Issues

@IdClass, even not @Indexed, leads to NPE on bootstrap

Upgrade to Hibernate ORM 5.4.13.Final

Simplify and improve ordering and parallelism of Elasticsearch indexing

Always enable delayed commits in the Lucene backend

Upgrade to Elasticsearch 7.6.1

java.lang.IndexOutOfBoundsException when an Elasticsearch bulk work fails

Fix SearchQueryResultLoadingOrTransformingIT failures caused by incorrectly expecting ordered hits

Upgrade to Lucene 8.5.0

Stop testing against JDK13 in the CI build

Clean up any change to the permission of index files during tests

Elasticsearch search queries fetch the whole document source by default

Refactor FieldSearchSortIT to enable testing more complex cases

.desc().missing().last() / .asc().missing().first() places documents around 0 instead of last/first

upgrade maven-checkstyle-plugin to 3.1.1

ConcurrentModificationException in very specific scenarios when indexing multiple entities with addOrUpdate which contains @IndexEmbedded dependencies

Remove the unnecessary generic type parameter for DocumentElement

Failure to index reports the wrong document ID with Elasticsearch discriminator multi-tenancy

Entities whose indexing failed are not reported when even just one document ID cannot be converted

Expose a tool to build property keys

Create-or-validate schema management by default

Dynamic entities are ignored when creating a scope targeting Object.class

Customize enabled syntax elements in simple query strings

Prefix query in SimpleQueryString predicate is case sensitive

NullPointerException when using simpleQueryString() on non-analyzed, non-normalized fields

Test Hibernate Search with JDK15 regularly

Offline startup of the Elasticsearch backend

Restore support for concurrent Lucene work execution - at least during mass indexing

Mark experimental APIs/SPIs with an @Incubating annotation

@IndexedEmbedded should be repeatable

Initialization options and APIs

Implicit nested predicates

dropAndCreateSchema() as an alternative to purge() in the MassIndexer

Support a "degraded mode" where failure to contact Elasticsearch or index documents does not affect the whole application

Add configuration options for the size and number of indexing queues, and the max size of Elasticsearch bulks

Sorts on multi-valued fields

Current implementation of SpatialBridge prevents multi-valued locations