In version 1.3.172, a performance regression was introduced when fixing the issue 389 (when there is a multi-column primary key, H2 does not seem to always pick the right index). This was related to boosting an index that matches the "order by" column list (the wrong index was used in some cases).

Rename "getNoDimensions" to "getDimensionCount", because "getNoDimensions" is confusing for non-native English speaking people ("get no dimensions versus get yes dimensions" ?!)

In version 1.3.172, a performance regression was introduced when fixing the issue 389 (when there is a multi-column primary key, H2 does not seem to always pick the right index). This was related to boosting an index that matches the "order by" column list (the wrong index was used in some cases).

We need two methods - one for testing the Class<?> and one for testing an Object

remove testRandom(), Nicolas Fortin says this test case is now redundant, because the necessary logic is tested by the MVMap unit test

sessions are creating large transactions.

When running the Recover tool on very large (>6G) databases, some statistics were reported with negative numbers.

Queries "group by" that contain a subquery with an aggregate function returned the wrong result in some cases.

New feature from Davide Cavestro - allow using custom Java object serialization engines on a per-DB basis.

when the search key was a different cardinal type from the table index key.
e.g. where the one was INT and the other was LONG

make NonUniqueHashIndex directly subclass BaseIndex. There is no point in sub-classing HashIndex if it is going to override all of the logic anyhow.