobQueue::deduplicateRootJob()
 253          *
 254          * @param string $key A key that identifies the task
 255          * @return array Map of:
 256          *   - rootJobIsSelf    : true
 257          *   - rootJobSignature : hash (e.g. SHA1) that identifies the task
 258          *   - rootJobTimestamp : TS_MW timestamp of this instance of the task
 259          * @since 1.21
 260          */
 261         public static function newRootJobParams( $key ) {
 262                 return [
 263                         'rootJobIsSelf'    => true,
 264                         'rootJobSignature' => sha1( $key ),
 265                         'rootJobTimestamp' => wfTimestampNow()
 266                 ];
 267         }
 268
 269         /**
 270          * @see JobQueue::deduplicateRootJob()
 271          * @return array
 272          * @since 1.21
 273          */
 274         public function getRootJobParams() {
 275                 return [
 276                         'rootJobSignature' => isset( $this->params['rootJobSignature'] )
 277                                 ? $this->params['rootJobSignature']
 278                                 : null,
 279                         'rootJobTimestamp' => isset( $this->params['rootJobTimestamp'] )
 280                                 ? $this->params['rootJobTimestamp']
 281                                 : null
 282                 ];
 283         }
 284
 285         /**
 286          * @see JobQueue::deduplicateRootJob()
 287          * @return bool
 288          * @since 1.22
 289          */
 290         public function hasRootJobParams() {
 291                 return isset( $this->params['rootJobSignature'] )
 292                         && isset( $this->params['rootJobTimestamp'] );
 293         }
 294
 295         /**
 296          * @see JobQueue::deduplicateRootJob()
 297          * @return bool Whether this is job is a root job
 298          */
 299         public function isRootJob() {
 300                 return $this->hasRootJobParams() && !empty( $this->params['rootJobIsSelf'] );
 301         }
 302
 303         /**
 304          * @param callable $callback A function with one parameter, the success status, which will be
 305          *   false if the job failed or it succeeded but the DB changes could not be committed or
 306          *   any deferred updates threw an exception. (This parameter was added in 1.28.)
 307          * @since 1.27
 308          */
 309         protected function addTeardownCallback( $callback ) {
 310                 $this->teardownCallbacks[] = $callback;
 311         }
 312
 313         /**
 314          * Do any final cleanup after run(), deferred updates, and all DB commits happen
 315          * @param bool $status Whether the job, its deferred updates, and DB commit all succeeded
 316          * @since 1.27
 317          */
 318         public function teardown( $status ) {
 319                 foreach ( $this->teardownCallbacks as $callback ) {
 320                         call_user_func( $callback, $status );
 321                 }
 322         }
 323
 324         /**
 325          * Insert a single job into the queue.
 326          * @return bool True on success
 327          * @deprecated since 1.21
 328          */
 329         public function insert() {
 330                 JobQueueGroup::singleton()->push( $this );
 331                 return true;
 332         }
 333
 334         /**
 335          * @return string
 336          */
 337         public function toString() {
 338                 $paramString = '';
 339                 if ( $this->params ) {
 340                         foreach ( $this->params as $key => $value ) {
 341                                 if ( $paramString != '' ) {
 342                                         $paramString .= ' ';
 343                                 }
 344                                 if ( is_array( $value ) ) {
 345                                         $filteredValue = [];
 346                                         foreach ( $value as $k => $v ) {
 347                                                 $json = FormatJson::encode( $v );
 348                                                 if ( $json === false || mb_strlen( $json ) > 512 ) {
 349                                                         $filteredValue[$k] = gettype( $v ) . '(...)';
 350                                                 } else {
 351                                                         $filteredValue[$k] = $v;
 352                                                 }
 353                                         }
 354                                         if ( count( $filteredValue ) <= 10 ) {
 355                                                 $value = FormatJson::encode( $filteredValue );
 356                                         } else {
 357                                                 $value = "array(" . count( $value ) . ")";
 358                                         }
 359                                 } elseif ( is_object( $value ) && !method_exists( $value, '__toString' ) ) {
 360                                         $value = "object(" . get_class( $value ) . ")";
 361                                 }
 362
 363                                 $flatValue = (string)$value;
 364                                 if ( mb_strlen( $value ) > 1024 ) {
 365                                         $flatValue = "string(" . mb_strlen( $value ) . ")";
 366                                 }
 367
 368                                 $paramString .= "$key={$flatValue}";
 369                         }
 370                 }
 371
 372                 $metaString = '';
 373                 foreach ( $this->metadata as $key => $value ) {
 374                         if ( is_scalar( $value ) && mb_strlen( $value ) < 1024 ) {
 375                                 $metaString .= ( $metaString ? ",$key=$value" : "$key=$value" );
 376                         }
 377                 }
 378
 379                 $s = $this->command;
 380                 if ( is_object( $this->title ) ) {
 381                         $s .= " {$this->title->getPrefixedDBkey()}";
 382                 }
 383                 if ( $paramString != '' ) {
 384                         $s .= " $paramString";
 385                 }
 386                 if ( $metaString != '' ) {
 387                         $s .= " ($metaString)";
 388                 }
 389
 390                 return $s;
 391         }
 392
 393         protected function setLastError( $error ) {
 394                 $this->error = $error;
 395         }
 396
 397         public function getLastError() {
 398                 return $this->error;
 399         }
 400 }