JIRA: https://jira.spring.io/browse/INT-3423
Consider a message containing "one one" and a splitter
with expression "payload.split(' ')", only one message
is emitted.
The default Collection created by SpEL is a HashSet.
Change the ExpressionEvaluatingSplitter to expect a
List instead of a Collection.
Conflicts:
spring-integration-core/src/test/java/org/springframework/integration/splitter/SpelSplitterIntegrationTests.java
JIRA: https://jira.spring.io/browse/INT-3410
When running a fixed thread pool with a bound queue,
and CALLER_RUNS execution rejection policy, it
was possible to deadlock the IO selector thread.
Add a `CompositeExecutor` to use different threads for
IO to those used for message assembly.
Add a `CallerBlocksPolicy` to block the invoking thread
(for a specified time) if the pool is exhausted.
Add documentation.
Polishing
Also add another control-bus test.
Conflicts:
spring-integration-core/src/test/java/org/springframework/integration/config/xml/ControlBusTests.java
Resolved.
JIRA: https://jira.spring.io/browse/INT-3316
The `SimpleTypeConverter` delegate uses property editors
which are not thread safe. We protect against concurrent
access when the source is not a String, but when a String,
we delegate to the `STC`.
Synchronize access to the `STC.convertIfNecessary()` method.
JIRA: https://jira.springsource.org/browse/INT-3298
Emit a WARN log whenever the `ReleaseStrategyFactoryBean`
falls back to using the `SequenceSizeReleaseStrategy`.
This will happen if no method name is provided and no
annotated method is found, or a null reference is provided.
JIRA: https://jira.springsource.org/browse/INT-3300
The change introduced by https://jira.springsource.org/browse/INT-3147
(Registration of `MessageSource<?>` as a bean with suffix `.source`) broke
the case where `MessageSource<?>` is a reference to an existing bean.
Add check to the `AbstractPollingInboundChannelAdapterParser` for the type
of `source` and apply the appropriate logic.
Conflicts:
spring-integration-core/src/test/java/org/springframework/integration/config/ChannelAdapterParserTests.java
Resolved.
JIRA: https://jira.springsource.org/browse/INT-3252
Remove the use of `SecureRandom.generateSeed()` which is not
supported on some platforms.
Note: While this change does not reflect the custom IBM
code noted in the JIRA issue, it does revert the initial
SecureRandom generation to use the same code as
UUID.randomUUID(), so the expectation is that this will
work fine on zOS.
JIRA: https://jira.springsource.org/browse/INT-3240
Spring Integration 3.0 introduced registration of `MessageSource`s as beans
with an `id` based on the adapter `id` or, as a fallback, generated from
the `MessageSource` class name and the suffix `.source`.
But since that bean wasn't registered as a bean with the generated name (because the suffix was added),
subsequent elements get the same source bean name. This meant that
several anonymous `inbound-channel-adapter` with the same type (`file:`, `ftp:`, `twitter:` etc.)
caused an issue, when the same `MessageSource` bean was used by several `SourcePollingChannelAdapter` beans.
Fix the issue for fallback name generation by using the `SourcePollingChannelAdapterFactoryBean` class name.
Since that is registered, the source will always get a unique name, regardless of whether an id is provided.
Remove warning from BFCR if there is no HeaderChannelRegistry in
the bean factory. Change the log to DEBUG and remove the stack
trace.
Instead, update the Exception message when a channel name can't be
resolved to a channel to indicate that there is no registry.
JIRA: https://jira.springsource.org/browse/INT-3248
JIRA: https://jira.springsource.org/browse/INT-3235
Unable to programmatically wire header expressions or method-specific
properties into a GatewayProxyFactoryBean.
The GPFB has public methods that take arguments of
this type.
Make the class public instead of package-protected.
Add a test case.
JIRA: https://jira.springsource.org/browse/INT-3227
In MessagingTemplate.TemporaryReplyChannel, it was possible
the `clientHasReceived` boolean could be set by the other
thread after normal message receipt, but before the sender
tests it for the purpose of emitting a duplicate reply log
message.
Capture the boolean value of `clientHasReceived` before
counting down the latch.
JIRA: https://jira.springsource.org/browse/INT-3006
* Add integration properties `channels.maxUnicastSubscribers` and `channels.maxBroadcastSubscribers`
with default value to `Integer.MAX_VALUE`
* Add SpEL value resolution for those default values from parsers
* Deprecate similar properties from `ChannelInitializer`
INT-3006 `maxSubscribers` from channels' onInit()
* Move default init for `maxSubscribers` to the channels' `onInit()`
from `integrationProperties`, not from parsers
* extract `default` `integrationProperties` from static block in the `IntegrationProperties`
and use them as `default` on `IntegrationContextUtils#getIntegrationProperties`
* Add bean definition expression building to get the default values for `integration properties`
on bean building phase.
INT-3006: Add `taskScheduler.poolSize` property
INT-3006: Polishing
JIRA: https://jira.springsource.org/browse/INT-3216
Previously, After a group has been selected for completion
by the reaper, if the group is otherwise completed before
the reap AND expireGroupsOnCompletion is true AND a new
group with the same ID is created, ALL in the same millisecond,
AND the group creation timestamp is used for the reap criteria,
the new group will be incorrectly reaped.
Also, for a simple MGS, it was not necessary to re-fetch the group
if the group has already been marked complete.
- Defer reaping if the group is already complete (unless it's empty).
- Verify the timestamp hasn't changed (indicating a new group)
Add tests
- Verify no refetch of a complete SMG
- Verify no reap of a complete group (after refetch)
- Verify no reap when the timestamp changed
JIRA: https://jira.springsource.org/browse/INT-3055
* Add to HTTP Outbound Endpoint `uri-variables-expression`
* Add mutually exclusive check for `uri-variables-expression` with `<uri-variable>`
* Tests and Docs
INT-3055: Fixes
* `ExpressionEvalMap` now can apply `Map<String, Object>`,
but the value must have a `String` or `Expression` type.
Polishing
INT-3055 Doc Polishing
JIRA: https://jira.springsource.org/browse/INT-3213
Backwards reference from ....support.channel to
....channel.registry (BeanFactoryChannelResolver).
- Move the registry interface to ....support.channel
- Move the implementation to ....channel
JIRA: https://jira.springsource.org/browse/INT-3212
INT-3212 Fix Failing TCP Test
TcpNioConnections can deadlock when a fixed thread pool is being used.
The deadlock is detected after 60 seconds, but the test case failed
after 20 seconds - before the deadlock was detected.
Change the test to use a cached thread pool instead.
Also, increase concurrency by using threading in the test server
so each socket is handled on a separate thread.
Enchance connectionId to include both local and remote ports to
aid debugging.
Some minor formatting corrections.
JIRA: https://jira.springsource.org/browse/INT-3199
* Set `TypeDescriptor` to `Void.class` as default `targetParameterType` for falling-back
* Change `fall-thru` only to one method:
* Extract generic type from `Massage` parameter, so now
``` m1(Message<String> message);
m2(String payload);
```
are equal by `payload type` and similar POJO causes ambiguity configuration exception
INT-3199: Polishing and further fixing
* Address PR's comments
* Provide more friendly exception message for `Void.class` method
* Remove the logic around `@ServiceActivator`:
- one part was fixed within INT-3114
- another for `RequestReplyExchanger#exchange` doesn't make sense as it is good candidate
* Add fallback to `Iterator.class` method for `Iterable` payloads
* Add tests
Polishing
INT-3199: Add `handlerMessageMethods`
* Introduce separate properties:
- `handlerMessageMethods` for methods with `Message` parameter
- `handlerMethod`, if there is only one candidate after configuration
* Polishing several tests according new state of `MessagingMethodInvokerHelper`'s properties
Polishing
Revert `RequestReplyExchanger` fallback
Add `fallbackMessageMethods` processing
INT-3199 Polishing
Only revert to RequestReplyExchanger if there are ambiguous
fallback methods - if the target object is an RRE, it will always
have a candidate fallback (exchange).
Add more RRE tests.
Make ParametersWrapper a public inner class - it's only used
within the context of a process() call. Making it public allows
SpEL to access its properties without reflection.
Allows reply channel resolution after a message has been serialized
somewhere in a flow. Previously, the reply channel was lost.
With this change, the reply channel can be registered and the header
becomes a string (channel name) that can be serialized.
JIRA: https://jira.springsource.org/browse/INT-2269
INT-2269 Rename Registry to HeaderChannelRegistry
- Extract interface
- DefaultHeaderChannelRgistry
INT-2269 Polishing; PR Comments
- Add errorChannel registration as well.
INT-2269 HeaderChannelRegistry - Fix
The internal BridgeHandler in MessagingGatewaySupport did not have
a channel resolver. When a reply was explicitly routed to the gateway's
reply channel, the String representation of the reply channel could
not be resolved to a channel.
Set the BeanFactory on the bridge handler.
Add tests.
INT-2269 HeaderChannelRegistry - Fix JMS/Enricher
The JMS inbound gateway and ContentEnricher instantiate a
MessagingGatewaySupport internally it does not get a
reference to the BeanFactory. This means that the internal
BridgeHandler cannot resolve the String representation of
the reply channel to a channel.
Make ChannelPublishingJmsMessageListener BeanFactory aware, and
propagate the bean factory to the MGS.
Set the MGS bean factory in the ContentEnricher (which is already
BFA).
INT-2269: Polishing
Enable the use of multiple JSch channels over a
single connection.
JIRA: https://jira.springsource.org/browse/INT-3047
INT-3047: Polishing locks logic
INT-3047 SFTP Reset Cache; Quiesce Shared Sessions
Support resetCache() on the CachingSessionFactory
- reset a shared JSch session so it is reestablished on next use
- immediately close idle sessions
- close in-use sessions as they are returned
- close the channel when an SftpSession is closed and a shared JSch is being used
- physically close a shared JSch session only when the last channel is closed
Polishing
Forgot to save the ftp.xml.
Change CachedSession.epoch to created.
INT-3047 Polishing - PR Comments
- Also add missing docs for `preserve-timestamp`
INT-3047 Polish - PR Comments
Also change epoch to nanoseconds and use a simple != comparison with the
epoch in which a session was created.
JIRA: https://jira.springsource.org/browse/INT-3017
* Add parsers tests and integration tests
* Polishing Redis XSD
* Fix `JmsMessageDrivenEndpointParser` `LifeCycle` attributes
* Catch `RedisSystemException` after `this.boundListOperations.rightPop`.
It maybe an exception about 'connection closed'.
If the `RedisQueueMessageDrivenEndpoint` is 'active' this exception is rethrown,
otherwise just logged under error category.
* Add Redis queue components docs
JIRA: https://jira.springsource.org/browse/INT-3019
INT-3017 Doc Polishing
Previously, `MetadataStore` couldn't be configured for Twitter Adapters
- only a global one could be used.
The `metadataKey` was generated automatically with a 'difficult' value.
* Register all `MessageSource` for `SourcePollingChannelAdapter`
as beans with id based on adapter id and prefix '.source' (INT-3147)
* Polishing parser to get rid of explicit `MessageSource` beans. (INT-3147)
* Make Feed and Twitter adapters `id` attribute as required -
now it presents a `metadataKey` for `MetadataStore` (INT-3147)
* Add to Twitter adapters a reference attribute for `MetadataStore` (INT-3173)
* Add Twitter adapters `poll-skip-period` attribute (INT-3167)
* Add and implement `MetadataStore#remove` (INT-1941)
* Make `MetadataStore` as `@ManagedResource` (INT-1941)
* Polishing tests
JIRAs:
https://jira.springsource.org/browse/INT-3147https://jira.springsource.org/browse/INT-3167https://jira.springsource.org/browse/INT-3173https://jira.springsource.org/browse/INT-1941
INT-3147: Polishing and fixes
* add domain suffix to `metadataKey`
* change contract of `MetadataStore.remove`
* remove timeout window from `AbstractTwitterMessageSource`
* polishing and fix `SearchReceivingMessageSourceWithRedisTests`
INT-3147: Rebasing and polishing
INT-3147: fix 'metadata' package tangle
INT-3147 Doc Polishing
* Add `AmqpHeaders` headers
* Add `JavaType` headers mapping to `DefaultAmqpHeaderMapper`
* Add ability to Json Transformers to add/extract `JavaType` headers
* Make `JsonToObjectTransformer#targetClass` as non-required and do fallback
to extract type from `JavaType` headers
* Add `JavaType` extraction in the `JsonObjectMapper` implementors
JIRA: https://jira.springsource.org/browse/INT-2809
INT-2809: Polishing and refactoring
* Introduce `JsonHeaders`, `AbstractJacksonJsonObjectMapper`
* Move `TestPerson` and `TestAddress` to package level
* Now `JsonToObjectTransformer` supports not only `String` payload
* Remove json headers after transformation in the `JsonToObjectTransformer`
INT-2809 Minor Polishing
INT-2809 Docs and What's New
Wrap `FileInputStream` and `FileOutputStreams` with
`BufferedInputStream`, `BufferedOutputStream` respectively.
Also use `BufferedReader`, `BufferedWriter` to wrap `FileReader`, `FileWriter`.
JIRA: https://jira.springsource.org/browse/INT-2953
Previously, these tests examined specific stack frames to
verify the handler was directly invoked by the Dispatcher.
JIRA: https://jira.springsource.org/browse/INT-3188
Add a utility method to scan the stack trace instead to
verify that the Dispatcher frame exists before any
MessagingMethodInvokerHelper frame.
Previously, methods with primitive parameters were not matched
with primitive wrapper objects.
Check, if the argument is a primitive wrapper, that the paramer type
can be the equivalent primitive.
Add test case.
INT-3182 Polishing - PR Comments
Extract PrimitiveWrapper->Primitive map to ClassUtils.
JIRA: https://jira.springsource.org/browse/INT-3182
When a simple MessageHandler is referenced from a
`<service-activator/>`, it is wrapped in an anonymous
AbstractReplyProducingMessageHandler to facilitate proper handling of
the `<request-handler-advice-chain/>` and so that normal
'produced no reply` messages are logged.
Previously, this caused 2 MBeans to be exposed, one for tha actual
handler; one for the wrapper.
This added more stack frames between the dispatcher and the
handler (because 2 sets of statistics were captured).
The root cause was that the MBean exporter was not aware of the
relationship between these objects.
https://jira.springsource.org/browse/INT-3186
Add code to the IntegrationMBeanExporter to suppress adding
metrics to the wrapper.
This involves detecting that it is a wrapper and not exposing
it as an MBean and, when matching handlers with endpoints,
checking that the wrapped handler is the actual handler for the endpoint.
Update test to reflect the reduced number of stack frames between
the dispatcher and handler.
Also add tests to verify the detection of multiple references to
the same ARPMH.
INT-3186: `IntegrationMBeanExporter` optimization
- Close child context before parent to publish child close event to parent
- SpelRouter FB injection
- Don't pass null Method to TransactionAttributeSource.getTransactionAttribute()
When determining candidates for direct invocation, we need
to use the target of any existing proxy (such as JMX Metrics).
Copy ServiceActivatorDefaultFrameworkMethodTests to the JMX project
and adjust to work with JMX Proxies.
INT-3185: fix for AbstractSMHFBean#checkReuse
JIRA: https://jira.springsource.org/browse/INT-3185
https://jira.springsource.org/browse/INT-2961
onInit() relied on subclasses calling super.onInit() to apply the
advice chain and add the BeanFactory to the MessagingTemplate.
At least one extension (smpp) failed to do this, disabling these features.
Make onInit() final and call doInit() for subclass initialization (no-op in ARPMH).
INT-2961 Polishing
- do not increase visibility of doInit in subclasses.
- remove final modifiers from doInit so user subclasses can participate in initialization if needed.
https://jira.springsource.org/browse/INT-2629
Strategy for mapping gateway methods to map method arguments
to a message.
Default strategy works as today, but pluggable using attribute
'mapper' on <gateway/>.
A custom mapper should implement InboundMessageMapper<MethodArgsHolder>
where the MethodArgsHolder contains the Method object and the argument
values.
When using a custom mapper, the mapper is entirely responsible for
creating the Message - therefore 'payload-expression' attributes
and <header/> elements are not allowed.
INT-2629 Polishing - Javadocs
INT-2629 Polishing
- Introduce MethodArgsMessageMapper - higher level API for custom mapper.
INT-2629: Polishing docs
