JIRA: https://jira.spring.io/browse/INT-3994
If a header mapper is injected into the mail receiver, a mapped
message is result instead of a message with the raw `MimeMessage`.
`MimeMessage` properties are mapped as discrete headers; in addition
the raw email headers are provided as a multivalue map in the headers.
Handle Multipart
Since a Multipart content has a reference to the original MimeMessage,
convert to a simple byte[] by default.
Add an option in case the user wants to map the message but still
interpret the Multipart intact.
Namespace and Docs
Polishing
Polishing - PR Comments
JIRA: https://jira.spring.io/browse/INT-3991
* Add `IntegrationGraphController` `@RestController` over `IntegrationGraphServer` bean
* Add `@EnableIntegrationGraphController` and `<int-http:graph-controller>` to register
`IntegrationGraphController` if the `DispatcherServlet` is in classpath.
* Also register the `IntegrationGraphServer` bean from the same place, if that isn't presented in the application context yet.
* Allow to configure the "root" `path` for the `IntegrationGraphController` as a placeholder value via
`spring.integration.graph.controller.request.mapping.path` property
* Add tests for both `@EnableIntegrationGraphController` and `<int-http:graph-controller>` cases based on the `MockMvc`
JIRA: https://jira.spring.io/browse/INT-3989,
https://jira.spring.io/browse/INT-3990,
https://jira.spring.io/browse/INT-3988
INT-3989: Add `FileReadingMessageSource.WatchServiceDirectoryScanner`
* Deprecate top-level `WatchServiceDirectoryScanner` because of inconsistency around `Lifecycle` and shared `directory` property
* Copy/paste its logic into the `FileReadingMessageSource.WatchServiceDirectoryScanner` to hide that inconsistency, but still get a gain from the `WatchService` benefits
* Add support for the `StandardWatchEventKinds.ENTRY_MODIFY` and `StandardWatchEventKinds.ENTRY_DELETE` events in the `FileReadingMessageSource.WatchServiceDirectoryScanner`
* Introduce `useWatchService` option to switch to the internal `FileReadingMessageSource.WatchServiceDirectoryScanner`
* Make `CompositeFileListFilter` also as `ResettableFileListFilter`
* Deprecate weird `FileReadingMessageSource.onSend()` method and remove its usage from tests
* Modify `WatchServiceDirectoryScannerTests` for the new logic
* Document changes
Add `MODIFY` and `DELETE` test coverage
Optimize the `filesFromEvents()` logic replacing item with the fresh event sources.
Remove the item from the result set in case of `DELETE` event, because file removal generates both `MODIFY` and `DELETE` events.
* Add `FileReadingMessageSource.setWatchEvents` to allow to listen to the specific events,
not only `CREATE` or all of them.
* Modify tests and docs to reflect the new API
Address PR comments
* Improve `FileReadingMessageSource.setWatchEvents()`
* Add `file.exists()` before adding file from event.
Add more DEBUG logs into the `WatchServiceDirectoryScanner`
With the fact of those logs provide more optimizations:
* Don't register the same directory for watching:
- use the `ConcurrentMap<Path, WatchKey> pathKeys` to track registrations
- any modification within the directory causes the `ENTRY_MODIFY` for the directory as well.
So, skip such an event exactly for the directory during `walkDirectory`
* Add debug logs in case of `ENTRY_DELETE`
Since acks may arrive after the timeout and therefore won't be
delivered to the ackChannel, purge the channel to clear it
instead of waiting for the acks, before stubbing the template.
https://build.spring.io/browse/INT-B43-164/
* Fix CheckStyle for `System. out .println(` expression commenting out such a code line
* Fix generics in the `JdbcMessageStoreTests`
* Increase receive timeout in the `FileMessageHistoryTests`
* Change the `AsyncAmqpGatewayTests` logic to deal with the same channel for both `ack/nack`.
The `timeout` case without reply may end up with the `Consumer cancel`, therefore `nack` not `ack` like we expected
JIRA: https://jira.spring.io/browse/INT-3387,
https://jira.spring.io/browse/INT-3806
* Introduce
```
MessageGroupStore
void addMessagesToGroup(Object groupId, Message<?>... messages);
```
And implement it in all stores.
* Use new `addMessagesToGroup` where it is reasonable, e.g. `DelayHandler`
* Optimize test-case to use a new store method (where it is possible)
* Fix timing delays in the `JdbcMessageStoreTests`
* Introduce `PersistentMessageGroup`
* Add `AbstractMessageGroupStore#proxyMessageGroupForLazyLoad` to wrap the raw `MessageGroup` to the `PersistentMessageGroup` for lazy-load
* Rework `MessageGroupMetadata` do not be `immutable` and allow to store/restore in the `AbstractKeyValueMessageStore` only the `MessageGroupMetadata`
* Refactor `ResequencingMessageHandler` and `SequenceSizeReleaseStrategy` a bit for better performance when interact with the `MessageGroup`
* Add `AbstractMessageGroupStore#setLazyLoadMessageGroups` to switch off the `lazy-load` behavior and restore the previous full `MessageGroup` logic
* Add `What's New` note and `message-store.adoc` paragraph for the lazy-load functionality
`GroupType.PERSISTENT` and not lazy by default
PR Comments
Fix `JdbcMessageStoreTests` timing issues
Address PR comments
* Add performance test to the `ConfigurableMongoDbMessageGroupStoreTests`
* Add JavaDocs for the `MessageGroupFactory` methods
* Add `log4j.properties` into the `test` MongoDB module for better traceability
* Fix `JdbcMessageStore#getOneMessageFromGroup()` over the `doPollForMessage()` delegation.
The `jdbcTemplate.queryForObject()` requires exactly one and only one raw in `resultSet`
* Add performance test results into the `message-store.adoc`
JIRA: https://jira.spring.io/browse/INT-4001
Since all other Spring Framework components use the Jackson's `ObjectMapper` configuration by default as:
```
this.objectMapper.configure(MapperFeature.DEFAULT_VIEW_INCLUSION, false);
this.objectMapper.configure(DeserializationFeature.FAIL_ON_UNKNOWN_PROPERTIES, false);
```
It would be better to be consistent in Spring Integration, too.
* Add mention options to the `Jackson2JsonObjectMapper` default's `ObjectMapper`
* Add on more constructor into the `Jackson2JsonMessageParser` to allow to inject customized `Jackson2JsonObjectMapper`
* Modify a couple tests to reflect the change
* Reformat code style in the `AbstractJsonInboundMessageMapperTests` for better readability
**The Migration Guide note follows up**
JIRA: https://jira.spring.io/browse/INT-3999
Since the scheduled tasks may live for a long time it can finish
with the `OutOfMemory` if we use the direct reference to big objects, like `Message<?>`.
* Fix `AbstractCorrelatingMessageHandler` to deal only with the `groupId`
from the `scheduleGroupToForceComplete()` when we `schedule` `Runnable` for the `forceRelease` logic.
* Fix `DelayHandler` to deal only with `messageId` in the `releaseMessageAfterDelay()`, when we
`schedule` `Runnable` for the `releaseMessageAfterDelay`.
* Since the logic hasn't been changed for those components, there is no any new test.
There is just enough to be sure that all existing tests are fine.
**Cherry-pick to 4.0.x, 4.1.x, 4.2.x**
Optimise the release task for the `SimpleMessageStore` case
JIRA: https://jira.spring.io/browse/INT-3998
Do not modify interceptors from the `AbstractMessageChannel`.
They must be declared as beans, too.
Fix [UnusedImport] in the `EnableIntegrationTests`
Doc Polishing
https://build.spring.io/browse/INT-SONAR-1700/
We set up some stubbing on the `AsyncAmqpTemplate` to test nacks
which are not easy to generate on a real broker.
However, some acks might not have been consumed yet - this can cause
partial stubbing errors, depending on timing.
Be sure to consume all acks before stubbing the template.
JIRA: https://jira.spring.io/browse/INT-3996
One of the query provider implementations was missing. To avoid a maintenance
nightmare, simply reference the interface's javadoc which lists all known
implementations.
* Simple `JdbcChannelMessageStore` JavaDocs polishing
* Add `LogAdjustingTestSupport` to the `StompIntegrationTests` to trace failures in the future.
* Also change the test `Start/Stop` messages to the `WARN` level in the `LogAdjustingTestSupport`, because `DEBUG` doesn't work everywhere.
Maybe some logging systems mix misconfiguration, like we have in the WebSocket modules with the Embedded Tomcat
JIRA: https://jira.spring.io/browse/INT-3980
When we declare our components via `@Bean` (or `<bean>`) the requirements for
the `@Component` (`@MessageEndpoint`) looks redundant and sometimes even dangerous,
when we use `@ComponentScan`, too.
* Remove the `@Component` restriction logic from the `MessagingAnnotationPostProcessor`
* Ensure that tests pass
(We may consider to backport afterwards)
Add `requireComponentAnnotation` logic
* Introduce `spring.integration.messagingAnnotations.require.componentAnnotation` property for `spring.integration.properties` as `false` by default
* Add `MessagingAnnotationPostProcessor#setRequireComponentAnnotation()` to accept the value from the `spring.integration.messagingAnnotations.require.componentAnnotation`
* Fix `integraton` typo everywhere
* Document the change in the `What's New` and in the `configuration.adoc`
* Ensure that logic works properly (no messaging endpoints populated) with the `spring.integration.messagingAnnotations.require.componentAnnotation = true` and "unannotated" `AnnotatedEndpoint2` in the `AnnotatedEndpointActivationTests`
This PR also fixes: https://jira.spring.io/browse/INT-3962
Fix issues according Travis report
Doc Polishing
JIRA: https://jira.spring.io/browse/INT-3979
Previously we had a to instantiate all messages from the `MessageBuilder`
if we would like to return messages as splitted items.
For example to populate some item-specific headers.
From other hand the `MessageBuilder` as an item incorrectly remained as a `payload`.
* Add condition logic to the `AbstractMessageSplitter` to properly handle `AbstractIntegrationMessageBuilder`
and don't create an extra `Message`
* Apply the same logic to the `AbstractAggregatingMessageGroupProcessor`
(We may consider to backport it later)
JIRA: https://jira.spring.io/browse/INT-3984
The `PropertySourcesPlaceholderConfigurer implements BeanFactoryPostProcessor, PriorityOrdered`
meaning that it is run before any other regular `BeanFactoryPostProcessor`,
like the `IntegrationConfigurationBeanFactoryPostProcessor` has been before.
With such an order any `BeanDefinition` declaration from the `IntegrationConfigurationBeanFactoryPostProcessor` process
ended up without `PP` resolutions.
* Rework `IntegrationConfigurationBeanFactoryPostProcessor` to the `BeanDefinitionRegistryPostProcessor`,
which makes it be loaded as early as possible. See `PostProcessorRegistrationDelegate`:
````java
// Now, invoke the postProcessBeanFactory callback of all processors handled so far.
invokeBeanFactoryPostProcessors(registryPostProcessors, beanFactory);
invokeBeanFactoryPostProcessors(regularPostProcessors, beanFactory);
````
* Modify `EnableIntegrationTests` and `ChannelSecurityInterceptorSecuredChannelAnnotationTests` to ensure that the change
meets the `PP` and `SpEL` resolutions during bean definition phase.
* Fix the timing issue in the `JdbcMessageStoreChannelIntegrationTests`
JIRA: https://jira.spring.io/browse/INT-3978
Previously the `MessageHistory.write()` always used `MessageBuilder` to populated the `history` header.
Since the `MessageBuilder` doesn't care about the `source` message type, we might lose some important information,
like `AdviceMessage.inputMessage`
* Add conditional logic into the `MessageHistory` for all known `Message<?>` implementations
* Provide appropriate tests in the `MessageHistoryTests` for known `Message<?>` implementations
* Make `AdviceMessage` generic and fix all affected code on the matter
* Since `MutableMessage` is public already, fix `MongoDbMessageStore.DBObjectToMutableMessageConverter` to use
`MutableMessage` type directly.
Add WARN for non-standard `Message` type
JIRA: https://jira.spring.io/browse/INT-3909
Un`@Ignore` the `ClientWebSocketContainerTests` to ensure that `Tomcat-8.0.33` has the fix for the closed WebSocket session race condition.
Fix `StompIntegrationTests` test data
Clean up queue channels before each test, because of some remained extra messages during the web-socket protocol interactions,
like socket close events or similar
Fix `RedisStoreInboundChannelAdapterIntegrationTests` race conditions
Since Jedis driver is async, we not always receive a reply from the `TX-sync` component immediately.
And loop checking the Redis for the `key` after some `Thread.sleep()`
Fixes https://build.spring.io/browse/INT-MJATS41-593
* Fix `LoggingHandler` JavaDoc
* Fix `JdbcOutboundGatewayParserTests` timing and race condition issues
* Add `mock(BeanFactory.class)` to the `JpaOutboundChannelAdapterTests` for `JpaExecutor`
* Fix `JpaOutboundChannelAdapterTests` for the incorrect transaction usage
JIRA: https://jira.spring.io/browse/INT-3981
Since the `MessageHandlerChain` uses a `ReplyForwardingMessageChannel` in the end of invocation,
we don't need to have add an extra `<bridge>` for shifting just populated `replyChannel` header.
* Add a couple tests to prove that
* Fix the typos in the `ChainParserTests`
* Remove unnecessary `DirectFieldAccessor` usage in the `MessageHandlerChain`
checkstyle EmptyBlock
checkstyle fixRightCurly Script
checkstyle EmptyStatement
checkstyle RightCurly
checkstyle TailingWhite
checkstyle NeedBraces
Fix the line separator in the `fixRightCurly.gradle`
JIRA: https://jira.spring.io/browse/INT-3977
* Add `Level` ctor
* Add `setLogExpression(Expression)` and `setLogExpressionString(String)`
* Deprecate existing `setExpression(String)` in favor of those new
* Some refactor for redundant code around `evaluationContext`
* Fix tests according a new `LoggingHandler` logic
* Add JavaConfig sample to the Reference Manual
Update note on channel interceptors in docs
Update `channel.adoc` including `afterReceiveCompletion(..)` in the
list of methods that are not invoked when an interceptor is applied
to `PollableChannel`
Improve `aggregator.adoc` formatting in docs
Fix and improve `channel.adoc` formatting in docs
Convert tabs to spaces of `codec.adoc` in docs
Convert tabs to spaces and fix formatting of `content-enrichment.adoc` in docs
Convert tabs to spaces and fix formatting of `logging-adapter.adoc` in docs
Convert tabs to spaces on `metrics.adoc` in docs
Improve `resequencer.adoc` formatting in docs
Improve `splitter.adoc` formatting in docs
* Fix more typos in the Reference Manual
fixes
fixModifiers after fixFinal
Revert CachingSessionFactory
Class is spied in tests.
checkstyle - Import Rules
checkstyle InterfaceIsType
checkstyle InnerTypeLast
checkstyle OneStatementPerLine
CovariantEquals
OneTopLevelClass
* Revert `'\n'` -> `System.lineSeparator()` in the Gradle scripts to meet Git `autocrlf = true` on Windows
* Fix timing issue with the `LastModifiedFileListFilterTests`, when the `age = 1` might not be enough for the file object when we have some delay before checking
JIRA: https://jira.spring.io/browse/INT-3976
The class name in case of `StaticScriptSource` (for inline scripts) is required only for Groovy native engine.
Even if `GroovyScriptParser` provides an explicit `scriptClassName` option, we still can configure
`GroovyScriptExecutingMessageProcessor` with the `StaticScriptSource` from Java without the script class name.
Therefore the logic in the `GroovyScriptParser` isn't robust.
* Remove the `GroovyScriptParser` `scriptClassName` logic altogether
* Change the `GroovyScriptExecutingMessageProcessor` to rely on the Groovy native `scriptClassName` generation
in case of the `scriptSource` is provided without the class name option
* Fix the test for the actual generated script class name
JIRA: https://jira.spring.io/browse/INT-3975
Also, switch the default mapping to map all inbound headers, but
don't map `x-*` on outbound (by default).
This will provide apps access to important headers such as `x-death` by
default, while not propagating dangerous headers outbound.
Polishing - PR Comments
Simple code style polishing
JIRA: https://jira.spring.io/browse/INT-3591
Modify `TestFtpServer` and `FtpServerOutboundTests` for file with leading whitespace
to confirm the `FtpServerFactory` has been fixed in the `commons-net-3.4`
JIRA: https://jira.spring.io/browse/INT-3944
Polishing (PR comments) and Doc Polish
Polishing - PR Comments
Remove custom `async` boolean in favor of the now public setter.
Add exception to `JmsException` hierarchy.
INT-3944: Polishing
* Rename `asyncReplySupported` just to `async`. As well as its getter and setter. Plus fix docs on the matter
* Polishing for the `JmsOutboundGateway` according PR comments
* Rework `JmsOutboundGateway` to return `AbstractIntegrationMessageBuilder` instead of `Message`
since `AbstractMessageProducingHandler` rebuilds `Message` for a new one to copy headers from request
* Add `getPayload()` and `getHeaders()` to the `AbstractIntegrationMessageBuilder` to make the `Routing Slip`
users happy, when the `AbstractIntegrationMessageBuilder` is pushed to the `Routing Slip path` function
* Fix race condition in the `ChatMessageListeningEndpointTests`: the `stanza` parsing is done in the different Thread.
Doc Polishing (Routing Slip)
Fix timing issue in the `LastModifiedFileListFilterTests`:
`Thread.sleep()` not always reflects the reality.
Switch to the "past simulation" via explicit `File.setLastModified()`
