Oracle integration tests where failing when the docker container was freshly started,
as it happens every time on CI.
The reason was mainly a misuse of Awaitility.ignoreException(Class) which only ignores exception of exactly the class provided, not of subtypes.
Apart from fixing this for Oracle the complete logic for verifying the connection was moved to DataSourceConfiguration so other database use it as well in a consistent manner.
Original pull request: #237.
Rename RequiredFeature to EnabledOnFeature to align with JUnit terminology. Use RunWith instead of Spring rules. Reformat code. Update documentation.
Original pull request: #232.
An Oracle database is started in Docker using Testcontainers.
Tests that don't work yet are skipped using the TestDatabaseFeatures which allows central control, which database supports which feature.
A suitable oracle image is deployed on DockerHub as "springci/spring-data-oracle-xe-prebuild:18.4.0".
The official Oracle docker images as described here https://github.com/oracle/docker-images/blob/master/OracleDatabase/SingleInstance/README.md are not suitable since it needs about 15min to start and also TestContainers seems to break it by trying to us it to early.
The referenced image was created based on these instructions: https://github.com/oracle/docker-images/tree/master/OracleDatabase/SingleInstance/samples/prebuiltdb
The following features don't yet work for Oracle:
* Loading of date like properties.
* Ids with custom names.
* Array support.
* Boolean properties.
* BigDecimals and BigInteger is limited to what fits into Oracles NUMBER data type.
* Entities that use a join during load time, i.e. with a one to one relationship.
Original pull request: #232.
This changes the behavior of the JDBC driver from always returning `MySQL` for product name to a more involved logic which might result in it returning `MariaDB`.
Original pull request: #231.
Using Criteria.from(…) with multiple Criteria objects now uses properly AND combination along with group nesting to render a correct criteria. Previously, the INITIAL combinator in groups caused a mapping exception.
Let SqlIdentifier implement Streamable<SqlIdentifier> to allow iteration of composed identifiers.
Adapt dialect to iteration changes.
Rewrite simple if blocks to using ternary operators.
Original pull request: #221.
We now no longer double-wrap parameters for IN criteria. Previously, collection arguments were wrapped into another collection which caused double-wrapped lists.
Original pull request: #217.
Custom value types like outlined below can now server as id property
public class IdValue {
String id;
}
public class DomainType {
@Id
IdValue id;
String value;
// ...
}
Enabled access to identifier parts by String arguments in order to make usage with MyBatis feasable.
Properly pass identifier information to MyBatis for insert statements.
Original pull request: #218.
Using `execute` instead of `query` since we are not interested in the results.
Refactoring of the concurrency tests.
Make the concurrency tests run with all databases.
Added support for DB2.
Moved AnsiDialect to the main sources so it can be referenced in other Dialects.
Original pull request: #196.
