* getBeanDefinitionNames defensively returns a copy of the bean definition names array.
* copyConfigurationFrom provides an independent AutowireCandidateResolver instance and copies a ConversionService and dependency comparator configuration as well.
* findAutowireCandidates only considers a self reference fallback for direct dependency declarations, not as a collection element.
Issue: SPR-14897
Issue: SPR-14921
Issue: SPR-14965
In general, web.util is supposed to remain lowest-level, not depending on any other web.* package. Since web.client also has a support package, a corresponding web.server.support package seemed appropriate for a helper class depending on ServerWebExchange in web.server itself.
The recent refactoring lead to
java.lang.NoClassDefFoundError: io/reactivex/Completable
where only RxJava 1 is in the classpath. Most likely due to the
lack of prefix in Completable::complete with rx package to avoid
the RxJava 2 reference.
Main differences with the Spring MVC original implementation:
- Default redirect HTTP code is 303 See Other since we can assume
all HTTP clients support HTTP 1.1 in 2016
- No more http10Compatible property, use statusCode instead
- By default the redirect is relative to the context path
- A builder allow to set various properties if needed
- In UrlBasedViewResolver, a Function<String, RedirectView>
redirectViewProvider property allows to customize RedirectView
instances in a flexible way
Issue: SPR-14534
Simplify getAdapterFrom/To into a single getAdapter method that looks
for an exact match by type first and then isAssignableFrom.
Also expose shortcut methods in ReactiveAdapter to minimize the need
for access to the ReactiveTypeDescriptor.
Issue: SPR-14902
The ReactiveAdapterRegistry now detects the presence of Reactor. In
practice Reactor is required for the Spring Framework reactive support
and it is expected to be present.
The registry however is now capable of being neutral if Reactor is not
present on the classpath for example where other Spring projects may
not have the same assumptions about Reactor's presence.
Issue: SPR-14902
Collapse ReactiveAdapter hierarchy into a single class that simply
delegates to functions for converting to/from a Publisher.
A private ReactorAdapter extension automaticlaly wraps adapted, "raw"
Publisher's as Flux or Mono depending on the semantics of the target
reactive type.
Issue: SPR-14902
This commit polishes previous one by also accepting
generic types explicitly declared with a class that
extends DataBuffer allowing to write Flux<DefaultDataBuffer>
for example.
Issue: SPR-14952
This modifies the signature of
ReactiveHttpOutputMessage#writeAndFlush(...) in order to
be able to use Flux<Flux<DataBuffer>> objects as arguments of
this method.
Issue: SPR-14952
This commit polishes Kotlin nullable support by reusing
MethodParameter#isOptional() instead of adding a new
MethodParameter#isNullable() method, adds
Kotlin tests and introduces Spring Web Reactive
support.
Issue: SPR-14165
Where `isOptional` is used, also check for `isNullable` i.e.
values are not considered required if they are Kotlin nullables:
- spring-messaging: named value method arguments
- spring-web: named value method arguments
- spring-webmvc: request parts
This means that Kotlin client code no longer has to explicity specify
"required=false" for Kotlin nullables -- this information is inferred
automatically by the framework.
Issue: SPR-14165
In case the filter is also registered to the ERROR dispatcher, the
following happens:
* the filter is executed once for the regular execution
* the filter should be executed a second time when dispatched to error
Since the filter is a `OncePerRequestFilter`, the filter is only
executed once and won't be executed when handling the error.
This can lead to situations like spring-projects/spring-boot#7348
This commit makes this filter a simple `GenericFilterBean`.
Issue: SPR-14891
When processing a "close" notification from the server make an effort
to cancel any outstanding heartbeat but avoid going as far as acquiring
the responseLock since the server itself may already hold a lock of its
own leading to a potential deadlock.
The heartbeat task is now also further protected with an isClosed()
check in case the heartbeat does not get cancelled in a concurrent
scenario.
Issue: SPR-14917
Juergen Hoeller
Rossen Stoyanchev
Brian Clozel
Sebastien Deleuze
Daniel Fernández
Raman Gupta
Arjen Poutsma
Violeta Georgieva
Stephane Maldini
Stephane Nicoll