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
Prior to this commit, only the java.lang.reflect.Method was used to
identify an annotated method. As a result, if different annotations
were placed on different methods (method overriding, interface
implementation) only the first one (cached) was used.
LazyParamAwareEvaluationContext was affected by the exact
same problem and has been also fixed.
Issue: SPR-11692
Prior to this commit, the CacheResolver was not used by Spring's
caching abstraction. This commit provides the necessary configuration
options to tune how a cache is resolved for a given operation.
CacheResolver can be customized globally, at the operation level or at
the class level. This breaks the CachingConfigurer class and a support
implementation is provided that implements all methods so that the
default is taken if it's not overridden. The JSR-107 support has been
updated as well, with a similar support class.
In particular, the static and runtime information of a cache
operation were mixed which prevents any forms of caching. As the
CacheResolver and the KeyGenerator can be customized, every operation
call lead to a lookup in the context for the bean.
This commit adds CacheOperationMetadata, a static holder of all
the non-runtime metadata about a cache operation. This is used
as an input source for the existing CacheOperationContext.
Caching the operation metadata in an AspectJ aspect can have side
effects as the aspect is static instance for the current ClassLoader.
The metadata cache needs to be cleared when the context shutdowns.
This is essentially a test issue only as in practice each application
runs in its class loader. Tests are now closing the context properly
to honor the DisposableBean callback.
Issue: SPR-11490
This commit adds a putIfAbsent method to the Cache interface. This
method offers an atomic put if the key is not already associated in
the cache.
Issue: SPR-11400
Prior to this commit, an exception thrown by an @Async void method
was not further processed as there is no way to transmit that
exception to the caller.
The AsyncUncaughtExceptionHandler is a new strategy interface that
can be implemented to handle unexpected exception thrown during the
invocation of such asynchronous method.
The handler can be specified using either the XML namespace or by
implementing the AsyncConfigurer interface with the EnableAsync
annotation.
Issue: SPR-8995
Prior to this commit, common cache operation settings had to be
repeated for every operation: cache name(s), custom cache manager
and custom key manager.
This commit introduces the @CacheConfig annotation to bet set at
class-level (either directly or as a meta-annotation). As the cache
name(s) can be rationalized there, the "value" of the various
annotations are no longer mandatory.
CacheAnnotationParser has an API breakage to be able to retrieve
information at class-level.
Issue: SPR-11316
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
Prior to this commit, Spring supported meta-annotation attribute
overrides in custom composed annotations with reflection-based
annotation processing but not with ASM-based annotation processing.
This commit ensures that meta-annotation attribute overrides are
supported in AnnotationMetadataReadingVisitor.getAnnotationAttributes().
Issue: SPR-11574
Prior to this commit, AnnotationAttributesReadingVisitor treated Class
annotation attributes as Strings instead of Classes. As a result,
look-ups for Class attributes in meta-annotations failed with an
IllegalArgumentException.
This commit fixes this issue by consistently treating Class attributes
as Classes in AnnotationAttributesReadingVisitor.
Issue: SPR-11557
This commit introduces unit and integration tests that attempt to
reproduce the issue claimed by the reporter in SPR-11557. However, the
tests pass without any problems.
Issue: SPR-11557
Restored original singleton-only adaptInnerBeanName behavior, avoiding endless unique name calculation for every single prototype instance. Non-named inner BeanDefinition objects now suffixed with an identity hex code analogous to non-named XML bean definitions, largely avoiding naming collisions to begin with. After SPR-11246, post-processors can deal with unstable classes per bean name, so occasional collisions aren't a hard problem anymore.
Issue: SPR-11545
Prior to this commit, an array argument was not handled properly in
SimpleKey#equals and SimpleKey#hashCode. As a result, two method
invocations with the same array argument lead to two different keys
and therefore two different entries in the cache.
This commit uses deepEquals and deepHashCode to properly handle
methods that have arguments that are array types.
Issue: SPR-11505
