Different JPQL queries might be mapped to the same QueryKey wich causes severe problems when using query caching as the cached result from another query might be returned.

Take for example these 2 JPQL query (with their parameters in [])

select cacheTest from at.bluesource.CacheTest cacheTest WHERE cacheTest.boola = ?1 AND cacheTest.boolb = ?2 AND cacheTest.boolc = ?1	[1=true, 2=false]

select cacheTest from at.bluesource.CacheTest cacheTest WHERE cacheTest.boola = ?1 AND cacheTest.boolb = ?2 AND cacheTest.boolc = ?2	[1=true, 2=false]

The org.hibernate.cache.spi.QueryKey for those 2 JPQL queries generated by org.hibernate.loader.QueryLoader.generateQueryKey(..) (implemented in the superclass org.hibernate.loader.Loader) both use the same SQL string and the same query parameters

SQL string:

select cachetest0_.id as id1_36_, cachetest0_.boola as boola2_36_, cachetest0_.boolb as boolb_3_36_, cachetest0_.boolc as boolc_4_36_ where cachetest0_.boola=? and cachetest0_.boolb=? and cachetest0_.boolc=?

Parameters:

[1=true, 2=false]

Thus calling the first and the second JPQL query without any database modifications in between returns twice the result set for the first query whereas the result sets should differ (as they do when not using query caching)