mirror of
https://github.com/honeymoose/OpenSearch.git
synced 2025-02-09 22:45:04 +00:00
a615ab454a
The trigger engine did always create a new schedule data structure, when the watcher indexing listener called an add. However the indexing listener also called add, when the watch status was updated. This means, that upon a watch status update the watch got retriggered, potentially waiting a defined interval from the watch status update onwards, instead of waiting from the last run. This commit only updates the schedule in the trigger engine, if it actually has changed, otherwise the existing schedule will not be touched. This has two results 1. If a watch is updated by an execution, the existing interval will not be touched (meaning the scheduled time will not move forward). 2. If a watch is updated by a user, but the schedule is not changed, it will not be reset from the update (for example starting to count from 5 minutes again, if the interval was set to 5 minutes). Furthermore some minor cleanups were applied, making variables final in the ctor, preventing double creation of variables.