Implemented new prose test for Transactions #1824
Conversation
…ons are ignored when inside a transaction. JAVA-5684
|
|
||
| @DisplayName("Options Inside Transaction Prose Tests. 1. Write concern not inherited from collection object inside transaction") | ||
| @Test | ||
| void testWriteConcernInheritance() { |
There was a problem hiding this comment.
Having to use map / flatMap to pass the session into each Mono is painful. But appears to be the correct way to ensure that the session use is correctly ordered.
| import static com.mongodb.ClusterFixture.isSharded; | ||
| import static org.junit.Assert.assertTrue; | ||
| import static org.junit.Assume.assumeTrue; | ||
| import static org.junit.jupiter.api.Assertions.assertDoesNotThrow; |
There was a problem hiding this comment.
I took the opportunity to modernize the test to junit5
| session.startTransaction(); | ||
| addresses.add(collection.find(session, Document.parse("{}"))); | ||
| try (MongoCursor<Document> cursor = collection.find(session, Document.parse("{}")).cursor()) { | ||
| addresses.add(cursor.getServerAddress()); |
There was a problem hiding this comment.
The test wasn't to spec as it didn't actually check the server addresses, as it does in the prose test spec.
| session = client.startSession(); | ||
| public void testNewTransactionUnpinsSession() { | ||
| collection.insertOne(Document.parse("{}")); | ||
| try (ClientSession session = client.startSession()) { |
There was a problem hiding this comment.
Prefer try with resources (this code was from java 6 days).
| return isSharded(); | ||
| @DisplayName("Options Inside Transaction Prose Tests. 1. Write concern not inherited from collection object inside transaction") | ||
| @Test | ||
| void testWriteConcernInheritance() { |
…on cold publishers Only run the validation when subscribe has been called. JAVA-5985
| } | ||
|
|
||
| @Override | ||
| public Publisher<Void> commitTransaction() { |
There was a problem hiding this comment.
Writing the reactive streams prose test identified a bug. Validation was done when the publisher was created and not when the publisher was being used.
This complicated the usage of transactions, a fact that was hidden by the lack of prose tests! This fix is in the remit of JAVA-5895
|
|
||
| @BeforeEach | ||
| public void setUp() { | ||
| assumeTrue(isSharded()); |
There was a problem hiding this comment.
Should we also check for server version as per the spec?
| assumeTrue(isSharded()); | |
| assumeTrue(isSharded()); | |
| assumeTrue(serverVersionAtLeast(4, 2)); |
| .applyConnectionString(multiMongosConnectionString); | ||
|
|
||
| client = MongoClients.create(MongoClientSettings.builder(builder.build()) | ||
| .applyToSocketSettings(builder1 -> builder1.readTimeout(5, TimeUnit.SECONDS)) |
There was a problem hiding this comment.
Why do we need the SocketSettings?
| collection.insertOne(Document.parse("{}")); | ||
| session = client.startSession(); | ||
| public void testNewTransactionUnpinsSession() { | ||
| collection.insertOne(Document.parse("{}")); |
There was a problem hiding this comment.
we can add assumeTrue(serverVersionAtLeast(4, 2)); // spec requires 4.1.6
| try { | ||
| collection.insertOne(Document.parse("{}")); | ||
| session = client.startSession(); | ||
| collection.insertOne(Document.parse("{}")); |
| try (ClientSession session = client.startSession()) { | ||
| session.startTransaction(); | ||
| collection.insertOne(session, Document.parse("{ _id : 1 }")); | ||
| session.commitTransaction(); |
There was a problem hiding this comment.
The spec test doesn't commit the transaction, an oversight? (same for the previous test)
| @DisplayName("Options Inside Transaction Prose Tests. 1. Write concern not inherited from collection object inside transaction") | ||
| @Test | ||
| void testWriteConcernInheritance() { | ||
| assumeTrue(serverVersionAtLeast(4, 4)); |
There was a problem hiding this comment.
Why 4.4 only? Isn't this supposed to run against
transactions require a 4.0+ server when non-sharded and 4.2+ when sharded
| void testWriteConcernInheritance() { | ||
| assumeTrue(serverVersionAtLeast(4, 4)); | ||
| try (ClientSession session = client.startSession()) { | ||
| MongoCollection<Document> wcCollection = collection.withWriteConcern(new WriteConcern(0)); |
There was a problem hiding this comment.
| MongoCollection<Document> wcCollection = collection.withWriteConcern(new WriteConcern(0)); | |
| MongoCollection<Document> wcCollection = collection.withWriteConcern(WriteConcern.UNACKNOWLEDGED); |
| try (ClientSession session = client.startSession()) { | ||
| MongoCollection<Document> wcCollection = collection.withWriteConcern(new WriteConcern(0)); | ||
|
|
||
| assertDoesNotThrow(() -> { |
There was a problem hiding this comment.
I'm not sure at which point this test asserts that the write concern is not inherited ?
| + "and normal server selection is performed for the next operation.") | ||
| @Test | ||
| void testNewTransactionUnpinsSession() { | ||
| abort("There is no ability to get the server address with the reactive api"); |
There was a problem hiding this comment.
Why can't we call ?
client.getClusterDescription().getServerDescriptions()2 participants
Ensuring collection write concern options are ignored when inside a transaction.
JAVA-5684