mirror of
https://github.com/honeymoose/OpenSearch.git
synced 2025-02-06 13:08:29 +00:00
5f52bc4c9f
Fixes two bugs introduced by #57627: 1. We were not properly letting go of memory from the request breaker when the aggregation finished. 2. We no longer supported totally arbitrary stuff produced by the init script because we *assumed* that it'd be ok to run the script once and clone its results. Sadly, cloning can't clone *anything* that the init script can make, like `String` arrays. This runs the init script once for every new bucket so we don't need to clone.