Refactor AbstractMessageConverterMethodArgumentResolver and
BridgeMethodResolver to use ResolvableType in preference to deprecated
GenericTypeResolver calls.
Issue: SPR-10980
Add a new ResolvableType Class which encapsulates java.lang.reflect.Type,
providing access to supertypes, interfaces and generic parameters along
with the ability to ultimately resolve to a java.lang.Class.
ResolvableTypes may be obtained from fields, method parameters, method
returns, classes or directly from a java.lang.reflect.Type. Most methods
will themselves return ResolvableTypes, allowing easy navigation.
For example:
private HashMap<Integer, List<String>> myMap;
public void example() {
ResolvableType t = ResolvableType.forField(
getClass().getDeclaredField("myMap"));
t.getSuperType(); // AbstractMap<Integer, List<String>>;
t.asMap(); // Map<Integer, List<String>>
t.getGeneric(0).resolve(); // Integer
t.getGeneric(1).resolve(); // List
t.getGeneric(1); // List<String>
t.resolveGeneric(1, 0); // String
}
Issue: SPR-10973
A few noteworthy minor changes: LocaleContext.getLocale() may return null in special cases (not by default), which our own accessing classes are able to handle now. If there is a non-null TimeZone user setting, we're exposing it to all collaborating libraries, in particular to JSTL, Velocity and JasperReports. Our JSR-310 and Joda-Time support falls back to checking the general LocaleContext TimeZone now, adapting it to their time zone types, if no more specific setting has been provided. Our DefaultConversionService has TimeZone<->ZoneId converters registered. And finally, we're using a custom parseTimeZoneString method now that doesn't accept the TimeZone.getTimeZone(String) GMT fallback for an invalid time zone id anymore.
Issue: SPR-1528
Added extension to Future with capabilities for registering callbacks
when the future is complete.
- Added ListenableFuture, ListenableFutureCallback,
ListenableFutureCallbackRegistry, and ListenableFutureTask.
- Using ListenableFuture in AsyncRestOperations/AsyncRestTemplate.
- Added AsyncListenableTaskExecutor, implemented in
SimpleAsyncTaskExecutor.
- Added FutureAdapter and ListenableFutureAdapter.
Update AntPathMatcher Comparator to treat `/**` in the same way as
`null` paths.
Prior to this commit the pattern `/**` would be picked in preference
to patterns with 3 or more PathVariable (e.g. `/matches/{matchId}/
periods/{periodId}/teams/{teamId}/results`).
Issue: SPR-10550
Refine AntPathMatcher.combine rules to allow direct concatenation of
patterns when pattern1 does not contain '*.'. Prior to this commit
direct concatenation was allowed when pattern1 did not contain '.',
this prevented calls of the form:
pathMatcher.combine("/1.0", "/foo/test")
from working as expected.
This commit also applies some general cleanup to the `combine` method.
Issue: SPR-10554
This turned into a rather huge affair since it led to the introduction of a new AutowireCandidateResolver implementation in the spring-context module. That ACR impl is now being set through AnnotationConfigUtils; GenericApplicationContext and co do not set a default QualifierAnnotationAutowireCandidateResolver anymore (which has always been a smell anyway). At the same time, dependency ordering has moved from AutowiredAnnotationBeanPostProcessor to DefaultListableBeanFactory itself through a "dependencyComparator" strategy, applying to constructor dependencies and lazy resolution proxies as well.
Issue: SPR-10353
Also, StandardReflectionParameterNameDiscoverer calls "Parameter.isNamePresent()" now to return null (and pass on to the next discoverer) if no JDK 8 parameters are available. Note that this requires OpenJDK 8 b100 or higher to compile now.
Issue: SPR-10532
MimeType is available in core-spring and does not include support
for quality parameters and media used in HTTP content negotiation.
The MediaType sub-class in org.springframework.http adds q-parameters.
The method returning query parameters now returns only query string
parameters as opposed to any Servlet request parameter.
This commit also adds a ReadOnlyMultiValueMap.
Previously building with JDK > 1.8 b88 caused test failures due to errors
with custom compilers like Jibx and Jasper reports.
This commit adds a new TestGroup named CUSTOM_COMPILATION that allows the
CI server to continue to run these tests but allow committers to ignore
these tests.
Add ByteBufferConverter that is registered by default with the
DefaultConversionService. Allows conversion from/to a ByteBuffer and
byte[] or to any type that can be converted via a byte[].
Issue: SPR-10712
Prior to this commit the Spring Framework did not provide a public means
for scanning for available server ports. However, the Spring Framework
internally used a FreePortScanner in integration tests within its own
test suite. Furthermore, Spring Integration 2.2 provides similar support
in a SocketUtils class in the spring-integration-test module.
This commit introduces SocketUtils in spring-core to replace the
FreePortScanner which was previously only used internally within
Spring's test suite. This new implementation is inspired by both Spring
Framework's FreePortScanner and Spring Integration's SocketUtils and
consequently attempts to merge the best of both previous
implementations.
Issue: SPR-8032
Develop new org.springframework.core.io.Resource implementation
backed by java.nio.file.Path. Primarily developed to allow custom
file system implementations to be used with Spring.
Since the minimum requirement for Spring is still Java 6 the
existing FileSystemResource can't be retrofitted (and no #getPath
method can be added to the Resource interface).
Unlike FileSystemResource, PathResource delegates to the underlying
file system instead of StringUtils. It has therefore slightly
different semantics. First, when building relative resources via
createRelative the relative path will apply to this path (like URL or
Unix). Second, equality is delegated to the underlying file system
provider so it's case-insensitive on Windows.
Issue: SPR-10608
JOpt 4.4 has enumerable options, so this change can be made
if we upgrade. The only awkward thing is that JOpt allows
aliases for options, so we have to pick one to avoid double
counting. This implementation picks the last one in the list
which is the alphebtically last of the long options, if there
are any (e.g. "o1", "option1" returns "option1"). Most of the
time there will only be one or two aliases for each option so
it won't matter.
Issue: SPR-10579
Update AnnotationMetadata and MethodMetadata to extend from a new
AnnotatedTypeMetadata base interface containing the methods that are
common to both. Also introduce new getAllAnnotationAttributes methods
providing MultiValueMap access to both annotation and meta-annotation
attributes.
Existing classreading and standard implementations have been
refactored to support the new interface.