Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
15.89% |
41 / 258 |
|
14.29% |
5 / 35 |
CRAP | |
0.00% |
0 / 1 |
| Module | |
15.62% |
40 / 256 |
|
14.29% |
5 / 35 |
5288.26 | |
0.00% |
0 / 1 |
| name | n/a |
0 / 0 |
n/a |
0 / 0 |
0 | |||||
| id_field | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| table_name | n/a |
0 / 0 |
n/a |
0 / 0 |
1 | |||||
| table | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| full_sync_action_name | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| get_object_by_id | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| init_listeners | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| init_full_sync_listeners | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| init_before_send | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| set_defaults | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| reset_data | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| enqueue_full_sync_actions | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| estimate_full_sync_actions | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| get_full_sync_actions | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| count_actions | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| get_check_sum | |
0.00% |
0 / 8 |
|
0.00% |
0 / 1 |
12 | |||
| recursive_ksort | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
12 | |||
| still_valid_checksum | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
12 | |||
| enqueue_all_ids_as_action | |
0.00% |
0 / 24 |
|
0.00% |
0 / 1 |
42 | |||
| get_next_chunk | |
0.00% |
0 / 10 |
|
0.00% |
0 / 1 |
2 | |||
| get_last_item | |
0.00% |
0 / 8 |
|
0.00% |
0 / 1 |
2 | |||
| get_initial_last_sent | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| send_full_sync_actions | |
0.00% |
0 / 36 |
|
0.00% |
0 / 1 |
210 | |||
| adjust_chunk_size_if_stuck | |
0.00% |
0 / 35 |
|
0.00% |
0 / 1 |
182 | |||
| set_send_full_sync_actions_status | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
2 | |||
| send_action | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
| get_chunks_with_preceding_end | |
0.00% |
0 / 8 |
|
0.00% |
0 / 1 |
6 | |||
| get_metadata | |
0.00% |
0 / 13 |
|
0.00% |
0 / 1 |
6 | |||
| init_listeners_for_meta_type | |
100.00% |
3 / 3 |
|
100.00% |
1 / 1 |
1 | |||
| init_meta_whitelist_handler | |
100.00% |
3 / 3 |
|
100.00% |
1 / 1 |
1 | |||
| get_term_relationships | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| unserialize_meta | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
| get_objects_by_id | |
0.00% |
0 / 8 |
|
0.00% |
0 / 1 |
30 | |||
| get_min_max_object_ids_for_batches | |
0.00% |
0 / 34 |
|
0.00% |
0 / 1 |
42 | |||
| total | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
2 | |||
| get_where_sql | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| filter_objects_and_metadata_by_size | |
96.88% |
31 / 32 |
|
0.00% |
0 / 1 |
11 | |||
| 1 | <?php |
| 2 | /** |
| 3 | * A base abstraction of a sync module. |
| 4 | * |
| 5 | * @package automattic/jetpack-sync |
| 6 | */ |
| 7 | |
| 8 | namespace Automattic\Jetpack\Sync\Modules; |
| 9 | |
| 10 | use Automattic\Jetpack\Sync\Defaults; |
| 11 | use Automattic\Jetpack\Sync\Functions; |
| 12 | use Automattic\Jetpack\Sync\Listener; |
| 13 | use Automattic\Jetpack\Sync\Replicastore; |
| 14 | use Automattic\Jetpack\Sync\Sender; |
| 15 | use Automattic\Jetpack\Sync\Settings; |
| 16 | |
| 17 | if ( ! defined( 'ABSPATH' ) ) { |
| 18 | exit( 0 ); |
| 19 | } |
| 20 | |
| 21 | /** |
| 22 | * Basic methods implemented by Jetpack Sync extensions. |
| 23 | * |
| 24 | * @abstract |
| 25 | */ |
| 26 | abstract class Module { |
| 27 | /** |
| 28 | * Number of items per chunk when grouping objects for performance reasons. |
| 29 | * |
| 30 | * @access public |
| 31 | * |
| 32 | * @var int |
| 33 | */ |
| 34 | const ARRAY_CHUNK_SIZE = 10; |
| 35 | |
| 36 | /** |
| 37 | * Max query length for DB queries. |
| 38 | * |
| 39 | * @access public |
| 40 | * |
| 41 | * @var int |
| 42 | */ |
| 43 | const MAX_DB_QUERY_LENGTH = 15 * 1024; |
| 44 | |
| 45 | /** |
| 46 | * Max bytes allowed for full sync upload for the module. |
| 47 | * Default Setting : 7MB. |
| 48 | * |
| 49 | * @access public |
| 50 | * |
| 51 | * @var int |
| 52 | */ |
| 53 | const MAX_SIZE_FULL_SYNC = 7000000; |
| 54 | |
| 55 | /** |
| 56 | * Max bytes allowed for post meta_value => length. |
| 57 | * Default Setting : 2MB. |
| 58 | * |
| 59 | * @access public |
| 60 | * |
| 61 | * @var int |
| 62 | */ |
| 63 | const MAX_META_LENGTH = 2000000; |
| 64 | |
| 65 | /** |
| 66 | * Sync module name. |
| 67 | * |
| 68 | * @access public |
| 69 | * |
| 70 | * @return string |
| 71 | */ |
| 72 | abstract public function name(); |
| 73 | |
| 74 | /** |
| 75 | * The id field in the database. |
| 76 | * |
| 77 | * @access public |
| 78 | * |
| 79 | * @return string |
| 80 | */ |
| 81 | public function id_field() { |
| 82 | return 'ID'; |
| 83 | } |
| 84 | |
| 85 | /** |
| 86 | * The table name. |
| 87 | * |
| 88 | * @access public |
| 89 | * |
| 90 | * @return string|bool |
| 91 | * @deprecated since 3.11.0 Use table() instead. |
| 92 | */ |
| 93 | public function table_name() { |
| 94 | _deprecated_function( __METHOD__, '3.11.0', 'Automattic\\Jetpack\\Sync\\Module->table' ); |
| 95 | return false; |
| 96 | } |
| 97 | |
| 98 | /** |
| 99 | * The table in the database with the prefix. |
| 100 | * |
| 101 | * @access public |
| 102 | * |
| 103 | * @return string|bool |
| 104 | */ |
| 105 | public function table() { |
| 106 | return false; |
| 107 | } |
| 108 | |
| 109 | /** |
| 110 | * The full sync action name for this module. |
| 111 | * |
| 112 | * @access public |
| 113 | * |
| 114 | * @return string |
| 115 | */ |
| 116 | public function full_sync_action_name() { |
| 117 | return 'jetpack_full_sync_' . $this->name(); |
| 118 | } |
| 119 | |
| 120 | // phpcs:disable VariableAnalysis.CodeAnalysis.VariableAnalysis.UnusedVariable |
| 121 | |
| 122 | /** |
| 123 | * Retrieve a sync object by its ID. |
| 124 | * |
| 125 | * @access public |
| 126 | * |
| 127 | * @param string $object_type Type of the sync object. |
| 128 | * @param int $id ID of the sync object. |
| 129 | * @return mixed Object, or false if the object is invalid. |
| 130 | */ |
| 131 | public function get_object_by_id( $object_type, $id ) { |
| 132 | return false; |
| 133 | } |
| 134 | |
| 135 | /** |
| 136 | * Initialize callables action listeners. |
| 137 | * Override these to set up listeners and set/reset data/defaults. |
| 138 | * |
| 139 | * @access public |
| 140 | * |
| 141 | * @param callable $callable Action handler callable. |
| 142 | */ |
| 143 | public function init_listeners( $callable ) { |
| 144 | } |
| 145 | |
| 146 | /** |
| 147 | * Initialize module action listeners for full sync. |
| 148 | * |
| 149 | * @access public |
| 150 | * |
| 151 | * @param callable $callable Action handler callable. |
| 152 | */ |
| 153 | public function init_full_sync_listeners( $callable ) { |
| 154 | } |
| 155 | |
| 156 | /** |
| 157 | * Initialize the module in the sender. |
| 158 | * |
| 159 | * @access public |
| 160 | */ |
| 161 | public function init_before_send() { |
| 162 | } |
| 163 | |
| 164 | /** |
| 165 | * Set module defaults. |
| 166 | * |
| 167 | * @access public |
| 168 | */ |
| 169 | public function set_defaults() { |
| 170 | } |
| 171 | |
| 172 | /** |
| 173 | * Perform module cleanup. |
| 174 | * Usually triggered when uninstalling the plugin. |
| 175 | * |
| 176 | * @access public |
| 177 | */ |
| 178 | public function reset_data() { |
| 179 | } |
| 180 | |
| 181 | /** |
| 182 | * Enqueue the module actions for full sync. |
| 183 | * |
| 184 | * @access public |
| 185 | * |
| 186 | * @param array $config Full sync configuration for this sync module. |
| 187 | * @param int $max_items_to_enqueue Maximum number of items to enqueue. |
| 188 | * @param boolean $state True if full sync has finished enqueueing this module, false otherwise. |
| 189 | * @return array Number of actions enqueued, and next module state. |
| 190 | */ |
| 191 | public function enqueue_full_sync_actions( $config, $max_items_to_enqueue, $state ) { |
| 192 | // In subclasses, return the number of actions enqueued, and next module state (true == done). |
| 193 | return array( null, true ); |
| 194 | } |
| 195 | |
| 196 | /** |
| 197 | * Retrieve an estimated number of actions that will be enqueued. |
| 198 | * |
| 199 | * @access public |
| 200 | * |
| 201 | * @param array $config Full sync configuration for this sync module. |
| 202 | * @return int Number of items yet to be enqueued. |
| 203 | */ |
| 204 | public function estimate_full_sync_actions( $config ) { |
| 205 | // In subclasses, return the number of items yet to be enqueued. |
| 206 | return null; |
| 207 | } |
| 208 | |
| 209 | // phpcs:enable VariableAnalysis.CodeAnalysis.VariableAnalysis.UnusedVariable |
| 210 | |
| 211 | /** |
| 212 | * Retrieve the actions that will be sent for this module during a full sync. |
| 213 | * |
| 214 | * @access public |
| 215 | * |
| 216 | * @return array Full sync actions of this module. |
| 217 | */ |
| 218 | public function get_full_sync_actions() { |
| 219 | return array(); |
| 220 | } |
| 221 | |
| 222 | /** |
| 223 | * Get the number of actions that we care about. |
| 224 | * |
| 225 | * @access protected |
| 226 | * |
| 227 | * @param array $action_names Action names we're interested in. |
| 228 | * @param array $actions_to_count Unfiltered list of actions we want to count. |
| 229 | * @return array Number of actions that we're interested in. |
| 230 | */ |
| 231 | protected function count_actions( $action_names, $actions_to_count ) { |
| 232 | return count( array_intersect( $action_names, $actions_to_count ) ); |
| 233 | } |
| 234 | |
| 235 | /** |
| 236 | * Calculate the checksum of one or more values. |
| 237 | * |
| 238 | * @access protected |
| 239 | * |
| 240 | * @param mixed $values Values to calculate checksum for. |
| 241 | * @param bool $sort If $values should have ksort called on it. |
| 242 | * @return int The checksum. |
| 243 | */ |
| 244 | protected function get_check_sum( $values, $sort = true ) { |
| 245 | // Associative array order changes the generated checksum value. |
| 246 | if ( $sort && is_array( $values ) ) { |
| 247 | $this->recursive_ksort( $values ); |
| 248 | } |
| 249 | return crc32( |
| 250 | wp_json_encode( |
| 251 | Functions::json_wrap( $values ), |
| 252 | 0 // phpcs:ignore Jetpack.Functions.JsonEncodeFlags.ZeroFound -- No `json_encode()` flags because we don't want disrupt the checksum algorithm. |
| 253 | ) |
| 254 | ); |
| 255 | } |
| 256 | |
| 257 | /** |
| 258 | * Recursively call ksort on an Array |
| 259 | * |
| 260 | * @param array $values Array. |
| 261 | */ |
| 262 | private function recursive_ksort( &$values ) { |
| 263 | ksort( $values ); |
| 264 | foreach ( $values as &$value ) { |
| 265 | if ( is_array( $value ) ) { |
| 266 | $this->recursive_ksort( $value ); |
| 267 | } |
| 268 | } |
| 269 | } |
| 270 | |
| 271 | /** |
| 272 | * Whether a particular checksum in a set of checksums is valid. |
| 273 | * |
| 274 | * @access protected |
| 275 | * |
| 276 | * @param array $sums_to_check Array of checksums. |
| 277 | * @param string $name Name of the checksum. |
| 278 | * @param int $new_sum Checksum to compare against. |
| 279 | * @return boolean Whether the checksum is valid. |
| 280 | */ |
| 281 | protected function still_valid_checksum( $sums_to_check, $name, $new_sum ) { |
| 282 | if ( isset( $sums_to_check[ $name ] ) && $sums_to_check[ $name ] === $new_sum ) { |
| 283 | return true; |
| 284 | } |
| 285 | |
| 286 | return false; |
| 287 | } |
| 288 | |
| 289 | /** |
| 290 | * Enqueue all items of a sync type as an action. |
| 291 | * |
| 292 | * @access protected |
| 293 | * |
| 294 | * @param string $action_name Name of the action. |
| 295 | * @param string $table_name Name of the database table. |
| 296 | * @param string $id_field Name of the ID field in the database. |
| 297 | * @param string $where_sql The SQL WHERE clause to filter to the desired items. |
| 298 | * @param int $max_items_to_enqueue Maximum number of items to enqueue in the same time. |
| 299 | * @param boolean $state Whether enqueueing has finished. |
| 300 | * @return array Array, containing the number of chunks and TRUE, indicating enqueueing has finished. |
| 301 | */ |
| 302 | protected function enqueue_all_ids_as_action( $action_name, $table_name, $id_field, $where_sql, $max_items_to_enqueue, $state ) { |
| 303 | global $wpdb; |
| 304 | |
| 305 | if ( ! $where_sql ) { |
| 306 | $where_sql = '1 = 1'; |
| 307 | } |
| 308 | |
| 309 | $items_per_page = 1000; |
| 310 | $page = 1; |
| 311 | $chunk_count = 0; |
| 312 | $previous_interval_end = $state ? $state : '~0'; |
| 313 | $listener = Listener::get_instance(); |
| 314 | |
| 315 | // Count down from max_id to min_id so we get newest posts/comments/etc first. |
| 316 | // phpcs:ignore Generic.CodeAnalysis.AssignmentInCondition.FoundInWhileCondition, WordPress.DB.PreparedSQL.InterpolatedNotPrepared |
| 317 | while ( $ids = $wpdb->get_col( "SELECT {$id_field} FROM {$table_name} WHERE {$where_sql} AND {$id_field} < {$previous_interval_end} ORDER BY {$id_field} DESC LIMIT {$items_per_page}" ) ) { |
| 318 | // Request posts in groups of N for efficiency. |
| 319 | $chunked_ids = array_chunk( $ids, self::ARRAY_CHUNK_SIZE ); |
| 320 | |
| 321 | // If we hit our row limit, process and return. |
| 322 | if ( $chunk_count + count( $chunked_ids ) >= $max_items_to_enqueue ) { |
| 323 | $remaining_items_count = $max_items_to_enqueue - $chunk_count; |
| 324 | $remaining_items = array_slice( $chunked_ids, 0, $remaining_items_count ); |
| 325 | $remaining_items_with_previous_interval_end = $this->get_chunks_with_preceding_end( $remaining_items, $previous_interval_end ); |
| 326 | $listener->bulk_enqueue_full_sync_actions( $action_name, $remaining_items_with_previous_interval_end ); |
| 327 | |
| 328 | $last_chunk = end( $remaining_items ); |
| 329 | return array( $remaining_items_count + $chunk_count, end( $last_chunk ) ); |
| 330 | } |
| 331 | $chunked_ids_with_previous_end = $this->get_chunks_with_preceding_end( $chunked_ids, $previous_interval_end ); |
| 332 | |
| 333 | $listener->bulk_enqueue_full_sync_actions( $action_name, $chunked_ids_with_previous_end ); |
| 334 | |
| 335 | $chunk_count += count( $chunked_ids ); |
| 336 | ++$page; |
| 337 | // The $ids are ordered in descending order. |
| 338 | $previous_interval_end = end( $ids ); |
| 339 | } |
| 340 | |
| 341 | if ( $wpdb->last_error ) { |
| 342 | // return the values that were passed in so all these chunks get retried. |
| 343 | return array( $max_items_to_enqueue, $state ); |
| 344 | } |
| 345 | |
| 346 | return array( $chunk_count, true ); |
| 347 | } |
| 348 | |
| 349 | /** |
| 350 | * Given the Module Full Sync Configuration and Status return the next chunk of items to send. |
| 351 | * |
| 352 | * @param array $config This module Full Sync configuration. |
| 353 | * @param array $status This module Full Sync status. |
| 354 | * @param int $chunk_size Chunk size. |
| 355 | * |
| 356 | * @return array|object|null |
| 357 | */ |
| 358 | public function get_next_chunk( $config, $status, $chunk_size ) { |
| 359 | // phpcs:disable WordPress.DB.PreparedSQL.InterpolatedNotPrepared,WordPress.DB.DirectDatabaseQuery.DirectQuery |
| 360 | global $wpdb; |
| 361 | return $wpdb->get_col( |
| 362 | " |
| 363 | SELECT {$this->id_field()} |
| 364 | FROM {$this->table()} |
| 365 | WHERE {$this->get_where_sql( $config )} |
| 366 | AND {$this->id_field()} < {$status['last_sent']} |
| 367 | ORDER BY {$this->id_field()} |
| 368 | DESC LIMIT {$chunk_size} |
| 369 | " |
| 370 | ); |
| 371 | // phpcs:enable WordPress.DB.PreparedSQL.InterpolatedNotPrepared,WordPress.DB.DirectDatabaseQuery.DirectQuery |
| 372 | } |
| 373 | |
| 374 | /** |
| 375 | * Return last_item to send for Module Full Sync Configuration. |
| 376 | * |
| 377 | * @param array $config This module Full Sync configuration. |
| 378 | * |
| 379 | * @return array|object|null |
| 380 | */ |
| 381 | public function get_last_item( $config ) { |
| 382 | global $wpdb; |
| 383 | // phpcs:disable WordPress.DB.PreparedSQL.InterpolatedNotPrepared,WordPress.DB.DirectDatabaseQuery.NoCaching,WordPress.DB.DirectDatabaseQuery.DirectQuery |
| 384 | return $wpdb->get_var( |
| 385 | " |
| 386 | SELECT {$this->id_field()} |
| 387 | FROM {$this->table()} |
| 388 | WHERE {$this->get_where_sql( $config )} |
| 389 | ORDER BY {$this->id_field()} |
| 390 | LIMIT 1 |
| 391 | " |
| 392 | ); |
| 393 | // phpcs:enable WordPress.DB.PreparedSQL.InterpolatedNotPrepared,WordPress.DB.DirectDatabaseQuery.NoCaching,WordPress.DB.DirectDatabaseQuery.DirectQuery |
| 394 | } |
| 395 | |
| 396 | /** |
| 397 | * Return the initial last sent object. |
| 398 | * |
| 399 | * @return string|array initial status. |
| 400 | */ |
| 401 | public function get_initial_last_sent() { |
| 402 | return '~0'; |
| 403 | } |
| 404 | |
| 405 | /** |
| 406 | * Immediately send all items of a sync type as an action. |
| 407 | * |
| 408 | * @access protected |
| 409 | * |
| 410 | * @param array $config Full sync configuration for this module. |
| 411 | * @param array $status the current module full sync status. |
| 412 | * @param float $send_until timestamp until we want this request to send full sync events. |
| 413 | * @param int $started The timestamp when the full sync started. |
| 414 | * |
| 415 | * @return array Status, the module full sync status updated. |
| 416 | */ |
| 417 | public function send_full_sync_actions( $config, $status, $send_until, $started ) { |
| 418 | global $wpdb; |
| 419 | |
| 420 | if ( empty( $status['last_sent'] ) ) { |
| 421 | $status['last_sent'] = $this->get_initial_last_sent(); |
| 422 | } |
| 423 | |
| 424 | $limits = Settings::get_setting( 'full_sync_limits' )[ $this->name() ] ?? |
| 425 | Defaults::get_default_setting( 'full_sync_limits' )[ $this->name() ] ?? |
| 426 | array( |
| 427 | 'max_chunks' => null, |
| 428 | 'chunk_size' => null, |
| 429 | ); |
| 430 | |
| 431 | $limits = array( |
| 432 | 'max_chunks' => is_numeric( $limits['max_chunks'] ) ? (int) $limits['max_chunks'] : 10, |
| 433 | 'chunk_size' => is_numeric( $limits['chunk_size'] ) ? (int) $limits['chunk_size'] : 100, |
| 434 | ); |
| 435 | |
| 436 | $limits['chunk_size'] = $this->adjust_chunk_size_if_stuck( $status['last_sent'], $limits['chunk_size'], $started ); |
| 437 | |
| 438 | $chunks_sent = 0; |
| 439 | |
| 440 | // Store last_item in status to avoid re-running this expensive query on every invocation. |
| 441 | // The minimum ID does not change during a Full Sync. |
| 442 | if ( ! isset( $status['last_item'] ) ) { |
| 443 | $status['last_item'] = $this->get_last_item( $config ); |
| 444 | } |
| 445 | $last_item = $status['last_item']; |
| 446 | |
| 447 | while ( $chunks_sent < $limits['max_chunks'] && microtime( true ) < $send_until ) { |
| 448 | $objects = $this->get_next_chunk( $config, $status, $limits['chunk_size'] ); |
| 449 | |
| 450 | if ( $wpdb->last_error ) { |
| 451 | $status['error'] = true; |
| 452 | return $status; |
| 453 | } |
| 454 | |
| 455 | if ( empty( $objects ) ) { |
| 456 | $status['finished'] = true; |
| 457 | return $status; |
| 458 | } |
| 459 | // If we have objects as a key it means get_next_chunk is being overridden, we need to check for it being an empty array. |
| 460 | // In case it is an empty array, we should not send the action or increase the chunks_sent, we just need to update the status. |
| 461 | if ( ! isset( $objects['objects'] ) || array() !== $objects['objects'] ) { |
| 462 | $key = $this->full_sync_action_name() . '_' . crc32( wp_json_encode( $status['last_sent'], JSON_UNESCAPED_SLASHES ) ); |
| 463 | $result = $this->send_action( $this->full_sync_action_name(), array( $objects, $status['last_sent'] ), $key ); |
| 464 | if ( is_wp_error( $result ) || $wpdb->last_error ) { |
| 465 | $status['error'] = true; |
| 466 | return $status; |
| 467 | } |
| 468 | ++$chunks_sent; |
| 469 | } |
| 470 | |
| 471 | // Updated the sent and last_sent status. |
| 472 | $status = $this->set_send_full_sync_actions_status( $status, $objects ); |
| 473 | if ( $last_item === $status['last_sent'] ) { |
| 474 | $status['finished'] = true; |
| 475 | return $status; |
| 476 | } |
| 477 | } |
| 478 | |
| 479 | return $status; |
| 480 | } |
| 481 | |
| 482 | /** |
| 483 | * Adjust chunk size using adaptive logic and update transient for tracking stuck state. |
| 484 | * |
| 485 | * @param string $last_sent The current last_sent marker. |
| 486 | * @param int $default_chunk_size The default chunk size. |
| 487 | * @param int $started The timestamp when the full sync started. |
| 488 | * @return int Adjusted chunk size. |
| 489 | */ |
| 490 | private function adjust_chunk_size_if_stuck( $last_sent, $default_chunk_size, $started ) { |
| 491 | $transient_key = 'jetpack_sync_last_sent_' . $this->name() . '_' . $started; |
| 492 | $stuck_data = get_transient( $transient_key ); |
| 493 | $is_stuck = isset( $stuck_data['last_sent'] ) && $stuck_data['last_sent'] === $last_sent; |
| 494 | |
| 495 | // Preserve the adjusted chunk size and stuck count from the transient when stuck. |
| 496 | $stuck_count = $is_stuck && isset( $stuck_data['stuck_count'] ) ? $stuck_data['stuck_count'] : 0; |
| 497 | $adjusted_chunk_size = $is_stuck && isset( $stuck_data['adjusted_chunk_size'] ) ? $stuck_data['adjusted_chunk_size'] : $default_chunk_size; |
| 498 | |
| 499 | if ( $is_stuck && $stuck_data['adjusted_chunk_size'] === 1 ) { |
| 500 | // Refresh transient TTL to prevent expiry-driven reset cycles. |
| 501 | set_transient( $transient_key, $stuck_data, HOUR_IN_SECONDS ); |
| 502 | return 1; // If we are already at the minimum chunk size, do not adjust further. |
| 503 | } |
| 504 | |
| 505 | // We will adjust if it is stuck after 10 minutes. |
| 506 | if ( |
| 507 | $is_stuck && |
| 508 | ( time() - $stuck_data['timestamp'] ) >= 10 * MINUTE_IN_SECONDS |
| 509 | ) { |
| 510 | ++$stuck_count; |
| 511 | $adjusted_chunk_size = max( 1, (int) ( $default_chunk_size / ( 2 ** $stuck_count ) ) ); // Halve the chunk size for each stuck iteration. |
| 512 | |
| 513 | // Send one HTTP notification when chunk size reaches the minimum (1) |
| 514 | // so the stuck state is visible for monitoring. Intermediate cascade |
| 515 | // steps skip the HTTP request to avoid consuming the time budget. |
| 516 | if ( 1 === $adjusted_chunk_size ) { |
| 517 | $this->send_action( |
| 518 | 'jetpack_full_sync_stuck_adjustment', |
| 519 | array( |
| 520 | 'module' => $this->name(), |
| 521 | 'last_sent' => $last_sent, |
| 522 | 'stuck_count' => $stuck_count, |
| 523 | 'adjusted_chunk_size' => $adjusted_chunk_size, |
| 524 | ) |
| 525 | ); |
| 526 | } |
| 527 | } |
| 528 | |
| 529 | // Set or update the transient with the new last_sent, timestamp, and stuck_count. |
| 530 | // Reset the timestamp when not stuck or after an adjustment, so each new chunk size |
| 531 | // gets a 10-minute window to prove itself before halving further. |
| 532 | $previous_chunk_size = $stuck_data['adjusted_chunk_size'] ?? null; |
| 533 | $reset_timestamp = ! $is_stuck || $adjusted_chunk_size !== $previous_chunk_size; |
| 534 | set_transient( |
| 535 | $transient_key, |
| 536 | array( |
| 537 | 'last_sent' => $last_sent, |
| 538 | 'timestamp' => $reset_timestamp ? time() : $stuck_data['timestamp'], |
| 539 | 'stuck_count' => $stuck_count, |
| 540 | 'adjusted_chunk_size' => $adjusted_chunk_size, |
| 541 | ), |
| 542 | HOUR_IN_SECONDS |
| 543 | ); |
| 544 | |
| 545 | return $adjusted_chunk_size; |
| 546 | } |
| 547 | |
| 548 | /** |
| 549 | * Set the status of the full sync action based on the objects that were sent. |
| 550 | * Used to update the status of the module after sending a chunk of objects. |
| 551 | * Since Full Sync logic chunking relies on order of items being processed in descending order, we need to sort |
| 552 | * due to some modules (e.g. WooCommerce) changing the order while getting the objects. |
| 553 | * |
| 554 | * @access protected |
| 555 | * |
| 556 | * @param array $status This module Full Sync status. |
| 557 | * @param array $objects This module Full Sync objects. |
| 558 | * |
| 559 | * @return array The updated status. |
| 560 | */ |
| 561 | protected function set_send_full_sync_actions_status( $status, $objects ) { |
| 562 | |
| 563 | $object_ids = $objects['object_ids'] ?? $objects; |
| 564 | $status['last_sent'] = end( $object_ids ); |
| 565 | $status['sent'] += count( $object_ids ); |
| 566 | return $status; |
| 567 | } |
| 568 | |
| 569 | /** |
| 570 | * Immediately sends a single item without firing or enqueuing it |
| 571 | * |
| 572 | * @param string $action_name The action. |
| 573 | * @param array $data The data associated with the action. |
| 574 | * @param string $key The key to use for the action. |
| 575 | */ |
| 576 | public function send_action( $action_name, $data = null, $key = null ) { |
| 577 | $sender = Sender::get_instance(); |
| 578 | return $sender->send_action( $action_name, $data, $key ); |
| 579 | } |
| 580 | |
| 581 | /** |
| 582 | * Retrieve chunk IDs with previous interval end. |
| 583 | * |
| 584 | * @access protected |
| 585 | * |
| 586 | * @param array $chunks All remaining items. |
| 587 | * @param int $previous_interval_end The last item from the previous interval. |
| 588 | * @return array Chunk IDs with the previous interval end. |
| 589 | */ |
| 590 | protected function get_chunks_with_preceding_end( $chunks, $previous_interval_end ) { |
| 591 | $chunks_with_ends = array(); |
| 592 | foreach ( $chunks as $chunk ) { |
| 593 | $chunks_with_ends[] = array( |
| 594 | 'ids' => $chunk, |
| 595 | 'previous_end' => $previous_interval_end, |
| 596 | ); |
| 597 | // Chunks are ordered in descending order. |
| 598 | $previous_interval_end = end( $chunk ); |
| 599 | } |
| 600 | return $chunks_with_ends; |
| 601 | } |
| 602 | |
| 603 | /** |
| 604 | * Get metadata of a particular object type within the designated meta key whitelist. |
| 605 | * |
| 606 | * @access protected |
| 607 | * |
| 608 | * @todo Refactor to use $wpdb->prepare() on the SQL query. |
| 609 | * |
| 610 | * @param array $ids Object IDs. |
| 611 | * @param string $meta_type Meta type. |
| 612 | * @param array $meta_key_whitelist Meta key whitelist. |
| 613 | * @return array Unserialized meta values. |
| 614 | */ |
| 615 | protected function get_metadata( $ids, $meta_type, $meta_key_whitelist ) { |
| 616 | global $wpdb; |
| 617 | $table = _get_meta_table( $meta_type ); |
| 618 | $id = $meta_type . '_id'; |
| 619 | if ( ! $table ) { |
| 620 | return array(); |
| 621 | } |
| 622 | |
| 623 | $private_meta_whitelist_sql = "'" . implode( "','", array_map( 'esc_sql', $meta_key_whitelist ) ) . "'"; |
| 624 | |
| 625 | return array_map( |
| 626 | array( $this, 'unserialize_meta' ), |
| 627 | $wpdb->get_results( |
| 628 | // phpcs:disable WordPress.DB.PreparedSQL.InterpolatedNotPrepared, WordPress.DB.PreparedSQL.NotPrepared |
| 629 | "SELECT $id, meta_key, meta_value, meta_id FROM $table WHERE $id IN ( " . implode( ',', wp_parse_id_list( $ids ) ) . ' )' . |
| 630 | " AND meta_key IN ( $private_meta_whitelist_sql ) ", |
| 631 | // phpcs:enable WordPress.DB.PreparedSQL.InterpolatedNotPrepared, WordPress.DB.PreparedSQL.NotPrepared |
| 632 | OBJECT |
| 633 | ) |
| 634 | ); |
| 635 | } |
| 636 | |
| 637 | /** |
| 638 | * Initialize listeners for the particular meta type. |
| 639 | * |
| 640 | * @access public |
| 641 | * |
| 642 | * @param string $meta_type Meta type. |
| 643 | * @param callable $callable Action handler callable. |
| 644 | */ |
| 645 | public function init_listeners_for_meta_type( $meta_type, $callable ) { |
| 646 | add_action( "added_{$meta_type}_meta", $callable, 10, 4 ); |
| 647 | add_action( "updated_{$meta_type}_meta", $callable, 10, 4 ); |
| 648 | add_action( "deleted_{$meta_type}_meta", $callable, 10, 4 ); |
| 649 | } |
| 650 | |
| 651 | /** |
| 652 | * Initialize meta whitelist handler for the particular meta type. |
| 653 | * |
| 654 | * @access public |
| 655 | * |
| 656 | * @param string $meta_type Meta type. |
| 657 | * @param callable $whitelist_handler Action handler callable. |
| 658 | */ |
| 659 | public function init_meta_whitelist_handler( $meta_type, $whitelist_handler ) { |
| 660 | add_filter( "jetpack_sync_before_enqueue_added_{$meta_type}_meta", $whitelist_handler ); |
| 661 | add_filter( "jetpack_sync_before_enqueue_updated_{$meta_type}_meta", $whitelist_handler ); |
| 662 | add_filter( "jetpack_sync_before_enqueue_deleted_{$meta_type}_meta", $whitelist_handler ); |
| 663 | } |
| 664 | |
| 665 | /** |
| 666 | * Retrieve the term relationships for the specified object IDs. |
| 667 | * |
| 668 | * @access protected |
| 669 | * |
| 670 | * @todo This feels too specific to be in the abstract sync Module class. Move it? |
| 671 | * |
| 672 | * @param array $ids Object IDs. |
| 673 | * @return array Term relationships - object ID and term taxonomy ID pairs. |
| 674 | */ |
| 675 | protected function get_term_relationships( $ids ) { |
| 676 | global $wpdb; |
| 677 | |
| 678 | // phpcs:ignore WordPress.DB.PreparedSQL.NotPrepared |
| 679 | return $wpdb->get_results( "SELECT object_id, term_taxonomy_id FROM $wpdb->term_relationships WHERE object_id IN ( " . implode( ',', wp_parse_id_list( $ids ) ) . ' )', OBJECT ); |
| 680 | } |
| 681 | |
| 682 | /** |
| 683 | * Unserialize the value of a meta object, if necessary. |
| 684 | * |
| 685 | * @access public |
| 686 | * |
| 687 | * @param object $meta Meta object. |
| 688 | * @return object Meta object with possibly unserialized value. |
| 689 | */ |
| 690 | public function unserialize_meta( $meta ) { |
| 691 | $meta->meta_value = maybe_unserialize( $meta->meta_value ); |
| 692 | return $meta; |
| 693 | } |
| 694 | |
| 695 | /** |
| 696 | * Retrieve a set of objects by their IDs. |
| 697 | * |
| 698 | * @access public |
| 699 | * |
| 700 | * @param string $object_type Object type. |
| 701 | * @param array $ids Object IDs. |
| 702 | * @return array Array of objects. |
| 703 | */ |
| 704 | public function get_objects_by_id( $object_type, $ids ) { |
| 705 | if ( empty( $ids ) || empty( $object_type ) ) { |
| 706 | return array(); |
| 707 | } |
| 708 | |
| 709 | $objects = array(); |
| 710 | foreach ( (array) $ids as $id ) { |
| 711 | $object = $this->get_object_by_id( $object_type, $id ); |
| 712 | |
| 713 | // Only add object if we have the object. |
| 714 | if ( $object ) { |
| 715 | $objects[ $id ] = $object; |
| 716 | } |
| 717 | } |
| 718 | |
| 719 | return $objects; |
| 720 | } |
| 721 | |
| 722 | /** |
| 723 | * Gets a list of minimum and maximum object ids for each batch based on the given batch size. |
| 724 | * |
| 725 | * @access public |
| 726 | * |
| 727 | * @param int $batch_size The batch size for objects. |
| 728 | * @param string|bool $where_sql The sql where clause minus 'WHERE', or false if no where clause is needed. |
| 729 | * |
| 730 | * @return array|bool An array of min and max ids for each batch. FALSE if no table can be found. |
| 731 | */ |
| 732 | public function get_min_max_object_ids_for_batches( $batch_size, $where_sql = false ) { |
| 733 | |
| 734 | if ( ! $this->table() ) { |
| 735 | return false; |
| 736 | } |
| 737 | |
| 738 | $results = array(); |
| 739 | $table = $this->table(); |
| 740 | $current_max = 0; |
| 741 | $current_min = 1; |
| 742 | $id_field = $this->id_field(); |
| 743 | $replicastore = new Replicastore(); |
| 744 | |
| 745 | $total = $replicastore->get_min_max_object_id( |
| 746 | $id_field, |
| 747 | $table, |
| 748 | $where_sql, |
| 749 | false |
| 750 | ); |
| 751 | |
| 752 | while ( $total->max > $current_max ) { |
| 753 | $where = $where_sql ? |
| 754 | $where_sql . " AND $id_field > $current_max" : |
| 755 | "$id_field > $current_max"; |
| 756 | $result = $replicastore->get_min_max_object_id( |
| 757 | $id_field, |
| 758 | $table, |
| 759 | $where, |
| 760 | $batch_size |
| 761 | ); |
| 762 | if ( empty( $result->min ) && empty( $result->max ) ) { |
| 763 | // Our query produced no min and max. We can assume the min from the previous query, |
| 764 | // and the total max we found in the initial query. |
| 765 | $current_max = (int) $total->max; |
| 766 | $result = (object) array( |
| 767 | 'min' => $current_min, |
| 768 | 'max' => $current_max, |
| 769 | ); |
| 770 | } else { |
| 771 | $current_min = (int) $result->min; |
| 772 | $current_max = (int) $result->max; |
| 773 | } |
| 774 | $results[] = $result; |
| 775 | } |
| 776 | |
| 777 | return $results; |
| 778 | } |
| 779 | |
| 780 | /** |
| 781 | * Return Total number of objects. |
| 782 | * |
| 783 | * @param array $config Full Sync config. |
| 784 | * |
| 785 | * @return int total |
| 786 | */ |
| 787 | public function total( $config ) { |
| 788 | global $wpdb; |
| 789 | $table = $this->table(); |
| 790 | $where = $this->get_where_sql( $config ); |
| 791 | |
| 792 | // phpcs:ignore WordPress.DB.PreparedSQL.InterpolatedNotPrepared,WordPress.DB.DirectDatabaseQuery.DirectQuery,WordPress.DB.DirectDatabaseQuery.NoCaching |
| 793 | return (int) $wpdb->get_var( "SELECT COUNT(*) FROM $table WHERE $where" ); |
| 794 | } |
| 795 | |
| 796 | /** |
| 797 | * Retrieve the WHERE SQL clause based on the module config. |
| 798 | * |
| 799 | * @access public |
| 800 | * |
| 801 | * @param array $config Full sync configuration for this sync module. |
| 802 | * @return string WHERE SQL clause, or `null` if no comments are specified in the module config. |
| 803 | */ |
| 804 | public function get_where_sql( $config ) { // phpcs:ignore VariableAnalysis.CodeAnalysis.VariableAnalysis.UnusedVariable |
| 805 | return '1=1'; |
| 806 | } |
| 807 | |
| 808 | /** |
| 809 | * Filters objects and metadata based on maximum size constraints. |
| 810 | * It always allows the first object with its metadata, even if they exceed the limit. |
| 811 | * |
| 812 | * @access public |
| 813 | * |
| 814 | * @param string $type The type of objects to filter (e.g., 'post' or 'comment'). |
| 815 | * @param array $objects The array of objects to filter (e.g., posts or comments). |
| 816 | * @param array $metadata The array of metadata to filter. |
| 817 | * @param int $max_meta_size Maximum size for individual objects. |
| 818 | * @param int $max_total_size Maximum combined size for objects and metadata. |
| 819 | * @return array An array containing the filtered object IDs, filtered objects, and filtered metadata. |
| 820 | */ |
| 821 | public function filter_objects_and_metadata_by_size( $type, $objects, $metadata, $max_meta_size, $max_total_size ) { |
| 822 | $filtered_objects = array(); |
| 823 | $filtered_metadata = array(); |
| 824 | $filtered_object_ids = array(); |
| 825 | $current_size = 0; |
| 826 | |
| 827 | foreach ( $objects as $object ) { |
| 828 | $object_size = strlen( (string) maybe_serialize( $object ) ); |
| 829 | $current_metadata = array(); |
| 830 | $metadata_size = 0; |
| 831 | $id_field = $this->id_field(); |
| 832 | $object_id = (int) ( is_object( $object ) ? $object->{$id_field} : $object[ $id_field ] ); |
| 833 | |
| 834 | foreach ( $metadata as $key => $metadata_item ) { |
| 835 | if ( (int) $metadata_item->{$type . '_id'} === $object_id ) { |
| 836 | $metadata_item_size = strlen( (string) maybe_serialize( $metadata_item ) ); |
| 837 | if ( $metadata_item_size >= $max_meta_size ) { |
| 838 | $metadata_item->meta_value = ''; // Trim metadata if too large. |
| 839 | $metadata_item_size = strlen( (string) maybe_serialize( $metadata_item ) ); |
| 840 | } |
| 841 | $current_metadata[] = $metadata_item; |
| 842 | $metadata_size += $metadata_item_size; |
| 843 | |
| 844 | if ( ! empty( $filtered_object_ids ) && ( $current_size + $object_size + $metadata_size ) > $max_total_size ) { |
| 845 | break 2; // Exit both loops. |
| 846 | } |
| 847 | unset( $metadata[ $key ] ); |
| 848 | } |
| 849 | } |
| 850 | |
| 851 | // Always allow the first object with metadata. |
| 852 | if ( empty( $filtered_object_ids ) || ( $current_size + $object_size + $metadata_size ) <= $max_total_size ) { |
| 853 | $filtered_object_ids[] = strval( is_object( $object ) ? $object->{$id_field} : $object[ $id_field ] ); |
| 854 | $filtered_objects[] = $object; |
| 855 | $filtered_metadata = array_merge( $filtered_metadata, $current_metadata ); |
| 856 | $current_size += $object_size + $metadata_size; |
| 857 | } else { |
| 858 | break; |
| 859 | } |
| 860 | } |
| 861 | |
| 862 | return array( |
| 863 | $filtered_object_ids, |
| 864 | $filtered_objects, |
| 865 | $filtered_metadata, |
| 866 | ); |
| 867 | } |
| 868 | } |