Introduced ContentNeogtiationStrategy for resolving the requested
media types from an incoming request. The available implementations
are based on path extension, request parameter, 'Accept' header,
and a fixed default content type. The logic for these implementations
is based on equivalent options, previously available only in the
ContentNegotiatingViewResolver.
Also in this commit is ContentNegotiationManager, the central class to
use when configuring content negotiation options. It accepts one or
more ContentNeogtiationStrategy instances and delegates to them.
The ContentNeogiationManager can now be used to configure the
following classes:
- RequestMappingHandlerMappingm
- RequestMappingHandlerAdapter
- ExceptionHandlerExceptionResolver
- ContentNegotiatingViewResolver
Issue: SPR-8410, SPR-8417, SPR-8418,SPR-8416, SPR-8419,SPR-7722
PropertySourcesPropertyResolver#containsProperty now
calls #containsProperty on each underlying PropertySource instead of
calling #getProperty and checking for null.
Issue: SPR-9529
The reference manual currently documents the wrong file name for the
default data SQL script used by EmbeddedDatabaseBuilder. In addition,
the testing chapter of the reference manual does not link to the testing
section of the JDBC chapter.
- Updated the "Testing data access logic with an embedded database"
section of the reference manual appropriately.
- Added a new paragraph to the "JDBC Testing Support" section of the
testing chapter which cross references the "Testing data access logic
with an embedded database" section.
Issue: SPR-9467
Commit aa415d7c0c introduced
JacksonObjectMapperFactoryBean and associated tests, but with Windows
(CRLF) line endings instead of the conventional Unix (LF) line endings.
This commit converts these files to LF endings using the handy
`dos2unix` utility.
Issue: SPR-9125
Currently, if a factory method is parameterized and the corresponding
variable types are declared on the method itself instead of on the
enclosing class or interface, Spring always predicts the return type to
be Object, even if the return type can be explicitly inferred from the
method signature and supplied arguments (which are available in the bean
definition).
This commit introduces a new resolveParameterizedReturnType() method in
GenericTypeResolver that attempts to infer the concrete type for the
generic return type of a given parameterized method, falling back to the
standard return type if necessary. Furthermore,
AbstractAutowireCapableBeanFactory now delegates to
resolveParameterizedReturnType() when predicting the return type for
factory methods.
resolveParameterizedReturnType() is capable of inferring the concrete
type for return type T for method signatures similar to the following.
Such methods may potentially be static. Also, the formal argument list
for such methods is not limited to a single argument.
- public <T> T foo(Class<T> clazz)
- public <T> T foo(Object obj, Class<T> clazz)
- public <V, T> T foo(V obj, Class<T> clazz)
- public <T> T foo(T obj)
Issue: SPR-9493
The readme is heavy on links and hard to read in source form when using
typical inline Markdown links. The following changes have been made to
increase readability.
- Wrap lines at 80 characters
- Use 'implicit link name' shortcut syntax as described at [1]
- Add link to 'spring-mvc' Stack Overflow tag as well
- Add link to new 'lifecycle of an issue' wiki doc
[1]: http://daringfireball.net/projects/markdown/syntax
Some Spring Framework users and teams cannot use transitive dependency
management tools like Maven, Gradle and Ivy. For them, a `distZip` task
has been added to allow for creating a Spring Framework distribution
from source that includes all optional and required runtime
dependencies for all modules of the framework.
This 'dist-with-deps' zip is not published to the SpringSource
repository nor to the SpringSource community download page. It is
strictly an optional task introduced as a convenience to the users
mentioned above.
Detailed instructions for building this zip locally have been added to
the wiki and the README has been updated with a link to that new doc.
Issue: SPR-6575
The BeanFactory makes it easier to configure a customized Jackson
ObjectMapper (for example enable/disable certain features).
This bean factory is usually used with
MappingJacksonHttpMessageConverter or MappingJacksonJsonView.
See JavaDoc for examples.
Issue: SPR-9125
Previously MediaType could only parse double-quoted parameters without
raising an IllegalArgumentException. Now parameters can also be
single-quoted.
Issue: SPR-8917
HttpStatus cannot be created with an unknown status code. If a server
returns a status code that's not in the HttpStatus enum values, an
IllegalArgumentException is raised. Rather than allowing it to
propagate as such, this change ensures the actual exception raised is
a RestClientException.
Issue: SPR-9406
A recent commit made aopalliance optional for spring-aop, while
continuing to require it for spring-tx. On review, this is probably
overly aggressive, and for convenience aopalliance should remain
required for spring-aop. There are use cases for which aopalliance is
indeed optional, but core functionality such as <aop:scoped-proxy>
should never result in a ClassNotFoundException.
This commit returns the aopalliance dependency for spring-aop to
required status, and also explicitly notes the same dependency in other
modules that compile directly against aopalliance types.
Issue: SPR-9501
The junit:junit Maven artifact includes a bundled version of hamcrest
core. For projects that depend on later versions of hamcrest this causes
significant issues in terms of dependency management.
The spring-test module now depends on junit:junit-dep, thus allowing
developers to better manage their test dependencies on a more fine
grained level.
Also tidied up dependency issues regarding hamcrest-core and
hamcrest-all across the build.
Issue: SPR-6966
This commit introduces a test suite (Spr8849Tests) that demonstrates
the claims made in SPR-8849.
Specifically, if <jdbc:embedded-database id="xyz" /> is used to create
an embedded HSQL database in an XML configuration file and that
configuration file is imported in different sets of configuration files
that are used to load ApplicationContexts for different integration
tests, the embedded database will be initialized multiple times using
any nested <jdbc:script /> elements. If such a script is used to create
a table, for example, subsequent attempts to initialize the database
named "xyz" will fail since an embedded database named "xyz" already
exists in the JVM.
As a work-around, this test suite uses a SpEL expression to generate a
random string for each embedded database instance:
id="#{T(java.util.UUID).randomUUID().toString()}"
See the Javadoc in Spr8849Tests for further information.
Issue: SPR-8849
Previously, DatabasePopulatorUtils#execute looked up a Connection from
the given DataSource directly which resulted in the executed statements
not being executed against a transactional connection (if any) which in
turn resulted in the statements executed by the populator potentially
not being rolled back.
Now DataSourceUtils#getConnection is used to transparently take part in
any active transaction and #releaseConnection is used to ensure the
connection is closed if appropriate.
Issue: SPR-9457
Default HTTP error exceptions thrown from RestTemplate now include
response headers in addition to the response body. In particular, this
enables inspection of the Content-Type header allowing manual
deserialization of the response body without guessing as to the content
type.
- introduce HttpStatusCodeException#getResponseHeaders
- add constructor with headers param for HttpStatusCodeException,
HttpClientErrorException and HttpServerErrorException
- preserve exsisting constructor signatures
- mark HttpHeaders as Serializable
- generate new serialVersionUID where needed
Issue: SPR-7938
The following style updates have been made in anticipation of
substantive changes in subsequent commits:
- organize imports
- correct whitespace errors (leading spaces in code, tabs in Javadoc)
- wrap Javadoc at 90 chars; make imperative ("Return" vs. "Returns")
- use conventional constructor argument wrapping
A serialVersionUID has also been added to RestClientException and its
ResourceAccessException subclass for consistency with the rest of that
same exception hierarchy.
Issue: SPR-7938
- Explicitly specify compile-time dependencies on other spring-*
modules, primarily for accuracy in pom generation and ensuring
minimal dependencies for users of spring-aspects.
- Remove use of p: namespace from annotation-cache-aspectj.xml to
avoid parser-related test failures under Eclipse (likely due to
classpath differences between Gradle and Eclipse).
The spring-intstrument jar should have a 'Premain-Class:' manifest
entry in order to enable use as a Java agent. This entry was present in
versions 3.1.1 and earlier, but due to build infrastructure changes
starting in 3.2.x this entry was missed. It is now back in place as
expected.
Issue: SPR-9458
Gradle-generated poms thoroughly tested against 3.1.1 versions, with an
eye toward making as many spring-* dependencies optional as possible.
All spring-* modules now declare a Gradle dependency on any other
spring-* module where there is a direct compile-time usage of the
sources in that module. Previously, dependency declarations were
minimal, letting transitive resolution do most of the work. However,
this creates less than ideal poms and is generally not very
informative.
So for example, spring-jdbc uses spring-core, spring-beans and
spring-tx classes directly. Therefore it has the following declarations:
compile project(":spring-core")
compile project(":spring-beans")
compile project(":spring-tx")
spring-core depends on spring-asm, but spring-jdbc does not use
spring-asm classes directly. Therefore spring-jdbc does not declare a
dependency on spring-asm. Transitive resolution is fine in such a case.
As for optional dependencies, it is a matter of degrees what
constitutes optional. A rule of thumb is whether there are legitimate
and likely use cases in which the module can be used without needing
the dependency. spring-jdbc has only one compile-time dependency on
spring-context classes, and that's in JndiDataSourceLookup. It is
certainly reasonable to imagine using spring-jdbc without JNDI,
therefore the spring-context dependency is declared optional as
follows:
compile(project(":spring-context"), optional) // for JndiDataSourceLookup
In Spring 3.1 the spring-test Maven artifact did not have a required
dependency on spring-core, but there is practically no part of
spring-test that can be used without spring-core. Most test utilities
that are intended to be stand-alone utilities in fact use utility
classes from spring-core (e.g., ReflectionTestUtils). Even some of the
web mocks/stubs use spring-core (e.g., DelegatingServletInputStream).
In addition, the current Gradle build configuration for the spring-test
module is very simplistic -- in that it does not explicitly list any
optional dependencies such as the Servlet and Portlet APIs -- and it
defines a 'compile' dependency on spring-webmvc-portlet.
The resulting Maven dependencies in the generated POM are therefore not
what a typical consumer of the spring-test artifact would reasonably
expect.
To address these issues, the Gradle build configuration for the
spring-test module now explicitly defines the following 'compile'
dependencies:
- spring-core
- spring-webmvc, optional
- spring-webmvc-portlet, optional
- junit, optional
- testng, optional
- servlet-api, optional
- jsp-api, optional
- portlet-api, optional
- activation, provided
The only required dependency is now spring-core; all other dependencies
are 'optional'.
Issue: SPR-8861
Prior to this change, by-type lookups using DLBF#getBeanNamesForType
required traversal of all bean definitions within the bean factory
in order to inspect their bean class for assignability to the target
type. These operations are comparatively expensive and when there are a
large number of beans registered within the container coupled with a
large number of by-type lookups at runtime, the performance impact can
be severe. The test introduced here demonstrates such a scenario clearly.
This performance problem is likely to manifest in large Spring-based
applications using non-singleton beans, particularly request-scoped
beans that may be created and wired many thousands of times per second.
This commit introduces a simple ConcurrentHashMap-based caching strategy
for by-type lookups; container-wide assignability checks happen only
once on the first by-type lookup and are afterwards cached by type
with the values in the map being an array of all bean names assignable
to that type. This means that at runtime when creating and autowiring
non-singleton beans, the cost of by-type lookups is reduced to that of
ConcurrentHashMap#get.
Issue: SPR-6870
Rossen Stoyanchev
Chris Beams
Sam Brannen
Dmitry Katsubo
David J. Biesack
Maxim Valyanskiy
Oliver Gierke
Scott Andrews
Spring Buildmaster
Juergen Hoeller