1.34.0

Included commits: 1.33.1...1.34.0

Fixes

• Fixed duplication and deadlock issues.

1.33.1

Included commits: 1.33.0...1.33.1

Improvements

• Added phpstan/phpstan new ^2.* module version support.

1.33.0

Included commits: 1.32.0...1.33.0

Improvements

• Increased PHPUnit version.

1.32.0

Included commits: 1.31.0...1.32.0

Improvements

• Introduced HydrateEventsRequest transfer.
• Introduced HydrateEventsResponse transfer.
• Introduced EventBehaviorFacadeInterface::hydrateEventDataTransfer() facade method to get event data with timestamp.

1.31.0

Included commits: 1.30.1...1.31.0

Improvements

• Adjusted TriggerManager to avoid duplicate event messages with the same IDs.
• Adjusted EventBehavior to use correct PrimaryKey in propel module for event message if the table doesn't have primary key.
• Increased Kernel version dependency.
• Increased Propel version dependency.

1.30.1

Included commits: 1.30.0...1.30.1

Fixes

• Adjusted EventBehaviorFacade::triggerRuntimeEvents() to not trigger delete on spy_event_behavior_entity_change table when no events were triggered.

1.30.0

Included commits: 1.29.0...1.30.0

Improvements

• Added PHP 8.3 support.

Checklist

• I agree with the Code Contribution License Agreement in CONTRIBUTING.md

1.29.0

Included commits: 1.28.1...1.29.0

Improvements

• Introduced EventBehaviorConstants::MAX_EVENT_MESSAGE_DATA_SIZE to define the recommended maximum data size for event messages (in KB).
• Introduced EventBehaviorConfig::getMaxEventMessageDataSize() method.
• Adjusted EventBehavior::saveEventBehaviorEntityChange() to log a warning messages when event message data size exceeds the limit defined by EventBehaviorConfig::getMaxEventMessageDataSize().

1.28.1

Included commits: 1.28.0...1.28.1

Adjustments

• Adjusted EventBehaviorCommunicationFactory so it extends AbstractCommunicationFactory.

1.28.0

Included commits: 1.27.0...1.28.0

Adjustments

• Adjusted EventBehaviorFacade::triggerRuntimeEvents() so processed events are removed at the end of method execution by process ID instead of removing them in batches, this change partially fixes a lock issue with big parallel requests.