Prior to this commit it was impossible to have an individual
constructor argument in a @Nested (i.e., inner) test class injected via
@Autowired, @Qualifier, or @Value.
This is due to a bug in javac on JDK versions prior to 9, whereby
annotation lookups performed directly via the
java.lang.reflect.Parameter API fail for inner class constructors.
Specifically, the parameter annotations array in the compiled byte code
for the user's test class excludes an entry for the implicit enclosing
instance parameter for an inner class constructor.
This commit introduces a workaround in ParameterAutowireUtils for this
off-by-one error by transparently looking up annotations on the
preceding Parameter object (i.e., index - 1). In addition, this commit
relies on the change recently introduced in MethodParameter in order to
compensate for the same JDK bug (see SPR-16652).
Issue: SPR-16653
This commit introduces @Nested tests to verify support for constructor
injection when using the SpringExtension.
One of the tests is disabled on Java 8 due to a bug in javac that was
first fixed in JDK 9.
See https://github.com/junit-team/junit5/issues/1345 for details.
This commit picks up where SPR-14498 left off by adding support for an
explicit target type when using JsonPath to perform an assertion against
the response content using a Hamcrest Matcher.
Specifically, there is a new overloaded value(Matcher<T>, Class<T>)
method in JsonPathResultMatchers for use with Hamcrest matchers where
the target type (i.e., Class<T>) can be specified.
Issue: SPR-16587
After this commit the use of interval in tests is combined with
take(n).onBackpressureBuffer(n) to ensure emissions don't fail if the
fixed rate is exceeded (e.g. on slow CI server).
Tests that verify N number of items followed by verifyOnComplete()
should set the number of emissions to N.
Tests that verify N number of items followed by thenCancel() should
set the number of buffered to an arbitrary number greater than N.
Previously MockHttpServletRequestBuilder calculated the pathInfo from the
provided URL without decoding the value. This meant that the pathInfo
incorrectly included URL encoded values.
Now MockHttpServletRequestBuilder properly decodes the pathInfo.
Fixes: SPR-16453
MockHttpServletResponse no longer calls flushBuffer on calls to
getContentAsByteArray and getContentAsString. The flushing doesn't
actually do anything useful but does commit the response leading to
unexpected side effects.
Issue: SPR-16430
This commit separates the creation of a response so that it is executed
after the synchronized block inside which requests need to be matched
and validated (for order and count).
This allows a ResponseCreator to be slow or block if it has to.
Issue: SPR-16319
This commit adds new header assertions for `WebTestClient`.
`doesNotExist` tests that a given header is not present:
.expectHeader().doesNotExist("Cache-Control");
`contentTypeCompatibleWith` tests for MediaType compatibility:
.expectHeader().contentTypeCompatibleWith("text/*");
Issue: SPR-16285
This avoids "IllegalStateException: Expectations already declared" when
a MockRestServiceServer is used after one request throws an exception.
Issues: SPR-16132
This commit decouples ExchangeResult from knowledge about
WiretapClientHttpRequest/Response both of which are now private to
WiretapConnector.
ExchangeResult takes ClientHttpRequest/Response instead along with
promises for the serialized request and response content in the form
of MonoProcessor<byte[]>.
This sets up the possibility for an ExchangeResult to be created
outside and independent of WebTestClient, should we decide to make its
constructor public.
Issue: SPR-16101
MockMvc now properly detects the presence of an AsyncContext and
re-uses the response instance used to start it.
This commit also includes a minor fix in
ResponseBodyEmitterReturnValueHandler to ensure it does not disable
ETag related content buffering for reactive return values that do not
result in streaming (e.g. single value or collections).
Issue: SPR-16067
The WebTestClient now takes advantage of the support for decoding
response to Void.class in WebClient so that applications can use
expectBody(Void.class) to the same effect as using
response.bodyToMono(Void.class) as documneted on WebClient#exchange.
The top-level, no-arg returnResult method (added very recently) has been
retracted, since the use of returnResult at that level, i.e. without
consuming the response content, should be used mainly for streaming.
It shouldn't be used for "no content" scenarios.
Documentation and Javadoc have been udpated accordingly.
Use the ClientResponse methods bodyToMono and bodyToFlux rather than
passing in a BodyExtractor so that WebTestclient now also benefits from
the recently improved handling of Void.class.
Concrete server and client, reactive request and response
implementations should not have to be accessed outside their package.
They could be seen as private to their HttpHandler adapters and
ClientHttpConnector's respectively.
The one exception, WebSocket upgrades in spring-webflux, is an internal
framework use case, accommodated via downcast + accessors in the
abstract base classes.
Prior to this commit, asking for a `Void` type using any of the
`ClientResponse#bodyTo*` methods would immediately return an empty
`Publisher` without consuming the response body.
Not doing so can lead to HTTP connection pool inconsistencies and/or
memory leaks, since:
* a connection that still has a response body being written to it cannot
be properly recycled in the connection pool
* incoming `DataBuffer` might not be released
This commit detects when `Void` types are asked as body types and in
those cases does the following:
1. Subscribe to the response body `Publisher` to allow the connection to
be returned to the connection pool
2. `cancel()` the body `Publisher` if the response body is not empty; in
that case, we choose to close the connection vs. consume the whole
response body
Those changes imply that `ClientHttpResponse` and other related
contracts don't need a `close()` method anymore.
Issue: SPR-16018
This commit moves WebFluxUriComponentsBuilder.fromServerRequest to the
ServerRequest interface itself.
Consequently, the WebFluxUriComponentsBuilder is removes itself, as it
contained no other methods.
Issue: SPR-15953
This commit introduces a methodName() method to the ServerRequest,
returning the String name of the method. This method is useful for
non-standard HTTP methods.
This commit introduces the following changes.
1) It adds a new Spring @NonNull annotation which allows to apply
@NonNullApi semantic on a specific element, like @Nullable does.
Combined with @Nullable, it allows partial null-safety support when
package granularity is too broad.
2) @Nullable and @NonNull can apply to ElementType.TYPE_USE in order
to be used on generic type arguments (SPR-15942).
3) Annotations does not apply to ElementType.TYPE_PARAMETER anymore
since it is not supported yet (applicability for such use case is
controversial and need to be discussed).
4) @NonNullApi does not apply to ElementType.FIELD anymore since in a
lot of use cases (private, protected) it is not part for the public API
+ its usage should remain opt-in. A dedicated @NonNullFields annotation
has been added in order to set fields default to non-nullable.
5) Updated Javadoc and reference documentation.
Issue: SPR-15756
Prior to this commit, various dependencies in the spring-test module
pulled in commons-logging as a transitive dependency. Consequently,
the presence of commons-logging in the generated Eclipse classpath
overrode the intended use of the spring-jcl module causing JUL to be
used instead of log4j for tests executed within Eclipse, thereby
ignoring the configuration in src/test/resources/log4j2-test.xml.
This commit addresses this issue by excluding commons-logging in
spring-test.gradle for all Selenium dependencies.
Issue: SPR-15930
Before this commit, there was no way to signal the HTTP client that we
were done consuming the response. Without that, the underlying client
library cannot know when it is safe to release the associated resources
(e.g. the HTTP connection).
This commit adds new `close()` methods on both `ClientHttpResponse`
and `ClientResponse`. This methods is non-blocking and its behavior
depends on the library, its configuration, HTTP version, etc.
At the `WebClient` level, `close()` is called automatically if we
consume the response body through the `ResponseSpec` or the
`ClientResponse` itself.
Note that it is *required* to call `close()` manually otherwise; not
doing so might create resource leaks or connection issues.
Issue: SPR-15920
This commit introduces a test that demonstrates that custom JUnit 4
rules can be @Autowired into a test instance but that they will not be
applied by JUnit since JUnit only ever sees such fields as null and
therefore ignores them.
Issue: SPR-15927
Juergen Hoeller
igor-suhorukov
Sam Brannen
Rossen Stoyanchev
Arjen Poutsma
Bronwyn Perry-Huston
Yannik Hampe
Stephane Nicoll
Rob Winch
Ketan Padegaonkar
Neonailol
Brian Clozel
sdeleuze
anton_bondarenko
Eric Pabst