This commit fixes the `configureMessageConverters` and
`configureMessageConvertersList` behavior.
`configureMessageConverters` was not executing consumers in their order
of registration (but in the reverse order).
`configureMessageConvertersList` was not executing multiple consumers
and was instead executing the first consumer multiple times.
This commit fixes both issues.
Fixes gh-36332
get method skips containsKey and instead checks if the enumeration
has elements, which should give the same behavior other than for
headers without values.
See gh-36334
- Optimize get method for request headers
- Update keySet methods to use custom extension of AbstractSet
- Drop use of native Tomcat headers, which could be an issue for
request and response wrappers that override header methods.
The performance of Servlet adapters should be similar for the
commonly used methods, and it should be possible to optimize
further for the future in HttpHeaders (e.g. by adding a
set alternative to put), and requesting Servlet API refinements.
Closes gh-36334
Prior to this commit, the reactive `MultipartParser` and `PartGenerator`
types were leaking memory at runtime in specific cases:
* many HTTP clients must send multipart requests to be parsed and close
the connection while uploading
* the `PartGenerator` must be configured to write file parts to
temporary files on disk
* concurrency, upload speed must be important to trigger cases where the
file system is not fast enough to consume incoming buffers
The `MultipartParser` parses and emits `BodyToken` to its sink
(here, the `PartGenerator`). By definition, Reactor's `FluxSink` when
created with `Flux.create(FluxSink)` will use a "buffer" strategy and
will queue emitted elements if they cannot be consumed.
Here, the cancellation signal does dispose internal states in the
`MultiPartParser` and `PartGenerator` but does not clear the internal
queue in `FluxSink`.
This commit ensures that an operation is registered to release buffers
on the discard event.
Fixes gh-36262
Prior to this commit, the `MediaType` and `MimeType` "copy" constructors
would not leverage the fact that the existing instance has been
validated already (types, subtype and parameters have been checked
already for errors) and the entire validation would be performed again.
This would also allocate map instances in the process.
This commit ensures that the already validated information is reused
directly and that we avoid unnessecary operations and allocations for
such constructors.
Closes gh-36318
Prior to this commit, the findAutowiredAnnotation() method in
AutowiredAnnotationBeanPostProcessor fully supported finding
@Autowired as a meta-annotation; however, the isRequired() method in
QualifierAnnotationAutowireCandidateResolver only found @Autowired as
a "directly present" annotation without any support for
meta-annotations.
For consistency and to avoid bugs, this commit revises
QualifierAnnotationAutowireCandidateResolver so that we always support
@Autowired as a meta-annotation.
Closes gh-36315
This commit revises AutowiredAnnotationBeanPostProcessor so that
determineRequiredStatus(MergedAnnotation<?>) only looks up the required
attribute once.
Closes gh-36314
The builder for `HttpMessageConverters` allows for auto-detection of
message converters on the classpath and their default registration when
`registerDefaults()` is called. Once called, there is no way to undo
this.
This commit adds a new `disableDefaults()` method to disable the default
registration and take full control over the list of message converters.
Closes gh-36303
Prior to this commit, AnnotationConfigUtils looked up @Lazy as a
meta-annotation at arbitrary depths (e.g., when used as a
meta-meta-annotation); however,
ContextAnnotationAutowireCandidateResolver only found @Lazy as a
"directly present" meta-annotation.
For consistency, this commit revises
ContextAnnotationAutowireCandidateResolver so that it also finds @Lazy
as a meta-annotation at arbitrary depths.
Closes gh-36306
Prior to this commit, ValidationAnnotationUtils looked up @Validated
as a meta-annotation at arbitrary depths (e.g., when used as a
meta-meta-annotation) in determineValidationGroups() but only as a
"directly present" meta-annotation in determineValidationHints().
For consistency, this commit revises determineValidationHints() so that
it finds @Validated as a meta-annotation at arbitrary depths as well.
Closes gh-36274
Prior to this commit, the `HttpEntityMethodProcessor` would create a new
`ServletServerHttpRequest` input message to parse the native Servlet
request, but would not reuse it for reading the request body using the
message converters.
In gh-32471, we applied a change that updates HTTP headers accordingly
when request parameters are read. But not reusing the input message
means that we are losing this update when instantiating the resulting
`HttpEntity`.
This commit ensures that `HttpEntityMethodProcessor` uses the input
message it just created when decoding the request body.
Fixes gh-36298
Prior to this commit, an attempt to restart the
StompBrokerRelayMessageHandler (SBRMH) resulted in an
IllegalStateException stating that the ReactorNettyTcpClient was still
in the process of "Shutting down." The reason is that a
ReactorNettyTcpClient cannot be restarted after it has been closed, and
that is by design.
To address that issue, this commit introduces an
`internallyManagedTcpClient` flag in SBRMH that is used to track
whether SBRMH is in charge of managing the TcpClient internally or if
the TcpClient was supplied by the user and is therefore managed
externally.
If SBRMH manages the TcpClient internally, isPauseable() returns
`true`, and the `tcpClient` field is set to null (in stopInternal())
when the handler is stopped. Consequently, a new internally managed
TcpClient will be created when the handler is restarted.
If the TcpClient is managed externally, the handler is not considered
to be "pauseable", and a reference to the externally managed TcpClient
is retained after the handler has been stopped. If the
ApplicationContext is subsequently restarted, the externally managed
TcpClient will be reused -- which may or may not work, depending on the
implementation of the TcpClient. Note, however, that this has always
been the behavior of SBRMH with regard to stop/start scenarios for
externally managed TcpClients.
Closes gh-36266
Stéphane Nicoll
rstoyanchev
Juergen Hoeller
海子 Yang
Niravil
Brian Clozel
Sam Brannen
Christian Schuster