What I'd really like to do is to nail down the use cases we need to support here. I see:
- non-JTA (connection pooling + native JDBC Connection transaction handling) - we will need variants here to account for combinations of:
- ExtractedDatabaseMetaData#supportsDataDefinitionInTransaction
- ExtractedDatabaseMetaData#doesDataDefinitionCauseTransactionCommit
- JTA - we will need variants here to account for combinations of:
- CMT versus BMT
- initially active
- initially inactive
- ExtractedDatabaseMetaData#supportsDataDefinitionInTransaction
- ExtractedDatabaseMetaData#doesDataDefinitionCauseTransactionCommit
|