Branch: refs/heads/master Home: https://github.com/hibernate/hibernate-search Commit: 277eb3b68a108ab78df669fe38fbe2fb2f9681db https://github.com/hibernate/hibernate-search/commit/277eb3b68a108ab78df6... Author: Yoann Rodière <yoann(a)hibernate.org&gt; Date: 2020-09-24 (Thu, 24 Sep 2020) Changed paths: M integrationtest/backend/tck/src/main/java/org/hibernate/search/integrationtest/backend/tck/search/query/SearchQueryFetchIT.java Log Message: ----------- HSEARCH-4019 Rig scoring in SearchQueryFetchIT so that even sharded ES indexes yield the same sort Sharded indexes would usually yield less-than-perfect scoring when it comes to repeated terms, because the impact of term repetition on scoring depends on the frequency of that term in a shard, and so it depends on to which shard each document was routed, which is not easily predictable. By assigning a different term and boost to each document, we ensure that the documents will be in the correct order.