Move `ApplicationHome`, `ApplicationPid` and `ApplicationTemp` to the
`system` package. Since `system` package is now much lower level, the
existing `FileWriter` implementations also needed to move to prevent
package tangles.
Fixes gh-8614
Add a new `WebServerApplicationContext` interface that provides a common
abstraction for all application contexts that create and manage the
lifecycle of an embedded `WebServer`.
Allows server namespaces to become a first-class concept (rather
subverting `ConfigurableWebApplicationContext.getNamespace()`) and
allow us to drop `getServerId()` from `WebServerInitializedEvent`.
Also helps to improve `ManagementContextAutoConfiguration` and
`ManagementContextFactory`.
Fixes gh-11881
By default, AbstractTestExecutionListeners have an order of lowest
precedence. This means that it is impossible to write a listener with
lower precedence that any listener that's using the default order.
This commit updates Boot's 6 AbstractTestExecutionListeners to order
them explicitly. MockitoTestExecutionListener performs injection of
Mockito mocks and spies into the test instance. It now has an order of
2050 giving it slightly lower precedence than the dependency injection
test execution listener (2000).
The remaining 5 listeners have all been ordered with lowest precedence
- 100. This leaves them near their current lowest precedence position
while creating some room for any listeners that require lower
precedence.
Closes gh-11796
Previously, the order of the entries in a TestJarFile was determined
by the underlying file system rather than by the order in which
they were added. This could lead to unpredicatable ordering and
failures in tests that verify archive entry ordering.
This commit updates TestJarFile to add entries to the archive in
insertion order.
See gh-11695
See gh-11696
Previously, the Repackager would write entries in the following
order:
- Libraries that require unpacking
- Existing entries
- Application classes
- WEB-INF/lib jars in a war
- Libraries that do not require unpacking
- Loader classes
Libraries that require unpacking were written before existing entries
so that, when repackaging a war, an entry in WEB-INF/lib would not
get in first and prevent a library with same location from being
unpacked. However, this had the unwanted side-effect of changing
the classpath order when an entry requires unpacking.
This commit reworks the handling of existing entries and libraries
that require unpacking so that existing entries can be written first
while also marking any that match a library that requires unpacking
as requiring unpacking.
Additionally, loader classes are now written first. They are the
first classes in the jar that will be used so it seems to make sense
for them to appear first. This aligns Maven-based repackaging
with the Gradle plugin's behaviour and with the structure documented
in the reference documentation's "The Executable Jar Format" appendix.
The net result of the changes described above is that entries are
now written in the following order:
- Loader classes
- Existing entries
- Application classes
- WEB-INF/lib jars in a war marked for unpacking if needed
- Libraries
Closes gh-11695
Closes gh-11696
Previously, the ordering of the entries in an archive produced by
BootJar was different to the ordering of the entries in an archive
produced by BootWar. The latter placed application classes before
any nested jars, whereas the former was the other way around.
This commit updates BootJar to use the same ordering as BootWar and
adds tests to verify that the ordering is the following:
1. Loader classes
2. Application classes (BOOT-INF/classes or WEB-INF/classes)
3. Nested jars (BOOT-INF/lib or WEB-INF/lib)
4. Provided nested jars in a war (WEB-INF/lib-provided)
The tests also verify that the position of a library is not affected
by it requiring unpacking.
See gh-11695
See gh-11696
Stop running apply-plugin tests as part of the build since during a
release the version number will change and the jar will not be
available.
Fixes gh-11857
Update a couple of the `spring-boot-gradle-plugin` sample gradle flies
so that they include the running classpath. The additional lines are
contained within a tag which is ultimately filtered from the final
documentation.
Fixes gh-11857
While Spring Mobile support has been removed from Spring Boot, the
auto-configuration has been relocated to a separate module that uses
the same keys.
Flagging those keys as deprecated means that the IDE will be confused
when the extra jar is present on the classpath as it advertizes, as
it should, support fo them.
Closes gh-11844
Add MeterFilter to restrict the maximum number of web client URI tags
created. Prior to this commit, if a user was manually building URIs for
use with a RestTemplate (rather than using uriVariables) the JVM could
run out of memory.
Fixes gh-11338
Co-authored-by: Phillip Webb <pwebb@pivotal.io>
Update micrometer auto-configuration so that a `CompositeMeterRegistry`
is only created when more than one `MeterRegistry` bean is declared.
When a composite is crated, it is marked as `@Primary` so that it
can be directly injected. Meter registries can now be defined directly
as beans, and auto-configuration can back off in the usual way.
The `MeterRegistryConfigurer` is now called `MeterRegistryCustomizer`
and is generically types so it's easy to apply customizations to a
particular `MeterRegistry` implementation.
Fixes gh-11799
Co-authored-by: Jon Schneider <jschneider@pivotal.io>
By default, JUL configures a single root handler. That handler is a
ConsoleHandler. Previously, we removed all root handlers from JUL but
this is problematic in environments where other handlers are
registered with JUL and those handlers need to be retained.
0679d436 attempted to fix the problem by leaving the root handlers in
place and only adding and removing the bridge handler. This resulted
in log output from Tomcat (and anything else that uses JUL) being
duplicated.
This commit makes another attempt at tackling the problem. It attempts
to detect JUL's default configuration (a single root handler that's a
ConsoleHandler) and only removes the handler if it appears to be from
the default configuration. For environments where default JUL
configuration is being used, this will prevent duplicate logging and
for environments where custom JUL configuration is being used, this
will prevent that configuration from being undone.
Closes gh-8933
Previously, the test in MetricsAutoConfigurationIntegrationTests was
testing the functionality of WebMvcMetricsFilter to verify that the
auto-configuration had registered the filter for async dispatches.
This test was complex and covered the same code as a test in
WebMvcMetricsFilterTests.
This commit reworks the test to examine the dispatcher types on the
filter registration directly instead.
Closes gh-11826
This commit makes sure that a "cache.time-to-live" property is not
generated for endpoints that do not have a main read operation (i.e. a
read operation with no parameter or only nullable parameters).
This matches the endpoint feature that provides caching for only such
operation.
Closes gh-11703
This commit enables a more flexible Liquibase/Flyway configuration by
allowing for a combination of the provider's and the primary
DataSource's configuration to be used. This gives developers the
flexibility to specify only a user or a url and having
Liquibase/Flyway fall back to individual datasource properties rather
than ignoring the Liquibase/Flyway properties and falling back to the
default data source.
See gh-11751
Add an operation on PropertyMapper that takes care of casting. Returns
a source for the requested type if the current value is of the right
type.
Closes gh-11788
There's an extra ClassLoader in the hierarchy of the TCCL in Groovy
2.4 vs Groovy 2.5 so we require an extra getParent() call to avoid
being able to load classes that are visible to the launched URL
class loader.
Closes gh-11745
Previously, the logging system was cleaned up in response to the
root context's ContextClosedEvent being received. This event is
published early in a context's close processing. As a result, the
logging system is in cleaned up state while, for example, disposable
beans are being destroyed.
This commit reworks the logic that triggers logging system clean up
to use a disposable bean instead. Disposable beans are called in
reverse-registration order. The logging clean up bean is registered as
early as possible so that it should be the last disposable bean to
be called.
Closes gh-11676
Update `@SpringBootTest` `WebTestClient` support so that the bean
definition is only registered when the user has not defined or
auto-configured their own.
See gh-10556
Update `@SpringBootTest` `TestRestTemplate` support so that the bean
definition is only registered when the user has not defined or
auto-configured their own.
See gh-10556
Move the "testdb" naming logic to `DataSourceProperties` and expose
the `deduceDatabaseName` method so they can be used in
auto-configuration.
See gh-11719
Previously, Hikari's pool name was auto-configured with the value of
`spring.datasource.name` that defaults to `testdb`, which brings some
confusion.
This commit removes the default `testdb` value on
`spring.datasource.name` as it is a sane default only for an embedded
datasource. It is applied whenever applicable instead.
Closes gh-11719
On CI, Cassandra running inside the Docker container sometimes fails
to start or the start times out. This has nothing to do with Boot so
we attempt to protect our tests from the flakiness of the container
by allowing 3 startup attempts.
This commit adds InvalidEndpointRequestException as a technology
agnostic way to signal that an endpoint request is invalid. When such
exception is thrown, the web layer translates that to a 400.
Rather than overriding the reason, this commit makes sure to reuse the
error infrastructure.
Closes gh-10618
Rename `spring-boot-configuration-analyzer` to
`spring-boot-deprecated-properties-support`.
Also renamed classes to match and polished some of the code.
See gh-11301
This commit makes sure that `CharacterEncodingFilter` is ordered with
the `Ordered.HIGHEST_PRECEDENCE` and that other filters, potentially
reading the request body, are ordered after.
In this particular case, both `WebMvcMetricsFilter` and
`ErrorPageFilter` are now ordered at `Ordered.HIGHEST_PRECEDENCE + 1` to
avoid cases where the request body is read before the encoding
configuration is taken into account.
Closes gh-11607
This commit adds a new `spring-boot-configuration-analyzer` module that
can be added to any app to analyze its environment on startup.
Each configuration key that has a matching replacement is temporarily
transitioned to the new name with a `WARN` report that lists all of
them.
If the project defines configuration keys that don't have a replacement,
an `ERROR` report lists them with more information if it is available.
Closes gh-11301
As explained in the "Split package compatibility" section of
https://blog.jetbrains.com/kotlin/2017/09/kotlin-1-2-beta-is-out/
kotlin-stdlib-jdk7 and kotlin-stdlib-jdk8 are the recommended
dependencies to use with Kotlin 1.2 for Java 9+ compatibility.
Closes gh-11716
Add `@ControllerEndpoint` and `@RestControllerEndpoint` annotations that
can be used to develop a Spring-only request mapped endpoint. Both
Spring MVC and Spring WebFlux are supported.
This feature is primarily for use when deeper Spring integration is
required or when existing Spring Boot 1.5 projects want to migrate to
Spring Boot 2.0 without re-writing existing endpoints. It comes at the
expense of portability, since such endpoints will be missing from
Jersey.
Fixes gh-10257
Create a `PathMappedEndpoint` interface that allows any `ExposedEndpoint`
to provide root path details. The `EndpointPathResolver` interface has
been renamed to `PathMapper` and is now only used during endpoint
discovery.
`EndpointPathProvider` has been replaced with `PathMappedEndpoints`
which simply finds relevant path mapped endpoints.
Fixes gh-10985
Refactor several areas of the actuator endpoint code in order to make
future extensions easier. The primary goal is to introduce the concept
of an `ExposableEndpoint` that has technology specific subclasses and
can carry additional data for filters to use. Many other changes have
been made along the way including:
* A new EndpointSupplier interface that allows cleaner separation of
supplying vs discovering endpoints. This allows cleaner class names
and allows for better auto-configuration since a user can choose to
provide their own supplier entirely.
* A `DiscoveredEndpoint` interface that allows the `EndpointFilter`
to be greatly simplified. A filter now doesn't need to know about
discovery concerns unless absolutely necessary.
* Improved naming and package structure. Many technology specific
concerns are now grouped in a better way. Related concerns are
co-located and concepts from one area no longer leakage into another.
* Simplified `HandlerMapping` implementations. Many common concerns have
been pulled up helping to create simpler subclasses.
* Simplified JMX adapters. Many of the intermediary `Info` classes have
been removed. The `DiscoveredJmxOperation` is now responsible for
mapping methods to operations.
* A specific @`HealthEndpointCloudFoundryExtension` for Cloud Foundry.
The extension logic used to create a "full" health endpoint extension
has been made explicit.
Fixes gh-11428
Fixes gh-11581
Previously, the server was created with out an explicitly configured
address. This lead to it using any local address which will prefer
IPv6 (::0) if it's available. By contrast, the client was created
with a base URL that specified localhost as the host. This meant the
the client would prefer to connect to IPv4. Normally this wouldn't
cause a problem as nothing would be listening on the port in the IPv4
stack so the client would then connect to the server being tested
using the IPv6 stack. However, if another process was listening to the
port in the IPv4 stack, the client would connect to the wrong server.
This could lead to an unexpected 404 response (if the wrong server
was an HTTP server) or a hang if it was not.
There's a chance, although I think it's unlikely, that the problem
described above is the cause of gh-10569. I think it's unlikely as
the hang tracked by gh-10569 only occurs when running the WebFlux
endpoint integration tests using Reactor Netty. If it was the problem
described above, there's no reason that I can think of why we
wouldn't have also seen it with the Web MVC endpoint integration
tests.
As `validation-api` 2 is available by default, this commit adds the
integration test as a regular test case. The integration test is kept to
exercise what happens in a standard project.
See gh-11512
This commit prevents the default error view from rendering itself if the
response has been committed already. In this case, it is impossible to
change the HTTP response status and write a proper response - trying to
do so often results in a `IllegalStateException` since the response body
has already been written to.
Fixes gh-11580
There is a suspicion that the use of epoll is causing the intermittent
failures being tracked by gh-10569. This commit disables the use of
epoll to see if it improves the situation.
See gh-10569
Previously, the default Logback configuration set the console
appenders charset to UTF-8. This was inconsistent with the Logback
file appender and Log4j2's console and file appenders, all of which
used the platform's default.
This commit removes the configuration of the Logback console
appender's charset. This means that it will use the platform's
default charset, aligning it with the Logback file appender and
both Log4j2 appenders.
Closes gh-11611
Previously, the ServletContext was configured after any
ServletContextInitializer beans had been initialized. This meant that
any configuration class that provided such a bean would be initialized
before the ServletContext was configured. If the configuration class
used the ServletContext in its initializtaion that it would see it in
its default, unconfigured state.
This commit reworks the configuration of the ServletContext so that
it happens before any ServletContextInitializer beans are initialized.
Closes gh-10699
Phillip Webb
Andy Wilkinson
Madhura Bhave
dreis2211
Johnny Lim
Stephane Nicoll
Tony Findeisen
Jon Schneider
Vedran Pavic
Jon Schneider
Gary Russell
Nikolay Rybak
igor-suhorukov
Dominic Gunn
yulin
Huang YunKun
Raja Kolli
Eddú Meléndez
Brian Clozel
andrey.onufreyko
sdeleuze
Arnaud Cogoluègnes