Prior to this commit, a DataSize input string could not be parsed if it
contained any whitespace.
With this commit, a DataSize input string can contain leading, trailing,
or 'in between' whitespace. For example, the following will be parsed
to the same DataSize value.
- "1024B"
- "1024 B"
- " 1024B "
- " 1024 B "
Closes gh-28643
Commit d6768ccc18 introduced a regression in the support for merging
annotation attributes in a multi-level annotation hierarchy.
This commit addresses that issue by ensuring that a merged annotation
is once again synthesized if a meta-annotation in the annotation
hierarchy declares attributes that override attributes in the target
annotation.
Closes gh-28716
When the ProtocolResolver SPI was introduced in Spring Framework 4.3,
support for protocol resolvers was added in DefaultResourceLoader's
getResource() implementation; however, GenericApplicationContext's
overridden getResource() implementation was not updated accordingly.
Prior to this commit, if a GenericApplicationContext was configured
with a custom ResourceLoader, registered protocol resolvers were
ignored.
This commit ensures that protocol resolvers are honored in
GenericApplicationContext even if a custom ResourceLoader is used.
Closes gh-28703
Prior to this commit, meta-annotations were unnecessarily synthesized
when attempting to synthesize a MergedAnnotation retrieved via the
MergedAnnotations.from(AnnotatedElement, ...).get(<annotationType>) API.
This is a regression in our merged annotation support that was
introduced when the MergedAnnotations API replaced our previous support.
This commit fixes this by revising the logic in TypeMappedAnnotation's
createSynthesizedAnnotation() method so that a meta-annotation is
returned unmodified if it is not synthesizable.
This commit also updates BootstrapUtilsTests, since @BootstrapWith
should never have been synthesized, and Class#getCanonicalName() is
only used in the toString() implementation of an annotation synthesized
by Spring or normal annotations on Java 19+ (see
https://bugs.openjdk.org/browse/JDK-8281462).
Closes gh-28704
Prior to this commit, ObjectToObjectConverter considered the return
type of non-static `to[targetType.simpleName]()` methods but did not
consider the return type of static `valueOf(sourceType)`,
`of(sourceType)`, and `from(sourceType)` methods.
This led to scenarios in which `canConvert()` returned `true`, but a
subsequent `convert()` invocation resulted in a
ConverterNotFoundException, which violates the contract of the
converter.
This commit addresses this issue by taking into account the return type
of a static valueOf/of/from factory method when determining if the
ObjectToObjectConverter supports a particular conversion. Whereas the
existing check in `determineToMethod()` ensures that the method return
type is assignable to the `targetType`, the new check in
`determineFactoryMethod()` leniently ensures that the method return
type and `targetType` are "related" (i.e., reside in the same type
hierarchy).
Closes gh-28609
To avoid unnecessary eager initialization of DataSize.PATTERN, this
commit initializes it lazily in the first invocation of DataSize.parse
by moving PATTERN to a private static nested class.
Closes gh-28560
With a Java 8 baseline in place for quite some time now, it no longer
makes sense to refer to features such as annotations as "Java 5
annotations".
This commit also removes old `Tiger*Tests` classes, thereby avoiding
duplicate execution of various tests.
This commit adds a warning to the Javadoc for the
TYPE_HIERARCHY_AND_ENCLOSING_CLASSES search strategy in
MergedAnnotations with regard to the scope of the search
algorithm.
See gh-28079
