This commit is a backport of SPR-11836.
Prior to this commit, the ordering of classpath entries in the
generated Eclipse .classpath files did not align with the intended
dependency management configured in the Gradle build. Specifically,
project dependencies often came before dependencies on third-party
libraries required for the given project, causing the project not to
compile in Eclipse.
This commit fixes this issue by introducing new functionality in the
ide.gradle script that moves all project dependencies to the end of the
generated classpath.
Issue: SPR-12947, SPR-11836
Using the "quotes" substitution group by default leads to side effect
when the "*" character is used. This is especially true for AOP pointcut
or for MVC mappings.
Plain verbatim might work most of the time unless you intend to highlight
a piece of code or a comment.
Issue: SPR-12456
(cherry picked from commit 0d00b674f4)
SPR-12354 applied new checks to make sure that served static resources
are under authorized locations.
Prior to this change, serving static resources from Servlet 3 locations
such as "/webjars/" would not work since those locations can be within
one of the JARs on path. In that case, the checkLocation method would
return false and disallow serving that static resource.
This change fixes this issue by making sure to call the
`ServletContextResource.getPath()` method for servlet context resources.
Note that there's a known workaround for this issue, which is using a
classpath scheme as location, such as:
"classpath:/META-INF/resources/webjars/" instead of "/webjars".
Issue: SPR-12432
(cherry picked from commit 161d3e3)
- remove leading '/' and control chars
- improve url and relative path checks
- account for URL encoding
- add isResourceUnderLocation final verification
Issue: SPR-12354
Prior to this commit, finding out how many application contexts had
been loaded within a test suite required the use of reflection and a
bit of hacking.
This commit addresses this issue by logging ContextCache statistics
whenever an application context is loaded by the Spring TestContext
Framework (TCF).
The log output can be enabled by setting the
"org.springframework.test.context.cache" logging category to DEBUG.
Issue: SPR-12409
(cherry picked from commit da04362)
This commit updates the documentation of the task:executor element to
reference the keep-alive flag.
Issue: SPR-12407
(cherry picked from commit 6534d00)
Also includes order preservation for resource results with Tiles 2 as well as retrieval failure logging with Tiles 3.
Issue: SPR-12362
(cherry picked from commit cb86036)
Prior to this commit, the ResolvableType static cache was holding a lot
of duplicates for simple types. We are using too much metadata to compute
the key when the class has no generic information. so setFoo(String foo)
and setBar(String bar) would result in two entries in the cache because
the TypeProvider is different. On a very simple application 65% of the
entries in the cache were duplicate.
When the type is a Class with no generic information, the ResolvableType
instance is a simple wrapper around it so we might just as well not cache
it at all as the cost of finding it back from the cache is higher than
creating that simple wrapper.
This commit adds an explicit check; if the type is a simple Class we just
return a "resolved" ResolvableType instance for it. On a few test cases,
this reduces the size of the cache by 85%
Issue: SPR-12275
(cherry picked from commit 6f1acdd)
This commit updates the default location of the SockJS' client library.
The previous location is being retired by the project maintainers.
The new default location is backed by several CDN providers:
* https://cdn.jsdelivr.net/sockjs/0.3.4/sockjs.min.js
See sockjs/sockjs-client#198
Issue: SPR-12254
(cherry picked from commit a6e1c53)
Per the Javadoc for the SocketUtils() constructor, SocketUtils can be
instantiated as a Spring Bean in XML configuration files; however,
SocketUtils is currently abstract which prevents such usage.
This commit removes the 'abstract' declaration thereby allowing
SocketUtils to be instantiated as a Spring bean.
Issue: SPR-12169
(cherry picked from commit 559e81b)
Prior to this commit, setting the parameters used to build the caches
was fragile in static mode as the caches were created right when the
setCacheNames setter was called.
This commit provides a better handling of such arguments and also
provide a way to restore the dynamic mode if necessary.
Issue: SPR-12120
(cherry picked from commit d47c543)
Prior to this commit, when there was a lot of entries in the
ResolvableType.cache HashMap, getting a simple value could
take a lot of time due to a lot of calls to ResolvableType.equals().
ResolvableType.equals() used this.type, getSource(),
this.variableResolver.getSource() and this.componentType, but
ResolvableType.hashCode() used only this.type.
With this commit, ResolvableType.hashCode() now uses the same
fields than ResolvableType.equals().
Performance on the spring-resolvabletype-benchmark project:
- 8000 us before this commit
- 120 us with this commit
Issue: SPR-12122
(cherry picked from commit 7ea69fb)
This commit updates the Gradle build to ensure that *all* source
artifacts are including in published 'sources' jars.
To achieve this, we are now assembling 'sources' jars from
'sourceSets.main.allSource' and not explicitly including or excluding
anything by default.
Issue: SPR-12085
(cherry picked from commit 07629a3c7c)
Also allows for direct setting of a ServerContainer and for custom triggering of endpoint registration.
Issue: SPR-12109
(cherry picked from commit 11805b6)
Prior to this commit, spring-build-src and spring-framework-bom
internal artifacts were published as part of the distribution bundle.
This commit excludes those projects so that those unnecessary artifacts
are no longer shipped.
Issue: SPR-12087
This commit fixes cases where part of the URI was mistaken for the
userinfo when:
* the URI did not contain any path
* the query string contained the "@"
Issue: SPR-11970
(cherry picked from commit a4484bb)
Set `-Xlint:-processing` to disable annotation processing warnings.
Required when building on JDK 1.8.0_20-b05 due to JDK-8039469.
Issue: SPR-11973
(cherry picked from commit ea16ce0)
Prior to this change, AbstractHttpReceivingTransportHandler had a direct
dependency on a Jacckson Exception (checking that exception in a catch
clause). This can cause issues for applications that don't have that
dependency.
This commit removes that direct dependency, still logging the
appropriate log messages using a parent exception (IOException) and
reflection.
Issue: SPR-11963
(cherry picked from commit e549103)
Prior to this commit, the ServletResponseHttpHeaders.get method
would throw an NPE when used under Wildfly 8.0.0.Final and 8.1.0.Final.
This can be traced to WFLY-3474, which throws an NPE when calling
HttpServletResponse.getHeaders("foo") and that header has not
been defined prior to that.
This would cause NPE being thrown by AbstractSockJsService when
checking for CORS HTTP headers in the server HTTP response.
This commit surrounds that method call in AbstractSockJsService and
guards against this issue.
Issue: SPR-11919
(cherry picked from commit 24cdefb)
Prior to this commit, calls to getAllow would fail is setAllow was set
to an EmptyCollection right before.
java.lang.IllegalArgumentException: No enum constant
org.springframework.http.HttpMethod
This commit fixes this by testing the header value for an empty value
before trying to use it to get a value from the Enum.
Issue: SPR-11917
(cherry picked from commit 9919a98)
Sessions connected to a STOMP endpoint are expected to receive some
client messages. Having received none after successfully connecting
could be an indication of proxy or network issue. This change adds
periodic checks to see if we have not received any messages on a
session which is an indication the session isn't going anywhere
most likely due to a proxy issue (or unreliable network) and close
those sessions.
This is a backport for commit:
https://github.com/spring-projects/spring-framework/commit/a3fa9c979777e554efad0df429041767f05dfdb8
Issue: SPR-11884
This change introduces removeOnCancelPolicy on ThreadPoolTaskScheduler
and ScheduledExecutorFactoryBean and sets it to true for SockJS.
This ensures that cancelled tasks are removed immediately to avoid
the "unbounded retention of cancelled tasks" that is mentioned in
the Javadoc ScheduledThreadPoolExecutor:
"By default, such a cancelled task is not automatically removed from
the work queue until its delay elapses. While this enables further
inspection and monitoring, it may also cause unbounded retention of
cancelled tasks. To avoid this, set setRemoveOnCancelPolicy to true,
which causes tasks to be immediately removed from the work queue at
time of cancellation."
This is a backport for:
https://github.com/spring-projects/spring-framework/commit/7441f2301281049cff7224bf650646490388080e
Issue: SPR-11918
This commit also includes lazy resolution support for StandardServletMultipartResolver, along the lines of existing lazy mode in CommonsMultipartResolver.
Issue: SPR-11074
Issue: SPR-11730
(cherry picked from commit 82336c3)
Removal of cached destination is now moved outside the for loop
that removes subscriptions to avoid ConcurrentModificationException.
Also since updateCache is a LinkedHashMap with accessOrder=true,
a simple access with updateCache.get() modify the map.
By iterating over updateCache.entrySet(), we avoid this update.
Issue: SPR-11755
DestinationCache is now used for both plain and pattern
destinations. It stores internally the subscriptions map for each
cached destination. Subscriptions are initially created when there
is no cache for the requested destination, and are updated when
subscriptions change.
Issue: SPR-11657
This change adds support for Reactor 1.1 in spring-messaging in
addition to Reactor 1.0.1 -- whichever is present on the classpath is
used.
Note also the module name change:
reactor-tcp:1.0.1 -> reactor-net:1.1.0
Issue: SPR-11636
Prior to this commit, invoking the getMergedAnnotationAttributes()
method in AnnotationReadingVisitorUtils resulted in mutation of the
internal state of the ASM-based annotation metadata supplied to the
method.
This commit fixes this issue by making a copy of the original
AnnotationAttributes for the target annotation before merging attribute
values from the meta-annotation hierarchy.
This commit also introduces a slight performance improvement by
avoiding duplicate processing of the attributes of the target
annotation.
Issue: SPR-11710
(cherry picked from commit e1720d89fc)
Prior to this commit, only the java.lang.reflect.Method was used to
identify an annotated method. As a result, if different annotations
were placed on different methods (method overriding, interface
implementation) only the first one (cached) was used.
LazyParamAwareEvaluationContext was affected by the exact
same problem and has been also fixed.
Issue: SPR-11692
(cherry picked from commit df34bab)
Changed the value of ScriptUtils.EOF_STATEMENT_SEPARATOR from
"<<< END OF SCRIPT >>>" to "^^^ END OF SCRIPT ^^^" so that the angle
brackets do not have to be escaped in XML configuration files.
Issue: SPR-11687
(cherry picked from commit 1753f5d1ca)
Prior to Spring Framework 4.0.3, it was possible to supply a bogus
statement separator (i.e., a separator string that does not exist in
the configured SQL scripts) to ResourceDatabasePopulator with the side
effect that the entire contents of a script file would be interpreted
as a single SQL statement.
This undocumented feature was never intentional; however, some
developers came to rely on it. Unfortunately, changes made in
conjunction with SPR-9531 and SPR-11560 caused such scenarios to no
longer work.
This commit introduces first-class support for executing scripts which
contain a single statement that spans multiple lines but is not
followed by an explicit statement separator.
Specifically, ScriptUtils.EOF_STATEMENT_SEPARATOR may now be specified
as a 'virtual' statement separator to denote that a script contains a
single statement and no actual separator.
Issue: SPR-11687
(cherry picked from commit cc0ae3a881)
Once the new bootstrap strategy for the TestContext framework (TCF) is
introduced in Spring Framework 4.1, a TestContextBootstrapper will
assume full responsibility for determining what ContextLoader to use as
the default. Consequently, the previous support for supplying the class
name for a custom, default ContextLoader will no longer be applicable.
This commit therefore officially deprecates support for custom, default
ContextLoader class names via the following mechanisms:
- The TestContextManager(Class, String) constructor
- The getDefaultContextLoaderClassName(Class) method in
SpringJUnit4ClassRunner
Issue: SPR-11682
Since a dedicated bootstrap strategy will be introduced for the
TestContext framework in Spring Framework 4.1, this commit officially
deprecates TestContextManager.getDefaultTestExecutionListenerClasses()
as a forewarning to developers or frameworks that may be using this
obscure feature.
Issue: SPR-11679
Prior to this commit, there were numerous places in the reference
manual, where we see output similar to the following:
BeanDefinition`s with the `ApplicationContext
This commit addresses this issue by using unconstrained quotes
(e.g., ++XYZ++s) instead of backticks (e.g., `XYZ`s) when the formatted
text is immediately followed by an "s".
This commit also corrects a few typos and corrects natural English
pluralization of Java code elements in the reference manual where
appropriate -- for example, "@Controllers" becomes "@Controller classes,
etc.
Issue: SPR-11650
(cherry picked from commit 0eba1f818f)
Prior to this commit, @SubscribeMapping mapped methods (backed with
@SendTo* annotations, or not) would send MESSAGEs with the wrong
destination. Instead of using the original SUBSCRIBE destination, it
would use the lookup path computed from the configured prefixes in the
application.
This commit fixes this issue - now @SubscribeMapping MESSAGEs use the
original SUBSCRIBE destination.
Issue: SPR-11648
This change ensures correct default settings for message channel's
created through the websocket message-broker XML namespace even
when the channels are customized for other reasons -- e.g. to add
a channel interceptor.
Issue: SPR-11623
Ensure the Standard- and the JettyWebSocketSession can return the
Principal and accepted WebSocket sub-protocol even after the
session is closed.
Issue: SPR-11621
Proactively notify all active WebSocket sessions when a shutdown is
progress. Sessions then can ignore further attempts to send messages
and also stop stop trying to flush messages right away.
Prior to this commit, the Spring TestContext Framework did not support
the declaration of both 'locations' and 'classes' within
@ContextConfiguration at the same time.
This commit addresses this in the following manner:
- ContextConfigurationAttributes no longer throws an
IllegalArgumentException if both 'locations' and 'classes' are
supplied to its constructor.
- Concrete SmartContextLoader implementations now validate the
supplied MergedContextConfiguration before attempting to load the
ApplicationContext. See validateMergedContextConfiguration().
- Introduced tests for hybrid context loaders like the one used in
Spring Boot. See HybridContextLoaderTests.
- Updated the Testing chapter of the reference manual so that it no
longer states that locations and classes cannot be used
simultaneously, mentioning Spring Boot as well.
- The Javadoc for @ContextConfiguration has been updated accordingly.
- Added hasLocations(), hasClasses(), and hasResources() convenience
methods to MergedContextConfiguration.
Issue: SPR-11634
(cherry picked from commit 1f017c4acb)
Prior to this commit, if multiple test methods were executed in a
subclass of AbstractTestNGSpringContextTests annotated with
@WebAppConfiguration, then injected Servlet API mocks would only
reference the mocks created for the first test method. Subsequent test
methods could therefore never reference the current mocks, and there
was a discrepancy between the state of the injected mocks and the mock
set in the RequestContextHolder.
This commit addresses this issue by ensuring that dependencies
(including updated mocks) are injected into the test instance before
the next test method if the ServletTestExecutionListener resets the
request attributes in RequestContextHolder.
Issue: SPR-11626
(cherry picked from commit c38600762d)
Although the need to map the ASYNC dispatcher type to a Filter was
already mentioned, it wasn't very prominent and can be quite critical
in some cases. This change addresses that.
Issue: SPR-10440
This reverts commit 3474afb165.
Unfortunately this change is likely to cause issues for applications
that use regular expressions in a URI variable. I think we will have
to leave at: if there are any dots in the last segment of the
request path, regardless of whether they're in a URI var or not,
the suffix pattern match is off.
Issue: SPR-11532
Improved the SAX to StAX (and vice-versa) bridge exposed via StaxUtils.
The old integration had some issues with namespace declaration
attributes, brought to light in a XMLUnit upgrade.
Issue: SPR-11549
Changes introduced in conjunction with issue SPR-11475 altered the
behavior of StandardAnnotationMetadata such that annotations could be
detected on superclasses, specifically in the case where the
AnnotatedElementUtils.getAllAnnotationAttributes() method is invoked to
obtain multiple annotations of the same type (on the lowest level in the
class hierarchy), as is the case for @Profile and @Conditional.
This commit partially reverts these changes as follows:
- All methods in AnnotatedElementUtils now set the
traverseClassHierarchy to false, thereby effectively reverting the
changes made in commit 1d30bf83a0.
Note, however, that the changes made to AnnotationUtils remain in
place.
- Introduced tests in AnnotationMetadataTests that verify behavior
present in Spring Framework 4.0.2 and earlier.
- Updated tests in AnnotatedElementUtilsTests so that they pass against
the reverted changes (i.e., align with the behavior present in Spring
Framework 4.0.2 and earlier).
- Refined Javadoc in AnnotationMetadata with regard to annotations
being "present" vs. "defined".
- Refined Javadoc in AnnotatedTypeMetadata.
Issue: SPR-11475, SPR-11595
This commit validates that the payload type of the message is
assignable to the one declared in the method signature. If that
is not the case, a meaningful exception message is thrown with
the types mismatch.
Prior to this commit, only the Message interface could be defined
in the method signature: it is now possible to define a sub-class
of Message if necessary which will match as long as the Message
parameter is assignable to that type.
Issue: SPR-11584
BufferingStompDecoder message buffer size limit can now be configured
with JavaConfig MessageBrokerRegistry.setMessageBufferSizeLimit() or
with XML <websocket:message-brocker message-buffer-size="">.
Issue: SPR-11527
Before this change the StompDecoder decoded and returned only the first
Message in the ByteBuffer passed to it. So to obtain all messages from
the buffer, one had to loop passing the same buffer in until no more
complete STOMP frames could be decoded.
This chage modifies StompDecoder to return List<Message> after
exhaustively decoding all available STOMP frames from the input buffer.
Also an overloaded decode method allows passing in Map that will be
populated with any headers successfully parsed, which is useful for
"peeking" at the "content-length" header.
This change also adds a BufferingStompDecoder sub-class which buffers
any content left in the input buffer after parsing one or more STOMP
frames. This sub-class can also deal with fragmented messages,
re-assembling them and parsing as a whole message.
Issue: SPR-11527
This commit adds new MockHttpServletRequestBuilder constructors
with an URI parameter in addition to the URL template + URL variables
existing ones.
It gives more control on how the URL is built, allowing for example to
use URL variables containing '/' character with proper encoding.
Issue: SPR-11441
Prior to this commit, the ShallowEtagHeaderFilter did not use the
content length given by the content generator to set the
ByteArrayOutputStream's buffer size.
This can lead to performance issues for large content since the buffer
grows as the content is being written.
This commit adds a new ByteArrayOutputStream variant called
ResizableByteArrayOutputStream. This implementation has public methods
for modifying the internal buffer size and does not synchronize on
buffer access.
This commit also make use of this new variant in
ShallowEtagHeaderFilter.
Issue: SPR-8271
When a send timeout is detected, the WebSocket session is now closed
with a custom close status that indicates so. This allows skipping
parts of the close logic that may cause further hanging.
Issue: SPR-11450
Before this change SockJsSession implementations of WebSocketSession
used synchronization around its method implementations protecting
internal state and ensuring only a single thread is sending messages
at a time.
A WebSocketSession is generally expected to be used from one thread
at a time and now that application messages are sent through
ConcurrentWebSocketSessionDecorator, there is no concern about
application messages sent from the different threads.
While there are some remaining concerns, those can be addressed
without using the synchronized keyword. This change removes it from
the methods of all SockJS session implementations.
Issue: SPR-11450
Before this change the decorator ensured that for a specific WebSocket
session only one thread at a time can send a message. Other threads
attempting to send would have their messages buffered and each time
that occurs, a check is also made to see if the buffer limit has been
reached or the send time limit has been exceeded and if so the session
is closed.
This change adds further protection to ensure only one thread at a time
can perform the session limit checks and attempt to close the session.
Furthermore if the session has timed out and become unresponsive,
attempts to close it may block yet another thread. Taking this into
consideration this change also ensures that state associated with the
session is cleaned first before an attempt is made to close the session.
Issue: SPR-11450
Since we now wrap the WebSocketSession with a concurrent decorator, the
synchronized keyword around message sending needed to be removed.
Issue: SPR-11586
- AnnotationAttributesReadingVisitor no longer processes annotations
from the java.lang.annotation package.
- Simplified logic in AnnotationReadingVisitorUtils
getMergedAnnotationAttributes().
Issue: SPR-11574
If a payload is present but conversion returns null (meaning no
converter knows how to convert), raise a MessageConversionException
that provides information about the type we were trying to convert
to and the message itself whose headers (namely content-type) contain
crucial information required to debug the problem.
Issue: SPR-11577
Prior to this commit, configuring a custom handshakeHandler when setting
up a stomp-endpoint with SockJS would not be taken into account:
<websocket:stomp-endpoint path="/foo">
<websocket:handshake-handler ref="customHandler"/>
<websocket:sockjs/>
</websocket:stomp-endpoint>
This commit fixes this by creating and registering a
WebsocketTransportHandler (with this handshakeHandler) as a
transportHandler override for the SockJSService.
Issue: SPR-11568
When no @Payload is provided, it is equivalent to @Payload with default
attribute values. Since the default value of required=true, then
an argument that's not annotated is required.
A payload that is required will now throw an appropriate exception
regardless of if a conversion is required or not.
isEmptyPayload now takes the payload instead of the message
so that both the original payload and the converted payload, if
necessary, share the same logic.
JSR-303 validation is now consistently applied.
Issue: SPR-11577
Prior to this commit, the codebase was using a mix of log4j.xml
and log4j.properties for test-related logging configuration. This
can be an issue as log4j takes the xml variant first when looking
for a default bootstrap configuration.
In practice, some modules declaring the properties variant were
taking the xml variant configuration from another module.
The general structure of the configuration has also been
harmonized to provide a standard console output as well as an
easy way to enable trace logs for the current module.
The clientInboundChannel and clientOutboundChannel now use twice
the number of available processors by default to accomodate for some
degree of blocking in task execution on average.
In practice these settings still need to be configured explicitly in
applications but these should serve as better default values than
the default values in ThreadPoolTaskExecutor.
Issue: SPR-11450
Prior to this commit, Spring supported meta-annotation attribute
overrides in custom composed annotations with reflection-based
annotation processing but not with ASM-based annotation processing.
This commit ensures that meta-annotation attribute overrides are
supported in AnnotationMetadataReadingVisitor.getAnnotationAttributes().
Issue: SPR-11574
This change exposes the WebSocketSession attributes through a message header.
The StompSubProtocolHandler adds this to incoming messages.
For now messaging handling methods can access the map via @Header, e.g.:
@Header(StompHeaderAccessor.SESSION_ATTRIBUTES) Map<String, Object> attrs) {
Issue: SPR-11566
This commit sets the DB_CLOSE_ON_EXIT flag to false for embedded H2
databases loaded using H2EmbeddedDatabaseConfigurer (i.e., via Spring's
<jdbc:embedded-database /> XML namespace, EmbeddedDatabaseBuilder,
EmbeddedDatabaseFactory, and EmbeddedDatabaseFactoryBean).
Issue: SPR-11573
This commit improves the configurability of EmbeddedDatabaseBuilder by
exposing the following new configuration methods.
- setDataSourceFactory(DataSourceFactory)
- addScripts(String...)
- setScriptEncoding(String)
- setSeparator(String)
- setCommentPrefix(String)
- setBlockCommentStartDelimiter(String)
- setBlockCommentEndDelimiter(String)
- continueOnError(boolean)
- ignoreFailedDrops(boolean)
If more fine grained control over the configuration of the embedded
database is required, users are recommended to use
EmbeddedDatabaseFactory with a ResourceDatabasePopulator and forego use
of the builder.
Issue: SPR-11410
Prior to this commit, AnnotationAttributesReadingVisitor treated Class
annotation attributes as Strings instead of Classes. As a result,
look-ups for Class attributes in meta-annotations failed with an
IllegalArgumentException.
This commit fixes this issue by consistently treating Class attributes
as Classes in AnnotationAttributesReadingVisitor.
Issue: SPR-11557
In Tiles v3 integration, underscores in filenames are
intended to identify the definition locale. This behavior
is now documented in order to avoid unexpected results
with filenames like tiles_definitions.xml.
This commit also updates Tiles v2 references to Tiles v3
in the Spring reference documentation.
Issue: SPR-11491
This commit updates reference documentation in order
to specify that ServletServerContainerFactoryBean is for
server side use and WebSocketContainerFactoryBean
for client side use.
Issue: SPR-11565
Use a custom ConfigReader to enforce the use of SynchronousDispatcher
and no other dispatchers otherwise created by default. Avoids the
creation thread pools never to be used.
Ignore DISCONNECT messages if already disconnected. This can occur if
the client explicitly sends a DISCONNECT frame and then closes the
socket quickly. The closing of the WebSocket sessions also sends a
DISCONNECT upstream to ensure the broker is aware.
In ScriptUtils and related classes, SQLExceptions are now caught and
wrapped in ScriptExceptions wherever feasible.
Affected "throws" declarations have also been revised as appropriate.
Issue: SPR-11564
This commit continues the work began in SPR-9531 as follows.
- ScriptException now extends DataAccessException.
- DatabasePopulator.populate() now explicitly throws ScriptException.
- Introduced UncategorizedScriptException.
- DatabasePopulatorUtils.execute() now throws an
UncategorizedScriptException instead of a
DataAccessResourceFailureException.
Issue: SPR-11564
Even though Netty is a required dependency of reactor-tcp at present,
there is no hard dependency in the spring-messaging Reactor-based
implementation.
Configure explicitly use of SynchronousDispatcher instead of the one
used otherwise by default (RingBufferDispatcher). As a result TCP
optations are now scoped to Netty's threads.
Remove Environment field. It is no longer required to shut it down
since we're now using SynchronousDispatcher by default.
Replace connection.in() with connection.consume() when composing
connection handling. The former creates a Stream for further composing,
e.g. via map(), filter() but all we need is to read a message.
Provide additional constructor that aceepts a pre-configured Reactor
TcpClient instance.
Issue: SPR-11531
Add accessor for brokerAvailable in AbstractBrokerMessageHandler
Ensure brokerAvailable is set even if eventPublisher is not
Add tests BrokerMessageHandlerTests
Turn off brokerAvailable when StompBrokerRelayMessageHandler stops
Actually stop message handling when brokerAvailable=false
Improve log messages
Issue: SPR-11563
Changes made in conjunction with SPR-9531, introduced a regression with
regard to support for using a single newline character as the statement
separator within SQL scripts. Investigation of the cause of this issue
resulted in the discovery of another, similar issue: support for
multiple newlines as a statement separator has been broken for years
but has gone unnoticed until now.
The reason that both of these issues have gone unnoticed is a result of
the fact that the test suite only executes SQL script integration tests
against HSQL DB, and HSQL does not care if two statements occur on the
same line; whereas, the H2 database will throw an exception if multiple
statements are included on the same line when executing an update.
This commit addresses both of these issues and provides further
enhancements to Spring's SQL script support as follows.
- ScriptUtils now properly checks if the supplied script contains the
custom statement separator or default separator before falling back
to the 'fallback' separator (i.e., newline).
- Introduced FALLBACK_STATEMENT_SEPARATOR constant in ScriptUtils.
- ScriptUtils.readScript() no longer omits empty lines from the input
file since a statement separator string may in fact be composed of
multiple newline characters.
- Introduced overloaded variants of splitSqlScript() and
executeSqlScript() in ScriptUtils with smaller argument lists for
common use cases.
- Extracted AbstractDatabasePopulatorTests from DatabasePopulatorTests
and introduced concrete HsqlDatabasePopulatorTests and
H2DatabasePopulatorTests subclasses for testing against HSQL and H2.
- Split ScriptUtilsTests into ScriptUtilsUnitTests and
ScriptUtilsIntegrationTests for faster builds.
Issue: SPR-11560
This commit introduces unit and integration tests that attempt to
reproduce the issue claimed by the reporter in SPR-11557. However, the
tests pass without any problems.
Issue: SPR-11557
Restored original singleton-only adaptInnerBeanName behavior, avoiding endless unique name calculation for every single prototype instance. Non-named inner BeanDefinition objects now suffixed with an identity hex code analogous to non-named XML bean definitions, largely avoiding naming collisions to begin with. After SPR-11246, post-processors can deal with unstable classes per bean name, so occasional collisions aren't a hard problem anymore.
Issue: SPR-11545
This commit continues the work in the previous commit as follows:
- Introduced an exception hierarchy for exceptions related to SQL
scripts, with ScriptException as the base.
- CannotReadScriptException and ScriptStatementFailedException now
extend ScriptException.
- Introduced ScriptParseException, used by ScriptUtils.splitSqlScript().
- DatabasePopulatorUtils.execute() now explicitly throws a
DataAccessException.
- Polished Javadoc in ResourceDatabasePopulator.
- Overhauled Javadoc in ScriptUtils and documented all constants.
- Added missing @author tags for original authors in ScriptUtils and
ScriptUtilsTests.
- ScriptUtils.splitSqlScript() now asserts preconditions.
- Deleted superfluous methods in ScriptUtils and changed method
visibility to private or package private as appropriate.
- Deleted the ScriptStatementExecutor introduced in the previous
commit; ScriptUtils.executeSqlScript() now accepts a JDBC Connection;
JdbcTestUtils, AbstractTransactionalJUnit4SpringContextTests, and
AbstractTransactionalTestNGSpringContextTests now use
DatabasePopulatorUtils to execute a ResourceDatabasePopulator instead
of executing a script directly via ScriptUtils.
- Introduced JdbcTestUtilsIntegrationTests.
Issue: SPR-9531
Prior to this commit neither ResourceDatabasePopulator nor
JdbcTestUtils properly supported multi-line comments (e.g., /* ... */).
Secondarily there has developed a significant amount of code
duplication in these two classes that has led to maintenance issues
over the years.
This commit addresses these issues as follows:
- Common code has been extracted from ResourceDatabasePopulator and
JdbcTestUtils and moved to a new ScriptUtils class in the
spring-jdbc module.
- Relevant test cases have been migrated from JdbcTestUtilsTests to
ScriptUtilsTests.
- ScriptUtils.splitSqlScript() has been modified to ignore multi-line
comments in scripts during processing.
- ResourceDatabasePopulator supports configuration of the start and end
delimiters for multi-line (block) comments.
- A new test case was added to ScriptUtilsTests for the new multi-line
comment support.
Issue: SPR-9531
After this change dots inside URI variables in a request mapping
pattern are ignored and no longer considered an indication that
the pattern contains a suffix itself.
Issue: SPR-11532
Prior to this commit, the `relay-port` attribute of the
`<websocket:stomp-broker-relay />` tag was of type `xsd:int`.
This prevents developers from using `PropertyPlaceholderConfigurer`,
even though this configuration key is a good candidate for such use
(this value depends on prod/staging/etc environment).
This commit changes that type to `xsd:string`.
Issue: SPR-11537
After this change if a content-length header is provided but is less
than 0 or cannot be parsed as a number, it is ignored and the body
is read sequentially, i.e. until we reach a null byte terminator.
This provides better protection against clients that may set the
content-length header in error.
Issue: SPR-11528
Fix NPE exception when closing TcpConnection
Ensure a ConnectionHandler is cleared when a TcpConnection is closed
(at the same time), logging an exception if the closing fails.
Improve error messages.
Issue: SPR-11531
Since HttpClient 4.3, custom configuration such as the connection
timeout and the socket timeout are set in a RequestConfig object
stored in the HttpContext.
Unfortunately, older HttpClients are not supporting this
infrastructure and new clients throw an exception when the
deprecated API is used.
This commit detects if the client is an "old" implementation and
set the configuration through the deprecated means to restore
full backward compatibility with these features.
Issue: SPR-11442
Also uses addAll instead of iteration over untyped collection now, supporting optimized addAll in target collection type, and avoids repeated getElementTypeDescriptor calls.
Issue: SPR-11479
This commit provides a first attempt at defining a standard code
formatting scheme for the Spring Framework in Intellij IDEA.
These are the major changes compared to standard settings:
* default indent option to use tab character instead of space for
all languages
* one space before the left brace of an array initializer
* keep when reformating: multiple expressions in one line,
simple blocks in one line
* method declaration parameters: do not align when multiline
* else, catch and finally on new line
* keep one space before } (solely use to keep the space between the
end of the last method and the end of the class)
* minimum blank line after class header 0 (instead of 1)
* Disabled Javadoc formatting
* class count to trigger static import to 50 (to prevent
import org.foo.*; instead of listing the classes of org.foo)
* changed the import sequence to import in the following order:
static imports, java.*, javax.*, others, org.springframework.*. Each
sequence is separated by a space
In addition to the target parameter values (SPR-9657), the target
parameter names must also be decoded to be able to match
them to the parameter names of incoming requests.
Issue: SPR-11504
When a node of an SPeL expression was a call to a bean referenced
in a method argument, the expression was resolved twice.
The resolved arguments are now specified to MethodValueRef instead
of resolving the arguments again in the constructor
Issue: SPR-11445
After this change, AbstractSockJsService does not add CORS headers if
the response already contains an "Access-Control-Allow-Origin" header.
Essentially it backs off assuming CORS headers are handled centrally
e.g. through a Filter.
In order to support this, the ServletServerHttpResponse now returns an
instance of HttpHeaders that also provides access to headers already
present in the HttpServletResponse.
Issue: SPR-11443
This commit aligns our include and exclude filters for test classes
with Gradle's standard patterns. Specifically, our patterns now end
with ".class" instead of ".*".
The aforementioned change makes the exclusion of inner classes
unnecessary. Thus, patterns for test classes ending with "TestCase" or
"TestSuite" have been deleted.
Furthermore, the include and exclude patterns previously used in the
spring-test module made it impossible for the
FailingBeforeAndAfterMethodsTests class in the 'testng' package to ever
be executed by the build. This has been addressed by renaming our JUnit
and TestNG variants of FailingBeforeAndAfterMethodsTests and moving the
TestNG variant into the 'junit' package so that it can be picked with
our standard include pattern for JUnit-based tests.
Fixed through falling back to the raw parameter type in the TypeDescriptor(MethodParameter) constructor, properly detecting the vararg array even in case of an unresolvable type variable, and through restoring getElementTypeDescriptor's original behavior for arrays, i.e. always returning a non-null descriptor.
Issue: SPR-11494
This change removes the use of a CountDownLatch to wait for the
asynchronously computed controller method return value. Instead we
check in a loop every 200 milliseconds if the result has been set.
If the result is not set within the specified amount of time to wait
an IllegalStateException is raised.
Additional changes:
- Use AtomicReference to hold the async result
- Remove @Ignore annotations on AsyncTests methods
- Remove checks for the presence of Servlet 3
Issue: SPR-11516
This commit adds tests for TransactionAwareCacheDecorator. In
particular, the put/evict behaviour when the operation is invoked in
the course of Spring-managed transaction.
- Deleted empty AbstractWebSocketClientTests class.
- AbstractServletHandlerMethodTests and AbstractHttpRequestTests are
now actually declared as abstract.
- The following classes are not abstract but currently have an
"Abstract" prefix and therefore get ignored by the Gradle build.
This commit renames each of these by deleting the "Abstract" prefix.
- AbstractFlashMapManagerTests
- AbstractMappingContentNegotiationStrategyTests
- AbstractSockJsServiceTests
- AbstractWebSocketHandlerRegistrationTests
Prior to this commit, an array argument was not handled properly in
SimpleKey#equals and SimpleKey#hashCode. As a result, two method
invocations with the same array argument lead to two different keys
and therefore two different entries in the cache.
This commit uses deepEquals and deepHashCode to properly handle
methods that have arguments that are array types.
Issue: SPR-11505
Prior to this commit, the test results for JUnit tests overwrote the
results for TestNG tests in the HTML test report generated by the
Gradle build.
This commit addresses this problem by introducing a new
'aggregateTestReports' task in the spring-test module and making the
'check' task depend on it.
To see aggregated reports in the spring-test module, execute either
`gradle build` or `gradle check`. Executing `gradle test` alone will
not trigger test report aggregation.
Issue: SPR-11509
The Servlet API does not provide notifications when a client
disconnects, see see https://java.net/jira/browse/SERVLET_SPEC-44.
Therefore network IO failures may occur simply because a client has
gone away. Before this change that could fill logs with unnecessary
stack traces.
After this change we make a best effort to identify such network
failures, on a per-server basis (tested with Jetty, Tomcat, Glassfish,
and WildFly), and log them under a separate log category.
A simple one-line message is logged at DEBUG level (i.e. no stack trace)
while a full stack trace is shown at TRACE level.
Issue: SPR-11438
The breakout box named "Constructor-based or setter-based DI?" in the
reference manual currently recommends the use of setter injection.
This commit refines this text to align with current best practices and
now favors constructor injection over setter injection.
Issue: SPR-11459
Allow Jaxb2RootElementHttpMessageConverter subclasses to customize
the {@link Marshaller} and the {@link Unmarshaller} created by the
message converter.
Issue: SPR-11488
Prior to this commit, one had to provide her own
RequestMappingHandlerMapping instance (i.e extend
WebMvcConfigurationSupport and override the requestMappingHandlerMapping
method) in order to customize path matching properties on that bean.
Since SPR-10163, XML config users can do that using the
<mvc:path-matching/> XML tag. This commit adds the same feature to MVC
Java config with a PathMatchConfigurer.
Issue: SPR-11486
Before this change CompositeMessageConverter had a ContentTypeResolver
field that was in turn set on all contained converters.
After this change that field is removed and effectively
CompositeMessageConverter is a simple container of other converters.
Each converter in turn must have been configured with a
ContentTypeResolver.
Doing so means it is less likely to have unexpected consequences when
configuring converters, the ContentTypeResolver set in the composite
converter overriding the one configured in a contained converter.
Also commit 676ce6 added default ContentTypeResolver initialization
to AbstractMessageConverter, which ensures that converters are still
straight forward to configure.
Issue: SPR-11462
Previously AbstractMessageConverter did not have a ContentTypeResolver
configured by default. However the Java config and XML namespace in
spring-messaging and spring-websocket always configured one.
This change ensures every AbstractMessageConverter is configured with an
instance of DefaultContentTypeResolver by default. This makes sense since
all the resolver does is make an attempt to find a content type to use
for matching. If it can't it returns null and it's up to the converter
to decide whether it can convert or not.
Issue: SPR-11462
AbstractMessageConverter now supports a strictContentTypeMatching mode
in which the content type of a message must be resolved to a (non-null)
value that matches one of the configured supported MIME types in order
for the converter to be used.
Issue: SPR-11463
This commit:
* adds a reference to the Spring Code Style wiki page in the main
CONTRIBUTING document
* updates the link to the Spring team in README
* adds a note regarding Intellij IDEA 13 issues
The previous commit introduced a dependency on
Class.getDeclaredAnnotation() which is a Java 8 API.
This commit refactors AnnotationUtils.findAnnotation(Class, Class, Set)
to use Class.getAnnotation() in conjunction with
isAnnotationDeclaredLocally() in order to achieve the same desired
behavior.
Issue: SPR-11475
Prior to this commit, the implementations of findAnnotation() in
AnnotationUtils and getAnnotationAttributes() in AnnotatedElementUtils
favored inherited annotations and inherited composed annotations over
composed annotations that are declared closer to the starting class
passed to these methods.
This commit addresses this issue as follows:
- Refactored AnnotationUtils to use getDeclaredAnnotation() and
getDeclaredAnnotations() instead of getAnnotation() and
getAnnotations() where appropriate.
- AnnotatedElementUtils.doProcess() supports a traverseClassHierarchy
flag to control whether the class hierarchy should be traversed,
using getDeclaredAnnotations() instead of getAnnotations() if the
flag is true.
- Overhauled Javadoc in AnnotatedElementUtils.
Issue: SPR-11475
Ant 1.9.2 is packaged with Gradle since release 1.10. Since the Spring
Framework build now uses Gradle 1.11, there is no longer a need for the
"javac1.7" build compiler work-around for the spring-oxm module.
This commit introduces a new isInJavaLangAnnotationPackage(Annotation)
method in AnnotationUtils. This method is now used in AnnotationUtils,
AnnotatedElementUtils, and MetaAnnotationUtils to ensure that search
algorithms do no search for meta-annotations on annotations in the
"java.lang.annotation" package.
The following are some empirical results from this change:
- The number of times that the findAnnotation(Class,Class,Set) method in
AnnotationUtils is recursively invoked while executing
AnnotationUtilsTests drops from 51 to 29.
- The number of times that the process(AnnotatedElement) method in
AnnotationUtils.AnnotationCollector is recursively invoked while
executing AnnotationUtilsTests.getRepeatableFromMethod() drops
from 16 to 2.
- The number of times that the doProcess() method in
AnnotatedElementUtils is recursively invoked while executing the
"getAnnotationAttributes() On MetaCycleAnnotatedClass with missing
target meta-annotation" test in AnnotatedElementUtilsTests drops
from 23 to 5.
- The number of times that the findAnnotationDescriptor(Class,Set,Class)
method in MetaAnnotationUtils is recursively invoked while executing
the "findAnnotationDescriptor() on MetaCycleAnnotatedClass with
missing target meta-annotation" test in MetaAnnotationUtilsTests drops
from 16 to 8.
Issue: SPR-11483
Clarify ability to use @MessageMapping methods on both @Controller as
well as @RestController.
Add section on configuring connections (including credentials) to the
message broker and clarify the use of the login/passcode headerers of
the STOMP CONNECT frame.
Add note on when to add the reactor-tcp dependency.
Issue: SPR-11464, SPR-11436, SPR-11449
Prior to this commit, the following methods in ContextLoaderUtils
treated a composed @ContextConfiguration annotation (i.e., a custom
annotation that is meta-annotated with @ContextConfiguration) as the
"declaring class" instead of the actual test class.
- resolveContextConfigurationAttributes()
- resolveContextHierarchyAttributes()
As a consequence, if @ContextConfiguration is used as a
meta-annotation, the meta-annotated class is stored as the
"declaringClass" in ContextConfigurationAttributes. Thus, when a
ContextLoader (or SmartContextLoader) attempts to detect default
resource locations or configuration classes, it does so for the
composed annotation class instead of for the declaring test class.
This commit ensures that ContextLoaders are supplied the declaring test
class instead of the composed annotation class in such use cases.
Issue: SPR-11455
Before this change tests on the CI server showed the following message:
Store limit is 102400 mb, whilst the data directory:
/opt/.../KahaDB only has 74810 mb of usable space
This change turns off store persistence and also explicitly sets the
limit on memory usage.
Prior to this commit, the findAnnotationDescriptor() and
findAnnotationDescriptorForTypes() methods in MetaAnnotationUtils only
supported a single level of meta-annotations. In particular, this kept
the following annotations from being used as meta-annotations on
meta-annotations:
- @ContextConfiguration
- @ContextHierarchy
- @ActiveProfiles
- @TestExecutionListeners
This commit alters the search algorithms used in MetaAnnotationUtils so
that arbitrary levels of meta-annotations are now supported for the
aforementioned test-related annotations.
Issue: SPR-11470
Prior to this commit, if @ActiveProfiles were used as a meta-annotation
on a composed annotation, then the composed annotation's class would be
passed to the ActiveProfilesResolver.resolve() method instead of the
test class, which breaks the contract for ActiveProfilesResolver.
This commit addresses this issue by ensuring that the actual test class
is always passed to ActiveProfilesResolver.resolve().
Issue: SPR-11467
Prior to this commit, AnnotationUtils.findAnnotation(Class, Class)
claimed to recursively search through annotations; however, only one
level of annotations was supported by the algorithm.
This commit alters the search algorithm so that nested meta-annotations
(i.e., meta-annotations on meta-annotations) are also supported.
Issue: SPR-11448
Update ConcurrentReferenceHashMap to protect against references that
have been garbage collected but for some reason do not appear as a
`pollForPurge` result.
Also added purgeUnreferencedEntries() method to allow for programmatic
cleanup.
Issue: SPR-11440
Prior to this commit, one couldn't set the virtualHost property on
StompBrokerRelayMessageHandler via JavaConfig, since
StompBrokerRelayRegistration's API didn't offer that possibility.
This commit adds a new method in StompBrokerRelayRegistration's fluent
API to set the virtualHost used by StompBrokerRelayMessageHandler.
Note: this property is already configurable via xml config.
Issue: SPR-11433
Prior to this commit, all 2xx HTTP responses were eligible for ETag
generation in ShallowEtagHeaderFilter. In some cases, this would use
CPU resources for no reason since HTTP clients would not use ETags.
This commit is an optimization and restricts ETags generation in cases
where (all conditions must be met):
- response has a 2xx status
- request is a GET
- response does not contain "no-store" in its "Cache-Control" header
Issue: SPR-11110
Before this change, when a client subscribed to a "user" destination
(e.g. /user/foo), actual messages received in response to that
subscription contained the server-translated, unique user destination
(e.g. /foo-user123).
This is not an issue for clients such as stomp.js since the
subscription is unique and sufficient to match subscription responses.
However, other STOMP clients do additional checks on the destination
of the subscription and the response.
This change ensures that messages sent to clients on user destionations
always contain a destination that matches the one on the original
subscription.
Issue: SPR-11423
This commit continues the work for fixing memory leaks resulting from
CGLIB subclass generation for beans relying on method injection.
- Set proxy callbacks on the CGLIB Factory (i.e., the instance) instead
of in the generated subclass (i.e., via the Enhancer).
- Convert private inner classes in CglibSubclassingInstantiationStrategy
to private static classes in order to avoid unnecessary coupling to
classes generated using CGLIB.
- Tidy up XmlBeanFactoryTests.
- Update logic in serializableMethodReplacerAndSuperclass() so that it
finally aligns with the decision made for SPR-356.
Issue: SPR-10785, SPR-356
This commit introduces a test in XmlBeanFactoryTests that verifies that
CGLIB generated subclasses for method injected beans are reused across
bean factories for identical bean definitions. In other words, by
verifying that the same CGLIB generated class is reused for identical
bean definitions, we can be certain that Spring is no longer generating
identical, duplicate classes that consume memory in the VM.
Issue: SPR-10785, SPR-11420
Prior to this commit, the inclusion of the 'overloaded' flag in the
implementations of equals() and hashCode() in MethodOverride could lead
to adverse effects in the outcome of equals() in AbstractBeanDefinition.
For example, given two bean definitions A and B that represent the
exact same bean definition metadata for a bean that relies on method
injection, if A has been validated and B has not, then A.equals(B) will
potentially return false, which is not acceptable behavior.
This commit addresses this issue by removing the 'overloaded' flag from
the implementations of equals() and hashCode() for MethodOverride.
Issue: SPR-11420
Before this change, issues surrounding the use of @Controller's in
combination with AOP proxying, resulted in an IllegalArgumentException
when trying to invoke the controller method.
This change detects such cases proactively and reports them with a
clear recommendation to use class-based proxying when it comes to
@Controller's. This is the most optimcal approach for controllers
in many respects, also allows @MVC annotations to remain on the
class.
The documentation has also been updated to have a specific section
on @Controller's and AOP proxying providing the same advice.
Issue:SPR-11281
Before this change the getPathWithinServletMapping method of
UrlPathHelper could not work properly when a default servlet mapping
(i.e. "/") was used in combination with urlDecode=false. The fact that
the getServletPath() method of HttpServletRequest always returns a
decoded path was getting in the way.
Although there is no way to check Servlet mappings through the Servlet
API, this change aims to detect the given scenario and returns the full
path following the context path thus avoiding URL decoding.
Note that the same can be achieved by setting urlDecode=false and
alwaysUseFullPath=true. However this change ensures that urlDecode
works properly without having to know that.
Issue: SPR-11101
mimeType is not a valid method on ContentResultMatche.
I have signed and agree to the terms of the SpringSource Individual
Contributor License Agreement.
- Consistent importing of org.junit.Assert.*;
- Proper declaration of expected exceptions via @Test(expected).
- Renamed SpEL ExpressionTestCase to AbstractExpressionTests.
- Formatting and test method naming conventions.
AssertionFailedError was thrown by JUnit 3.8. Since RedirectViewTests
has been upgraded to JUnit 4, now standard java.lang.AssertionErrors are
thrown. Thus, attempting to catch an AssertionFailedError is futile.
Of course, since these tests have not been failing, it is likely a moot
point, but changing the try-catch blocks to catch a possible
AssertionError can't be a bad thing.
Before this change MvcUriComponentsBuilder could not create a
UriComponentsBuilder for methods where the mapping has a URI variable
and no matching method argument for it.
For example a URI variable may be in the type-level mapping but not
all methods may have an @PathVariable argument for it.
This fix addresses the shortcoming such that MvcUriComponentsBuilder
expands the method argument values available to it and leaves remaining
URI variables to be further expanded via UriComponents.expand().
Issue: SPR-11391
This commit fixes some formatting issues in the 'TestContext Framework
support classes' section of the reference manual that were introduced
in the conversion from DocBook to AsciiDoc.
Headers provided to the SimpMessagingTemplate's convertAndSend methods
are now automatically moved into the "nativeHeaders" sub-map. This
ensures the headers will go out with the STOMP message and be received
by subscribers.
Issue: SPR-11387
This commit introduces static verify() and reset() methods in
AnnotationDrivenStaticEntityMockingControl for programmatic control
on the mock.
Issue: SPR-11395
Fix testPrintFormattedScopedAttributeResult test in org.springframework.web.servlet.tags.EvalTagTests class by making an assertion which is not locale dependent.
Issue: SPR-11392
This commit improves the test coverage for AbstractMethodMockingControl
by introducing tests that verify expected behavior for:
- reentrant method invocations via public methods
- reentrant method invocations via private methods
- test methods that do not set expectations or invoke playback()
- test methods that throw exceptions
For a more complete discussion of "after" vs. "after returning" advice
within AbstractMethodMockingControl, see the Javadoc in the tests.
This commit fixes the off-by-one regression accidentally introduced in
commit 55961544a7.
Specifically, this fix ensures that the correct recorded call is
indexed in the 'calls' list in the implementation of
AbstractMethodMockingControl.Expectations.nextCall().
In addition, this commit improves the Javadoc for
AbstractMethodMockingControl, @MockStaticEntityMethods, and
AnnotationDrivenStaticEntityMockingControl and introduces a proper
toString() implementation for the internal Expectations.Call class in
AbstractMethodMockingControl. Furthermore, code from the obsolete
Delegate test class has been inlined in
AnnotationDrivenStaticEntityMockingControlTests.
Issue: SPR-11385, SPR-10885
Prior to this commit several test classes named "*Test" were not
recognized as tests by the Gradle build. This is due to the configured
inclusion of '**/*Tests.*' which follows Spring's naming convention for
test classes.
This commit addresses this issue by:
- Renaming real test classes consistently to "*Tests".
- Renaming internal test classes to "*TestCase".
- Renaming @WebTest to @WebTestStereotype.
- Disabling broken tests in AnnoDrivenStaticEntityMockingControlTest.
- Modifying the Gradle build configuration so that classes ending in
either "*Tests" or "*Test" are considered test classes.
Issue: SPR-11384
SPR-11145 claims that ServletContextAware beans declared in an
ApplicationContext loaded for an integration test by the TestContext
framework (TCF) do not have their setServletContext() methods invoked
if the tests are executed manually using JUnit 4.11.
This commit verifies that such ServletContextAware beans are processed
properly regardless of how the test was launched. Specifically:
- A ServletContextAwareBean has been introduced.
- BasicAnnotationConfigWacTests has been retrofitted with a
ServletContextAwareBean in its context.
- ServletContextAwareBeanWacTests has been introduced to execute
BasicAnnotationConfigWacTests manually via JUnitCore.
Issue: SPR-11145
- Fix Javadoc in HibernateTemplate so that it compiles in Eclipse.
- Suppress generics warnings in HibernateTemplateTests.
- Remove unnecessary deprecation warning suppression in constructor for
LocalSessionFactoryBuilder.
Prior to this commit, RestTemplate's constructors were all initializing
default HTTPMessageConverters. Its API provides a way to replace
those converters with custom ones, but default converters are already
defined and initialized at that point, which can be an issue in some
cases (performance, classpath...).
This commits adds a new constructor for RestTemplate with a list
of message converters as argument. With this new constructor,
default message converters are never initialized.
Issue: SPR-11351
Note that this variant of HibernateTemplate is stripped down in terms of Session management options and just provides a current-session style with a fallback to a temporary session-per-operation. Furthermore, in the latter fallback mode, only read operations are supported, just like with a JPA EntityManager.
Issue: SPR-11291
Prior to this commit, parts of a multipart HTTP request could be
injected in @RequestPart and @RequestParam annotated arguments, when
using types:
* MultipartFile, Collection<MultiPartFile>
* javax.servlet.Part, Collection<Part>
This commits updates @RequestParam and @RequestPart argument resolvers
and now allows the array versions of those types:
* Part[]
* MultiPartFile[]
Note that the MockHtpServletRequest backing tests for standard
Servlets implementations now uses a MultiValueMap to store parts
(versus a simple hashmap).
Issue: SPR-11353
Stereotype annotations should support a 'value' attribute for
specifying the name of the Spring-managed component; however,
@RestController currently does not provide such an attribute.
This commit introduces a 'value' attribute in @RestController so that
developers can provide custom names for components annotated with
@RestController.
Issue: SPR-11360
JOpt 4.6 redeclared its nonOptionArguments() method from List<String> to List<?>, requiring us to select String arguments only as we do for regular option values already.
Issue: SPR-11359
According to the JDK's documentation and changelog, the Introspector itself safely handles weak references as of JDK 6 update 21 (which is what we require for Spring 4.0).
Issue: SPR-11356
Update reference documentation to make it clearer that only
`spring-core` has a direct dependency on `commons-logging`.
Also reference the 'empty jar' alternative option as described in
the SLF4J FAQ.
This change makes it possible to provide no configuration for the
DispatcherServlet when extending
AbstractAnnotationConfigDispatcherServletInitializer, and therefore
provide all config through the "root" context.
Issue: SPR-11357
This in particular allows for specifying "spring.getenv.ignore" and "spring.beaninfo.ignore" in a local way within the application, in case that JVM-level system properties are locked.
Issue: SPR-9014
Issue: SPR-11297
equalTo is not a valid method on JsonPathResultMatchers
I have signed and agree to the terms of the SpringSource Individual Contributor License Agreement.
Fixed through downcasting to AspectJ's ReflectionType and ReflectionBasedReferenceTypeDelegate, obtaining the myClass field there. We only fall back to regular class loading if we encounter any other kind of type.
Issue: SPR-11344
(cherry picked from commit c406c56)
Prior to this commit, a NoClassDefFoundError caught in
TestContextManager's retrieveTestExecutionListeners() method would be
handled differently for implicit default listeners (i.e., listeners not
declared via @TestExecutionListeners) and listeners explicitly declared
via @TestExecutionListeners. Specifically, a NoClassDefFoundError would
cause a test to fail for an explicitly declared TestExecutionListener
but not for an implicitly declared one.
This commit addresses this issue by:
- Always swallowing a NoClassDefFoundError for both implicitly and
explicitly declared TestExecutionListeners.
- Changing the log level from DEBUG to INFO to make such situations
more visible to the average end user.
Issue: SPR-11347
Assertions made in callbacks invoked on separate thread were not
reaching the test framework. This fix ensures failures are detected
and cause tests to fail.
This commit deletes methods in RollbackForRequiredEjbTxDaoTestNGTests
that were unnecessarily redeclared. The redeclaration is required for
JUnit's @FixMethodOrder support but not for TestNG's built-in support
for dependent methods.
Issue: SPR-6132
Prior to this commit, all MethodMessageHandlers tests were
implemented in a single class. Since SimpAnnotationMsgHandler
has been refactored with an abstract class, tests also
needed such a refactoring.
This commit creates test fixtures for AbstractMethodMessageHandler.
Issue: SPR-11191
Prior to this commit, it was not clear how to enable the support of matrix
variables in the mvc namespace. As the feature is disabled by default, added
something to highlight the part that explains how to configure it
Issue: SPR-11331
This allows us to avoid early initialization of footprint-heavy ConcurrentHashMaps, and reuses the existing postProcessingLock which will typically be active already when registerExternallyManaged* calls come in from the bean factory's post-processing phase.
Issue: SPR-11343
This commit introduces transactional integration tests executing
against both JUnit and TestNG in the TestContext framework (TCF) using
@TransactionAttribute in EJBs instead of Spring’s @Transactional
annotation.
These tests disprove the claims raised in SPR-6132 by demonstrating that
transaction support in the TCF works as expected when a transactional
EJB method that is configured with TransactionAttribute.REQUIRES_NEW is
invoked. Specifically:
- The transaction managed by the TCF is suspended while such an EJB
method is invoked.
- Any work performed within the new transaction for the EJB method is
committed after the method invocation completes.
- The transaction managed by the TCF is resumed and subsequently
either rolled back or committed as necessary based on the
configuration of @Rollback and @TransactionConfiguration.
The configuration for the JUnit-based tests is straightforward and self
explanatory; however, the configuration for the TestNG tests is less
intuitive.
In order for the TCF to function properly, the developer must ensure
that test methods within a given TestNG test (whether defined locally,
in a superclass, or somewhere else in the suite) are executed in the
proper order. In a stand-alone test class this is straightforward;
however, in a test class hierarchy (or test suite) with dependent
methods, it is necessary to configure TestNG so that all methods within
an individual test are executed in isolation from test methods in other
tests. This can be achieved by configuring a test class to run in its
own uniquely identified suite (e.g., by annotating each concrete
TestNG-based test class with @Test(suiteName = "< Some Unique Suite
Name >")).
For example, without specifying a unique suite name for the TestNG
tests introduced in this commit, test methods will be executed in the
following (incorrect) order:
- CommitForRequiredEjbTxDaoTestNGTests.test1InitialState()
- CommitForRequiresNewEjbTxDaoTestNGTests.test1InitialState()
- RollbackForRequiresNewEjbTxDaoTestNGTests.test1InitialState()
- RollbackForRequiredEjbTxDaoTestNGTests.test1InitialState()
- CommitForRequiredEjbTxDaoTestNGTests.test2IncrementCount1()
The reason for this ordering is that test2IncrementCount1() depends on
test1InitialState(); however, the intention of the developer is that
the tests for an individual test class are independent of those in
other test classes. So by specifying unique suite names for each test
class, the following (correct) ordering is achieved:
- RollbackForRequiresNewEjbTxDaoTestNGTests.test1InitialState()
- RollbackForRequiresNewEjbTxDaoTestNGTests.test2IncrementCount1()
- RollbackForRequiresNewEjbTxDaoTestNGTests.test3IncrementCount2()
- CommitForRequiredEjbTxDaoTestNGTests.test1InitialState()
- CommitForRequiredEjbTxDaoTestNGTests.test2IncrementCount1()
- CommitForRequiredEjbTxDaoTestNGTests.test3IncrementCount2()
- RollbackForRequiredEjbTxDaoTestNGTests.test1InitialState()
- RollbackForRequiredEjbTxDaoTestNGTests.test2IncrementCount1()
- RollbackForRequiredEjbTxDaoTestNGTests.test3IncrementCount2()
- CommitForRequiresNewEjbTxDaoTestNGTests.test1InitialState()
- CommitForRequiresNewEjbTxDaoTestNGTests.test2IncrementCount1()
- CommitForRequiresNewEjbTxDaoTestNGTests.test3IncrementCount2()
See the JIRA issue for more detailed log output.
Furthermore, @DirtiesContext(classMode = ClassMode.AFTER_CLASS) has
been used in both the JUnit and TestNG tests introduced in this commit
in order to ensure that the in-memory database is reinitialized between
each test class.
Issue: SPR-6132
Prior to this commit, several spring-messaging exceptions
had defined serialVersionUIDs. Those exception aren't
supposed to leave the system via Java serialization; also,
their deserialization is supported only against the same
version of Spring.
Issue: SPR-11339
Update ResolvableType to unwrap Serialization wrapped TypeVariables
before calling the equals method.
This protects against the recent change in OpenJDK 8 (build 124)
which changed the TypeVariableImpl equals method such that it only
matches against other TypeVariableImpl instances.
Issue: SPR-11342
Change SerializableTypeWrapper proxies to directly call equals() methods
on the underlying Type, rather than possibly generating more wrappers.
This should help to improve performance, especially as the equals()
method is called many times when the ResolvableType cache is checked.
Issue: SPR-11335
The interface is to be implemented in addition to
java.security.Principal when Principal.getName() is not globally unique
enough for use in user destinations.
Issue: SPR-11327
Note: A default orm.xml file is only being used when not co-located with a persistence.xml file, since we assume it is intended for use with that persistence.xml file only then.
Issue: SPR-11234
Issue: SPR-11260
The ServletTestExecutionListener is now prepended to the set of default
listeners in AbstractJUnit4SpringContextTests and
AbstractTestNGSpringContextTests.
Issue: SPR-11340
Prior to this commit TestNG tests would only be executed by the Gradle
build if they were located in the “testng” package. Tests in subpackages
would therefore be omitted from the build.
This commit ensures that all TestNG classes in the “testng” package and
any of its subpackages are executed in the Gradle build.
Furthermore, this commit ensures that the JUnit-based
FailingBeforeAndAfterMethodsTests test class is executed along with the
other JUnit tests even though it resides under the “testng” package.
Issue: SPR-11338
Prior to this commit, it was necessary to override
the HandlerMapping definition to change properties
like useSuffixPatternMatch, useSuffixPatternMatch...
Also, one couldn't set custom pathmatcher/pathhelper
on RequestMappingHandlerMapping via XML configuration.
This commits adds a new "mvc:annotation-driven"
subelement called "mvc:path-matching" for the tag
that allows to configure such properties:
* suffix-pattern
* trailing-slash
* registered-suffixes-only
* path-matcher
* path-helper
Note: this is a new take on this issue, since
96b418cc has been reverted by e2b99c3.
Issue: SPR-10163
This reverts commit 96b418cc8a,
"Make RequestMappingHandlerMapping xml config easier".
This implementation makes the mvc:annotation namespace less readable,
and future configuration items would add even more to this namespace.
Issue: SPR-10163
Allow HttpHeader instances to be returned directly from MVC controller
methods managed by HandlerMethodReturnValueHandler rather than needing
to be wrapped in a ResponseEntity.
Issue: SPR-11129
JXL had no release since 2009, and has some serious bugs remaining: e.g. JXL on Unix cannot read template files created on Windows, using the 'latest' JXL 2.6.12 release.
The @Header annotation in spring-messaging now resolves values from the
nested "nativeHeaders" map as well as top-level header map values.
In case of ambiguity (a value that exists in both), the top-level map
value is used and a warning message is printed. This is unlikly in most
cases but can be resolved by prefixing the header value with
"nativeHeadres.myHeader".
Issue: SPR-11326
DefaultUserDestinationResolver now uses the session id of
SUBSCRIBE/UNSUBSCRIBE messages rather than looking up all session id's
associated with a user.
Issue: SPR-11325
Provided overloaded versions of `forField` and `forMethodParameter` that
accept a `ResolvableType` implementation type (as opposed to a Class).
Primarily added to allow resolution against implementation types that
have been created programmatically using `forTypeWithGenerics`.
Issue: SPR-11218
This commit makes the logging in TransactionalTestExecutionListener
consistent for both starting and ending transactions. Specifically,
the current TestContext is now included in the informational log
statement when starting a new transaction.
Issue: SPR-11323
Prior to this commit, it was necessary to override
the HandlerMapping definition to change properties
like useSuffixPatternMatch, useSuffixPatternMatch...
This commits adds new attributes on the
mvc:annotation-driven XML tag that allows to configure
such flags:
* use-suffix-pattern-match
* use-trailing-slash-match
* use-registered-suffix-pattern-match
Issue: SPR-10163
sockjs-client expects a prelude to be written on every request with
streaming transports. The protocol tests don't make this clear and
don't expose this issue.
The test case for SPR-11183 (writing 20K messages in succession) did
expose the issue and this commit addresses it.
Issue: SPR-11183
When adding headers generically, MockHttpServletRequestBuilder now
recognizes Content-Type and updates the contentType field accordingly.
Issue: SPR-11308
This change adds a protected method to DefaultHandshakeHandler to
determine the user for the WebSocket session. By default it's
implemeted to obtain it from the request.
Issue: SPR-11228
The Jetty ServletWebUpgradeRequest implements getUserPrincipal to
return the Principal from the HttpServletRequest on the upgrade.
This change ensures we can fall back on that.
However the JettyRequestUpgradeStrategy still passes the user from
HttpServletRequest from the upgrade, in order to work with Jetty
9.0.x and avoid running into this 9.1.x issue:
https://bugs.eclipse.org/bugs/show_bug.cgi?id=423118
While we've had basic Quartz 2.2 support before, a few details were missing:
* LocalDataSourceJobStore's ConnectionProvider adapters need to provide an empty implementation of Quartz 2.2's new initialize method.
* SchedulerFactoryBean's "schedulerContextMap" needs to be explicitly declared with String keys, otherwise it can't be compiled against Quartz 2.2 (forward compatibility once we're dropping Quartz 1.x support). This doesn't hurt against older Quartz versions either, since the keys need to be Strings anyway.
Issue: SPR-11284
Before this change DefaultUserDestinationResolver provided a separate
destination prefix property for identifying "user" destinations in
subscription requests as opposed to in sent messages. Such a separate
property should not be needed.
Issue: SPR-11263
Configurers and BeanDefinitionParsers should use
OptionalValidatorFactoryBean instead of
LocalValidatorFactoryBean.
The Optional implementation catches and logs setup
exceptions, useful when a validation API is present on
the classpath but not the actual implementation.
Issue: SPR-11272
Payload parameters in @MessageMapping annotated
methods can now also be validated when annotated
with a Validation annotation (@Valid, @Validated...).
A default Validator is registered by the MessageBroker
Configurer, but it is possible to provide a list of custom
validators as well.
Issue: SPR-11185
Prior to this commit, some unit tests were using
Spring's Jackson 1.x implementations. Now Jackson
2.x implementations are the default ones used in
unit tests.
Even if Jackson 1.x support is deprecated, Jackson 1.x
unit tests are kept.
Issue: SPR-11121
Since the MessageHeaders constructor is makes a copy of the headers and
is protected against a null map, there is no need for the same to be
done in GenericMessage.
Issue: SPR-11268
Also simplified cache key 'hashCode' implementation, relying on 'equals' to differentiate between same method on different target classes.
Issue: SPR-11267
Also introduced a default ZonedDateTime-Calendar converter which is not covered by the default convention due to the 'from' method only being defined on GregorianCalendar.
Issue: SPR-11259
This change splits out an abstract base class from DefaultMockMvcBuilder
with StandaloneMockMvcBuilder switching to extend the new abstract class
(rather than DefaultMockMvcBuilder).
Issue: SPR-11238
This turned out to be a bug in the ASM-based AnnotationMetadata implementation where has/getAnnotatedMethods didn't consider meta-annotations., in contrast to its StandardAnnotationMetadata sibling.
Issue: SPR-10488
Update the MVC reference documentation with the new concepts and classes
introduced in Spring 4.0
This reverts commit 6f874d734a2a38431c74444894f54797ecf74933.
As a benefit over the "shared" flag, this allows for sharing a specific CacheManager in a system with potentially multiple CacheManagers involved, and it supports controlled shutdown by the EhCacheManagerFactoryBean that actually created the CacheManager.
Issue: SPR-11178
- Completed Javadoc for MetaAnnotationUtils.
- Added Javadoc notes to multiple annotations in the TCF, pointing out
which annotations can be used as meta-annotations.
Issue: SPR-11109
Fix a potential race condition with GenericMessagingTemplate's inner
TemporaryReplyChannel class.
Prior to this commit the `hasReceived` member variable was read after
calling `replyLatch.countDown()`.
Issue: SPR-11206
This commit fixes a bug introduced in the last commit.
ServletTestExecutionListener (STEL) now tracks whether it has already
populated the RequestContextHolder.
Issue: SPR-11144
The previous commit for issue SPR-11144 revealed a bug in
ServletTestExecutionListener (STEL). Specifically, STEL acted on the
fact that the ApplicationContext for a given TestContext was an
instance of WebApplicationContext. This behavior could potentially
break test code from previous releases of the Spring Framework that
relied on a custom setup of the RequestAttributes in the
RequestContextHolder with a custom WebApplicationContext ContextLoader.
This commit addresses this issue by ensuring that STEL only comes into
play if the test class is annotated with @WebAppConfiguration (for
prepareTestInstance() and beforeTestMethod()) or if the TestContext
attribute named RESET_REQUEST_CONTEXT_HOLDER_ATTRIBUTE is set to
Boolean.TRUE (for afterTestMethod()).
Issue: SPR-11144
Specifically, read and write methods are allowed to express property types with superclass/subclass relationships in both directions now.
Issue: SPR-11139
Specifically, if the current factory has Qualifier/ContextAnnotationAutowireCandidateResolver set up, it is important to pass it on to ancestor factories to get consistent qualifier matching results.
Issue: SPR-10966
XStream 1.4.5 doesn't check a given ConverterLookup for a ConverterRegistry implementation anymore, so we have to manually perform that check and assignment.
Issue: SPR-11147
Our per-bean caching in AutowiredAnnotationBeanPostProcessor and co relies on unique bean names, so this change fixes potential cache mismatch problems occuring there.
Issue: SPR-11131
Prior to this commit, the ServletTestExecutionListener did not
overwrite RequestAttributes in the RequestContextHolder if the
ApplicationContext associated with the given TestContext was not a
WebApplicationContext; however, the ServletTestExecutionListener would
clear the RequestAttributes after every test method execution,
regardless of whether the context was a WebApplicationContext or not.
This behavior breaks backwards compatibility with integration tests
that managed the RequestAttributes in RequestContextHolder themselves.
This commit addresses this issue by introducing a TestContext attribute
named RESET_REQUEST_CONTEXT_HOLDER_ATTRIBUTE in
ServletTestExecutionListener. This attribute is used internally within
ServletTestExecutionListener to ensure that the RequestContextHolder is
only cleared (i.e., reset) if the ServletTestExecutionListener actually
populated the RequestContextHolder.
Issue: SPR-11144
The Spring Framework reference now uses http://asciidoctor.org/[asciidoctor]. This document describes how to contribute documentation updates.
The Spring Framework reference now uses http://asciidoctor.org/[asciidoctor]. This
document describes how to contribute documentation updates.
== Building with Gradle
You can build the documentation using gradle using the `asciidoc` task. For example, from the project root execute the following command:
You can build the documentation using gradle using the `reference` task. For example, from
the project root execute the following command:
gradlew asciidoc
./gradlew reference
the output will be available at `spring-framework/build/asciidoc/index.html`
the output will be available at `spring-framework/build/reference/htmlsingle/index.html`
== Live editing
One of the nice features about using asciidoctor is the support for live editing.
You will find a Guardfile already present at `spring-framework/src/asciidoc/Guardfile`. Then ensure to follow the setup instructions within the http://asciidoctor.org/docs/editing-asciidoc-with-live-preview/[Editing AsciiDoc with Live Preview] document.
You will find a Guardfile already present at `spring-framework/src/asciidoc/Guardfile`.
Make sure first to follow the setup instructions within the
http://asciidoctor.org/docs/editing-asciidoc-with-live-preview/[Editing AsciiDoc with Live Preview]
document. Once you have done that, there are additional gems to install to make it work
(assuming that you are using http://livereload.com/[LiveReload]):
When running `guard start` within the `src/asciidoc/` folder, any changes to the `src/asciidoc/index.adoc` file will automatically be written at `src/asciidoc/build/index.html`.
gem install guard-rspec guard-livereload
When running `guard start` within the `src/asciidoc/` folder, any changes to the
`src/asciidoc/index.adoc` file will automatically be written at
`src/asciidoc/build/index.html`.
== Troubleshooting
* If you are using LiveReload, make sure to select _Allow access to file URLs_ in the
LiveEdit plugin options of your browser.
* The icon used to enable _LiveReload_ can be a bit confusing. The dot is empty when it is
disabled and full when the plugin is active. Make sure to enable it on the tab
displaying the `index.html` file.
* Ensure you are _not_ running guard start at all as two instances could not run at the
same time. To exit a current session in a clean way, just type e in the shell.
== Documentation notes
Some notes on documentation
* It is important to keep whitespaces to a minimum to make it simple to edit files. This means use an editor with line wrapping rather than manually inserting hard returns.
* Documentation is wrapped at 90 chars, ensure that you manually wrap your edits
* Tabs are used for indentation, do not use spaces
* Follow the existing style when inserting `source` blocks
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Sam Brannen