LUCENE-9068 moved construction of FuzzyQuery's automaton directly onto the query itself. However, as SOLR-14428 demonstrates, this ends up blowing up query caches that assume query objects are fairly small when calculating their memory usage. We should move automaton construction back into FuzzyTermsEnum, while keeping as much of the nice refactoring of LUCENE-9068 as possible.
- is related to
SOLR-14428 FuzzyQuery has severe memory usage in 8.5
- links to