X-Git-Url: http://wamblee.org/gitweb/?a=blobdiff_plain;f=test%2Fenterprise%2Fsrc%2Fmain%2Fjava%2Forg%2Fwamblee%2Fsupport%2Fpersistence%2Fpackage-info.java;fp=test%2Fenterprise%2Fsrc%2Fmain%2Fjava%2Forg%2Fwamblee%2Fsupport%2Fpersistence%2Fpackage-info.java;h=d38be5796fed215c5341d165ce549af23a5d4bbd;hb=28a9820d1b2fe0dbf9a05aed5a84b79f96cde743;hp=e30b0b324eb18f55a3a75b61d85acb27c36310ee;hpb=4b9f81b79e3d2e0f607a6a4126bf02b56b6a98ad;p=utils diff --git a/test/enterprise/src/main/java/org/wamblee/support/persistence/package-info.java b/test/enterprise/src/main/java/org/wamblee/support/persistence/package-info.java index e30b0b32..d38be579 100644 --- a/test/enterprise/src/main/java/org/wamblee/support/persistence/package-info.java +++ b/test/enterprise/src/main/java/org/wamblee/support/persistence/package-info.java @@ -13,5 +13,202 @@ * See the License for the specific language governing permissions and * limitations under the License. */ +/** + *
+ * This package provides test library for database testing in general and JPA testing + * specifically. As part of this it provides a means to transparently start an inmemory + * database from a junit test or connect to an external database. Also, given a persistence + * unit it is easy to start testing it in a junit test with only a few lines of code. + *
+ * + *+ * The main use cases are explained below: + *
+ *+ * See the design overview for a an overview of the design. + *
+ * + * + *+ * Database db = DatabaseBuilder.getDatabase(); + * DataSource dataSource = db.start(); + *+ * If nothing is specified in the user's environment, an inmemory database is started (derby). + * Using the datasource is just standard JDBC now. + * Stopping the database + *
+ * db.stop(); + *+ * + * + *
+ * Database db = DatabaseBuilder.getDatabase(DatabaseProvider.CAPABILITY_EXTERNAL); + *+ * This also requires a number of environment variables or system properties to be set, + * see {@link ExternalDatabase}. + * + * + * + *
+ * To execute code withing a JDBC transaction, use the {@link DatabaseUtils} and + * use the {@link DatabaseUtils#executeInTransaction(org.wamblee.support.persistence.DatabaseUtils.JdbcUnitOfWork)} + * method. + *
+ *+ * DatabaseUtils dbutils = new DatabaseUtils(dataSource); + * boolean result = dbutils.executeInTransaction( + * new JdbcUnitOfWork<Boolean>() { + * @Override + * public Boolean execute(Connection aConnection) throws Exception { + * ResultSet res = jpaTester.getDbUtils().executeQuery( + * aConnection, GROUP_QUERY, aGroup); + * return res.next(); + * } + * }); + *+ * {@link DatabaseUtils} also provides various other utility methods to work with JDBC queries. + * + * + *
To work with DBUnit, {@link DatabaseUtils#createDbTester(org.dbunit.dataset.filter.ITableFilterSimple)} + * must be used passing it in the tables to use in the form of a {@link org.dbunit.dataset.filter.ITableFilterSimple} + * object.
+ * + *+ * IDatabaseTester dbtester = dbutils.createDbTester(new ITableFilterSimple() { + * public boolean accept(String aTableName) throws DataSetException { + * return aTableName.startsWith("XYZ_"); + * } + * }); + *+ * + *
The reason for using a DatabaseUtils
instead of DBUnit directly is that
+ * DatabseUtils
will keep track of connections and close them when DatabaseUtils
+ * is closed
+ * First step is to create a {@link PersistenceUnitDescription} that matches the persistence unit you + * want to test. + *
+ *Second step is to make sure that all entities are listed explicitly in your + *
persistence.xml. Currently, class path scanning appears to fail when + * run from junit. + * Specifying all entities explicitly is not necessarily a bad thing as it is also more efficient. + * + * + *
+ * Now create a JpaTester
in your test code:
+ *
+ * @Before + * public void setUp() throws Exception { + * + * // First we create the JpaTester by telling us which persistence unit we + * // are going to test + * jpaTester = new JpaTester(new MyPersistenceUnit()); + * jpaTester.start(); + * } + *+ * + *
+ * Then in test code execute some JPA code within a unit of work: + *
+ *+ * jpaTester.getJpaBuilder().execute(new JpaUnitOfWork+ * + * + *() { + * public Void execute(EntityManager aEm) { + * MyEntity entity = new MyEntity("a", "b"); + * aEm.persist(entity); + * return null; + * } + * }); + *
The JPATester
provides access to all required object. It is usually convenient to
+ * get them directly from the JPATester
after initializing it:
+ *
+ * builder = jpaTester.getJpaBuilder(); + * dbtester = jpaTester.getDbUtils().createDbTester(new MyTables()); + * dbutils = jpaTester.getDbUtils(); + *+ * + * + *
Using {@link TransactionProxyFactory} it is possible to create a proxy for a given + * service interface to provide the semantics of 'requires new' transaction semantics. + *
+ * TransactionProxyFactory+ * + *factory = new TransactionProxyFactory ( + * jpaTester.getJpaBuilder(), Service.class); + * Service service = new ServiceImpl(factory.getTransactionScopedEntityManager()); + * Service proxy = factory.getProxy(service); + * proxy.execute(...); + *
+ * In the above example, the Service
POJO requires an {@link EntityManager} in its
+ * constructor and it is passed a transaction scoped entitymanager from the factory. This entitymanager
+ * is in fact a so-called contextual reference.
+ * Next, the proxy is obtained from the factory. Invoking any method on it will make sure a new
+ * transaction is started and a new entity manager is created for the scope of that transaction.
+ *