JIRA: https://jira.spring.io/browse/INT-4246
The `LockRegistryLeaderInitiator` currently does not revoke leadership
when leading instance is unable to acquire lock from the underlying lock.
This can result in multiple `LockRegistryLeaderInitiator` instances becoming
leaders in the exceptional situations such as lock timeouts.
* Handle leadership revoking when leading `LockRegistryLeaderInitiator`
is unable to acquire lock
Polish javadoc
**Cherry-pick to master**
JIRA: https://jira.spring.io/browse/INT-4243
Remove newline char from key exchange.
Fix CR and Add Windows Delete Diagnostics
Fix For Windows
More diagnostics.
It appears the new server doesn't close the file when the session is closed.
In the streaming test, consume the stream before closing the session.
Polishing - PR Comments
* More polishing according PR comments
Conflicts:
spring-integration-sftp/src/test/java/org/springframework/integration/sftp/SftpTestSupport.java
spring-integration-sftp/src/test/java/org/springframework/integration/sftp/dsl/SftpTests.java
spring-integration-sftp/src/test/java/org/springframework/integration/sftp/session/SftpRemoteFileTemplateTests.java
spring-integration-sftp/src/test/java/org/springframework/integration/sftp/session/SftpServerTests.java
spring-integration-sftp/src/test/java/org/springframework/integration/sftp/session/SftpSessionFactoryTests.java
JIRA: https://jira.spring.io/browse/INT-4238
Revert to using the sync client in the message-driven adapter so we can detect
subscription failures (the sync client throws an exception).
The only reason to use the async client was to timeout disconnects; this can
be achieved with the sync client and `disconnectForcibly`.
Also, the subscribe method updates the qos argument with the granted QOS values.
Detect and log if any QOS does not match the request.
Polishing
Polishing - PR Comments
Conflicts:
spring-integration-mqtt/src/test/java/org/springframework/integration/mqtt/MqttAdapterTests.java
* Remove all new tests since Paho lib has class signature check, so we can't mock its classes
JIRA: https://jira.spring.io/browse/INT-4229
* Make `SpelPropertyAccessorRegistrar` as `public` class and provide more API to customize it.
This class allows to register `PropertyAccessor` s for shared `EvaluationContext`
* Document how to configure SpEL functions and `PropertyAccessor` s with Java Config
**cherry-pick to 4.3.x**
JIRA: https://jira.spring.io/browse/INT-4237
It was possible to flush (close) the file while a write was in process; more likely when
`flushWhenIdle` is false.
This probably would not occur in the real world, just tests with short flush intervals,
but certainly possible.
There is already a lock used to prevent concurrent writes while appending; use the same
lock when flushing.
Conflicts:
spring-integration-file/src/main/java/org/springframework/integration/file/FileWritingMessageHandler.java
spring-integration-file/src/test/java/org/springframework/integration/file/FileWritingMessageHandlerTests.java
JIRA: https://jira.spring.io/browse/INT-4232
Currently to implement the synchronization logic when remote file is modified there is only one way: remove local file and provide a `localFilter` which can be clear from the removed file as well.
The Reference Manual claims that `FileSystemPersistentAcceptOnceFileListFilter` can let us pick up a fresh version of the remote file if we configure `preserveTimestamp = true`, but actually `AbstractInboundFileSynchronizer` just bypass that logic because it check a local file for existence.
* Modify `AbstractInboundFileSynchronizer` to compare `lastModified` of the remote file with local version if `preserveTimestamp == true`
* Also replace `AcceptOnceFileListFilter` to the `FileSystemPersistentAcceptOnceFileListFilter` since the first one doesn't care about `lastModified`.
If `preserveTimestamp` isn't in use the `FileSystemPersistentAcceptOnceFileListFilter` works fully similar to the `AcceptOnceFileListFilter`
**Cherry-pick to 4.3.x except test-case**
Make `SftpStreamingMessageSourceTests` compatible with Windows
Looks like SFTP embedded server provides different permissions for files on different OSs
Doc Polishing
Conflicts:
spring-integration-file/src/test/java/org/springframework/integration/file/remote/synchronizer/AbstractRemoteFileSynchronizerTests.java
spring-integration-ftp/src/test/java/org/springframework/integration/ftp/dsl/FtpTests.java
spring-integration-sftp/src/test/java/org/springframework/integration/sftp/inbound/SftpStreamingMessageSourceTests.java
src/reference/asciidoc/whats-new.adoc
Resolved.
JIRA: https://jira.spring.io/browse/INT-4227
Expose `MessagingAnnotationPostProcessor` for inheritors.
The `CustomMessagingAnnotationTests` demonstrates how custom Messaging Annotation can be registered and used
**Cherry-pick to 4.3.x**
Add `setupCustomPostProcessors()` and `addMessagingAnnotationPostProcessor` to the `MessagingAnnotationPostProcessor`
Expose some API for inheritors
Conflicts:
spring-integration-core/src/main/java/org/springframework/integration/config/annotation/MessagingAnnotationPostProcessor.java
JIRA: https://jira.spring.io/browse/INT-4236
Add methods to obtain status from the `SmartLifecycleRoleController`.
* Polishing according PR comments.
* Fix log messages in the `SmartLifecycleRoleController` from the Zookeeper mentioning
Conflicts:
src/reference/asciidoc/endpoint.adoc
Resolved.
Also Java 6 Compatibility.
JIRA: https://jira.spring.io/browse/INT-4230
The `get()` method only preserves the timestamp when an LS operation is done within it.
For MGET ops, we don't perform another LS, so the timestamp was not updated.
Add code to the recursive and non-recursive MGET methods to preserve the timestamp if so configured.
Conflicts:
spring-integration-file/src/main/java/org/springframework/integration/file/remote/gateway/AbstractRemoteFileOutboundGateway.java
spring-integration-ftp/src/test/java/org/springframework/integration/ftp/outbound/FtpServerOutboundTests-context.xml
spring-integration-sftp/src/test/java/org/springframework/integration/sftp/outbound/SftpServerOutboundTests-context.xml
JIRA: https://jira.spring.io/browse/INT-4228
Previously, if a bean definition for a channel exists, but has configuration issues,
the `AbstractMethodAnnotationPostProcessor` still went ahead and created a `DirectChannel`.
Only with TRACE logging was the root cause apparent.
This was because all `BeanException` s cause that behavior.
Now, only `NoSuchBeanDefinitionException` will cause auto-creation of channels.
Tested with mocks and a real Boot app, which now correctly reports
```
15:28:41.582 [main] DEBUG o.s.b.d.LoggingFailureAnalysisReporter - Application failed to start due to an exception
org.springframework.beans.factory.NoSuchBeanDefinitionException: No bean named 'rabbitConnectionFactory' available
```
* Resolve Mockito conflicts in the `MessagingAnnotationPostProcessorChannelCreationTests`
JIRA: https://jira.spring.io/browse/INT-3923Fixes: #2046
Previously, when connection is lost, the inbound adapter attempted to reconnect on
a schedule with a fixed delay.
If a connection was again lost, while the schedule is still running, we can end up
with another scheduled task running.
This is benign aside from the DEBUG log noise because the scheduled task tests the
connection before reconnecting.
However, if the `recoveryInterval` is short, it could consume CPU.
Change the reconnect to be a one-time scheduled task and reschedule if it fails to
reconnect.
Synchronize all access to the `connected` field.
Add a test case with a short recovery interval, before this fix, we see many logs
`Attempting reconnect`.
Conflicts:
spring-integration-mqtt/src/main/java/org/springframework/integration/mqtt/inbound/MqttPahoMessageDrivenChannelAdapter.java
spring-integration-mqtt/src/test/java/org/springframework/integration/mqtt/MqttAdapterTests.java
* Fix single-arg `Assert` method usage in the `AbstractCorrelatingMessageHandler`
JIRA: https://jira.spring.io/browse/INT-4221
* Upgrade to those versions of Spring project dependencies which potentially will provide similar fix
**Cherry-pick to 4.3.x**
JIRA: https://jira.spring.io/browse/INT-4217
Previously, SpEL method calls with required header parameters had the following form:
#target.messageAndHeader(message, headers['number'] != null ? headers['number'] : T(org.springframework.util.Assert).isTrue(false, 'required header not available: number'))
The SpEL compiler cannot compile this because the else clause of the ternary has no `exitDescriptor` to indicate the type.
Change the expression to use a function for required headers.
Also use an Elvis operator when possible.
Some examples of new expressions:
#target.optionalAndRequiredHeader(headers['prop'] ?: null, #requiredHeader(headers, 'num'))
#target.optionalAndRequiredDottedHeader(headers['dot1'] != null ? headers['dot1'].foo : null,
#requiredHeader(headers, 'dot2').baz
In the second case, we can't use an Elvis because we're accessing a `foo` property of the header.
__cherry-pick to 4.3.x__
- make `ParametersWrapper` `static`
- minor conflicts in imports
- remove the perf test
* Polishing: remove redundant annotation args for their default values usage
* Revert `java.util.Optional` import in favor of literal for Java < 8
* Make `AnnotatedTestService` as `public` to get compilable SpELs
* Remove performance test from the `MethodInvokingMessageProcessorTests`
JIRA: https://jira.spring.io/browse/INT-4216
Add test for running empty array through `ChainFileListFilter`
Assure no exceptions are throw when filter input is empty.
INT-4216: Add fix for ChainFileListFilter edge case
INT-4216: Fix checkstyle error on test class
Add copyright header to test file
* Polishing for `ChainFileListFilter` JavaDocs
* Move `CompositeFileListFilterTests` and `ChainFileListFilterIntegrationTests` to the `filters` package instead of `file` core
(cherry picked from commit 1f9d07a)
JIRA: https://jira.spring.io/browse/INT-4212
Add an option to flush after the `flushInterval`, regardless of intermediate writes.
Rename `lastFlush` to `firstWrite`
Polishing
**Cherry-pick to 4.3.x**
JIRA: https://jira.spring.io/browse/INT-4202
The `StoredProcOutboundGateway` uses `MessageBuilder` directly for procedure result without any conditions.
If procedure result is `null`, `MessageBuilder.withPayload` throws `java.lang.IllegalArgumentException: payload must not be null`.
* Since the super `AbstractReplyProducingMessageHandler` class takes care about `null` reply properly via its `requiresReply` property and really uses `MessageBuilder`
for reply, just fix `StoredProcOutboundGateway` to return procedure result as is from the `handleRequestMessage()` implementation
* Fix some typos and code style
* Increase some JDBC tests performance changing `Thread.sleep()` solution to proper `PollableChannel.receive()` or iterations over expected result
**Cherry-pick to 4.3.x**
Update `@Copyright` to 2017
Make `StoredProcOutboundGateway` as `requiresReply = true` by default
Revert `setRequiresReply(true)` change
Specify `setRequiresReply(true)` only if `expectSingleResult == true` and `setRequiresReply()` hasn't been called explicitly
Conflicts:
spring-integration-jdbc/src/main/java/org/springframework/integration/jdbc/StoredProcOutboundGateway.java
spring-integration-jdbc/src/test/java/org/springframework/integration/jdbc/JdbcMessageStoreChannelIntegrationTests.java
spring-integration-jdbc/src/test/java/org/springframework/integration/jdbc/config/StoredProcOutboundGatewayParserTests.java
Resolved.
JIRA: https://jira.spring.io/browse/INT-4203
Add channel names for easier use in DSL.
Add documentation (boot) example.
Normalize Expression Setters; add Javadocs
Polishing - SPR-15091
Tiny code style polishing
Conflicts:
spring-integration-core/src/test/java/org/springframework/integration/dsl/flows/IntegrationFlowTests.java
src/reference/asciidoc/handler-advice.adoc
spring-integration-core/src/test/java/org/springframework/integration/handler/advice/ExpressionEvaluatingRequestHandlerAdviceTests.java
* Restore removed methods in the `ExpressionEvaluatingRequestHandlerAdvice` and deprecate them
* Remove methods in the `ExpressionEvaluatingRequestHandlerAdvice` introduced since `5.0`
JIRA: https://jira.spring.io/browse/INT-4197
Since `Collection<Message<?>>` can be possible only if we are dealing with messages as a group (`MethodInvokingMessageListProcessor`), the case with the `List<?>` param for Messaging Annotation method should not be treated as candidate for `messages` collection.
* Add `this.canProcessMessageList` condition to avoid `messages` SpEL expression when we are not in the `MethodInvokingMessageListProcessor` environment
**Cherry-pick to 4.3.x**
* Fix Array creating formatting
* Add comment about `ReflectiveMethodExecutor` to the `testRouterWithListParam()`
* Send one test data as as `Collection` to be sure that `@Router` parameters mapping works well
JIRA: https://jira.spring.io/browse/INT-4201
name<->value alias was not processed correctly so using
`@Header(name = "baz") String baz` did not work.
Use `SynthesizingMethodParameter`.
(cherry picked from commit 0cbfd6e)
JIRA: https://jira.spring.io/browse/INT-4198
Enable setting properties like `needClientAuth` on the `SSLEngine` - when not using
NIO, this can be set on the server socket with a socket support implementation.
Add `nio-connection-support` to namespace.
Improved "Advanced Techniques" documentation, using this use case as an example.
Fail fast with NIO when SSL handshaking fails.
Polishing - PR Comments
More Polishing
* Final polishing
- fix several typos in log messages
- clean up `TcpConnectionFactoryFactoryBean` JavaDocs from redundant imports
- remove redundant `InitializationBean` functionality from the `DefaultTcpNetSSLSocketFactorySupport` as well
Conflicts:
spring-integration-ip/src/test/java/org/springframework/integration/ip/tcp/connection/SocketSupportTests.java
* Resolve Java 8 code style and revert `afterPropertiesSet()` with `@Deprecated`
Since we can have a result message in the `replyChannel` before the TX commit, a subsequent adapter `stop()` causes the current task interruption,
therefore TX rallback.
That's why we see `semaphore.acquire()` interruption and don't see message in the `afterCommitChannel`
**Cherry-pick to 4.3.x**
(cherry picked from commit bdab0aa)
There is tiny time window when `LockRegistryLeaderInitiator` can be stopped during `Context.yield()` invocation.
In this case the `LockRegistryLeaderInitiator` goes to the stopped state, but a new `leaderSelector` is submitted for election.
Therefore in the `LockRegistryLeaderInitiatorTests.competing()` the second initiator may not get be granted because there is uncontrolled `leaderSelector` task on background.
We can overcome it with always `this.initiator.stop()` for the `LockRegistryLeaderInitiatorTests`, but to be sure in the fix it would be better to leave as is.
Also add `@Rule Log4jLevelAdjuster` for future diagnostics
**Cherry-pick to 4.3.x**
(cherry picked from commit 56929cd)
