Previously, if a `@Cacheable` method was accessed with the same key by
multiple threads, the underlying method was invoked several times instead
of blocking the threads while the value is computed. This scenario
typically affects users that enable caching to avoid calling a costly
method too often. When said method can be invoked by an arbitrary number
of clients on startup, caching has close to no effect.
This commit adds a new method on `Cache` that implements the read-through
pattern:
```
<T> T get(Object key, Callable<T> valueLoader);
```
If an entry for a given key is not found, the specified `Callable` is
invoked to "load" the value and cache it before returning it to the
caller. Because the entire operation is managed by the underlying cache
provider, it is much more easier to guarantee that the loader (e.g. the
annotated method) will be called only once in case of concurrent access.
A new `sync` attribute to the `@Cacheable` annotation has been addded.
When this flag is enabled, the caching abstraction invokes the new
`Cache` method define above. This new mode bring a set of limitations:
* It can't be combined with other cache operations
* Only one `@Cacheable` operation can be specified
* Only one cache is allowed
* `condition` and `unless` attribute are not supported
The rationale behind those limitations is that the underlying Cache is
taking care of the actual caching operation so we can't really apply
any SpEL or multiple caches handling there.
Issue: SPR-9254
Even though the JSR-107 spec forbids to store null values, our cache
abstraction allows that behaviour with a special handled (and this is
the default behaviour).
While this was working fine with our own set of annotations, the
JSR-107 interceptor counterpart was interpreting the spec sensu strictu.
We now allow for that special case as well.
Issue: SPR-13641
While working on SPR-12532, an extra IdentityWrapper was added to work
around a backward compatible issue between commons pool 1.x and 2.x. This
issue (POOL-283) has actually been fixed in 2.4 and their IdentityWrapper
is using object equality so our wrapper is in the way.
Looking retrospectively, the code looks all fine without the workaround
and commons pool 2.4 or later so it has been removed.
Previously, a Bean implementing `AutoCloseable` (or `Closeable`) was
always destroyed regardless of its bean definition. In particular, the
documented way of disabling the destruction callback via an empty String
did not work.
AutoCloseable beans are now treated pretty much as any other bean: we
still use the presence of the interface to optimize the check of a
destroy method and we only auto-discover the method name to invoke if
the inferred mode is enabled.
Issue: SPR-13022
Prior to this commit, the GroovyBeanDefinitionReader claimed (via its
Javadoc) that it fully supported XML configuration files in addition to
its Groovy DSL; however, this was unfortunately inaccurate since XML
validation was disabled by default which led to certain features of XML
configuration not working. For example, it was impossible to define a
<qualifier> in an XML config file without specifying the 'type'
attribute (which has a default value defined in the spring-beans XSD).
This commit fixes this issue by ensuring that bean definitions in XML
resources are loaded with a "standard" XmlBeanDefinitionReader that is
created with default settings (i.e., with XML validation enabled). With
regard to backwards compatibility, bean definitions defined using the
Groovy DSL are still loaded with an XmlBeanDefinitionReader that has
XML validation disabled by default which is necessary for proper
parsing of the Groovy DSL.
Issue: SPR-12769
Previously, the exception-handler attribute was not taken care of when
task:annotation-driven is used in AspectJ mode. This commit provides the
expected behavior.
Issue: SPR-12619
Add support for annotation-based event listeners. Enabled automatically
when using Java configuration or can be enabled explicitly via the
regular <context:annotation-driven/> XML element. Detect methods of
managed beans annotated with @EventListener, either directly or through
a meta-annotation.
Annotated methods must define the event type they listen to as a single
parameter argument. Events are automatically filtered out according to
the method signature. When additional runtime filtering is required, one
can specify the `condition` attribute of the annotation that defines a
SpEL expression that should match to actually invoke the method for a
particular event. The root context exposes the actual `event`
(`#root.event`) and method arguments (`#root.args`). Individual method
arguments are also exposed via either the `a` or `p` alias (`#a0` refers
to the first method argument). Finally, methods arguments are exposed via
their names if that information can be discovered.
Events can be either an ApplicationEvent or any arbitrary payload. Such
payload is wrapped automatically in a PayloadApplicationEvent and managed
explicitly internally. As a result, users can now publish and listen
for arbitrary objects.
If an annotated method has a return value, an non null result is actually
published as a new event, something like:
@EventListener
public FooEvent handle(BarEvent event) { ... }
Events can be handled in an aynchronous manner by adding `@Async` to the
event method declaration and enabling such infrastructure. Events can
also be ordered by adding an `@Order` annotation to the event method.
Issue: SPR-11622
This commit removes the immediate package dependency cycle between the context and jmx packages. A specific callback arrangement will follow in time for 4.1 RC1; at this point, it's temporarily back to registration kicked off by afterPropertiesSet again.
Issue: SPR-8045
Reorganized class structure to match our code style (setter for
properties at the top of the class, public method before private
implementation).
Removed DisposableBean as it the lifecycle is already taking care
of removing MBeans on stop.
Cleaned test suite
Issue: SPR-8045
Prior to this commit, CacheResolver could not be configured through
the XML namespace (i.e. cache:annotation-driven). This is now the
case.
Issue: SPR-11490
This commit adds the necessary infrastructure to handle exceptions
thrown by a cache provider in both Spring's and JCache's caching
abstractions.
Both interceptors can be configured with a CacheErrorHandler that
defines several callbacks on typical cache operations. In particular,
handleCacheGetError can be implemented in such a way that an
exception thrown by the provider is handled as a cache miss by the
caching abstraction.
The handler can be configured with both CachingConfigurer and the
XML namespace (error-handler property)
Issue: SPR-9275
It is now possible to specify the CacheManager to use per operation.
The related cache annotation now has an extra attribute that defines
the name of the CacheManager bean to use. The cache manager that
was previously used is therefore a 'default' cache manager (i.e. the
one to use if no custom cache manager has been set on the operation).
Issue: SPR-8696
This commit adds an extra parameter to the base @Cache method
annotations: keyGenerator. This parameter holds the name of the
KeyGenerator bean to use to compute the key for that specific
caching endpoint.
This gives therefore a third way to customize the key. These are:
1. Default KeyGenerator (global for all endpoints)
2. The 'key' attribute of the annotation, giving the SpEL expression to use
3. The 'keyGenerator' attribute of the annotation
The annotation attributes are therefore exclusive. Trying to specify
them both will result in an IllegalStateException.
The KeyGenerator to use for a given operation is cached on startup
so that multiple calls to it does not resolve the instance to use over and
over again.
Issue: SPR-10629
Update `CacheAspectSupport` to expand any var-arg parameters before
calling `KeyGenerator` implementations. Prior to this commit var-args
would be passed to `KeyGenerator` implementations as a nested array,
often causing the same key to be generated regardless of the arguments.
Issue: SPR-10870
Allow the @CachePut 'key' SpEL to refer to the result of the method call
via a '#result' variable.
This change is particularly useful when working with JPA entities that
have generated @Id values since the ID will often not be available
until the entity has been saved.
Issue: SPR-10664
Allow @Cachable, @CachePut and equivalent XML configuration to provide
a SpEL expression that can be used to veto putting an item into the
cache. Unlike 'condition' the 'unless' parameter is evaluated after
the method has been called and can therefore reference the #result.
For example:
@Cacheable(value="book",
condition="#name.length < 32",
unless="#result.hardback")
This commit also allows #result to be referenced from @CacheEvict
expressions as long as 'beforeInvocation' is false.
Issue: SPR-8871
Prior to this commit many test utility classes and sample beans were
duplicated across projects. This was previously necessary due to the
fact that dependent test sources were not shared during a gradle
build. Since the introduction of the 'test-source-set-dependencies'
gradle plugin this is no longer the case.
This commit attempts to remove as much duplicate code as possible,
co-locating test utilities and beans in the most suitable project.
For example, test beans are now located in the 'spring-beans'
project.
Some of the duplicated code had started to drift apart when
modifications made in one project where not ported to others. All
changes have now been consolidated and when necessary existing tests
have been refactored to account for the differences.
Conflicts:
spring-beans/src/test/java/org/springframework/beans/factory/ConcurrentBeanFactoryTests.java
spring-beans/src/test/java/org/springframework/beans/factory/support/BeanFactoryGenericsTests.java
spring-beans/src/test/java/org/springframework/beans/support/PagedListHolderTests.java
In the course of this enhancement, the "cache.ehcache" and "cache.jcache" packages moved from spring-context to the spring-context-support module, expecting further transaction-related functionality. Also aligns with the presence of Spring's Quartz support in the spring-context-support module, since Quartz and EHCache are sort of sister projects at Terracotta now.
Issue: SPR-9966
java.util.concurrent's ScheduledExecutorService and its #schedule*
methods allow for an 'initialDelay' parameter in milliseconds.
Similarly, Spring's TaskExecutor abstraction allows for a concrete
'startTime' expressed as a Date. However, Spring's <task:scheduled> XML
element and @Scheduled annotation have, to date, not allowed for an
initial delay parameter that can be propagated down to the underlying
TaskScheduler/ScheduledExecutorService.
This commit introduces initial-delay and #initialDelay attributes to
task:scheduled and @Scheduled respectively, both indicating the number
of milliseconds to wait before the first invocation of the method in
question. Specifying a delay in this fashion is only valid in
conjunction with fixed-rate and fixed-delay tasks (i.e. not with cron
or trigger tasks).
The principal changes required to support these new attributes lie in
ScheduledTaskRegistrar, which previously supported registration of
tasks in the form of a Runnable and a Long parameter indicating (in the
case of fixed-rate and fixed-delay tasks), the interval with which the
task should be executed. In order to accommodate a third (and optional)
'initialDelay' parameter, the IntervalTask class has been added as a
holder for the Runnable to be executed, the interval in which to run
it, and the optional initial delay. For symmetry, a TriggerTask and
CronTask have also been added, the latter subclassing the former. And a
'Task' class has been added as a common ancestor for all the above.
One oddity of the implementation is in the naming of the new
setters in ScheduledTaskRegistrar. Prior to this commit, the setters
were named #setFixedDelayTasks, #setFixedRateTasks, etc, each accepting
a Map<Runnable, long>. In adding new setters for each task type, each
accepting a List<IntervalTask>, List<CronTask> etc, naturally the
approach would be to use method overloading and to introduce methods
of the same name but with differing parameter types. Unfortunately
however, Spring does not support injection against overloaded methods
(due to fundamental limitations of the underlying JDK Introspector).
This is not a problem when working with the ScheduledTaskRegistrar
directly, e.g. from within a @Configuration class that implements
SchedulingConfigurer, but is a problem from the point of view of the
ScheduledTasksBeanDefinitionParser which parses the <task:scheduled>
element - here the ScheduledTaskRegistrar is treated as a Spring bean
and is thus subject to these limitations. The solution to this problem
was simply to avoid overloading altogether, thus the naming of the new
methods ending in "List", e.g. #setFixedDelayTasksList, etc. These
methods exist primarily for use by the BeanDefinitionParser and are
not really intended for use by application developers. The Javadoc for
each of the new methods makes note of this.
Issue: SPR-7022
This renaming more intuitively expresses the relationship between
subprojects and the JAR artifacts they produce.
Tracking history across these renames is possible, but it requires
use of the --follow flag to `git log`, for example
$ git log spring-aop/src/main/java/org/springframework/aop/Advisor.java
will show history up until the renaming event, where
$ git log --follow spring-aop/src/main/java/org/springframework/aop/Advisor.java
will show history for all changes to the file, before and after the
renaming.
See http://chrisbeams.com/git-diff-across-renamed-directories
