Previously, background pre-init was started in response to the
application starting event. This meant that it would be running while
the logging turbo filter was in place and was rejecting all logging.
As a result, any logging performed during pre-init would be lost.
This commit changes background pre-init so that it is now started
in response to the application environment prepared event. By this
point, the logging turbo filter has been removed as the logging levels
have been configured from the environment.
Closes gh-21897
Previously, only UndertowOptions was used as the source of options for
both server and socket options, but it only contains server options.
As a result, attempting to configure any socket options defined by
XNIO's Options class would fail.
This commit updates the property-based configuration of options to use
UndertowOptions as the source for server options and XNIO's Options as
the source for socket options.
Fixes gh-22502
This commit fixes the auto-configuration of the Couchbase caching
support by configuring the cache once Spring Data Couchbase has gotten
a chance to be processed.
Closes gh-22542
The generated snippets were not declared as an output to the
`spring-boot-actuator-autoconfigure` test goal so when the test was
pulled from the cache, it did not contain the `generated-snippets`
directory. This directory is required as an input to the Asciidoctor
plugin.
See gh-22555
This commit introduces a new property to globally disable metrics
export. In integration tests, this property is automatically set to
disable everything but in-memory metrics.
This commit also introduces a `@AutoConfigureMetrics` annotation that
can be used for integration tests that require metrics export to operate
as they would in an application.
See gh-21658
This adds build caching and build scans.
The changes required disabling scans when using the maven invoker
plugin in order to not cause duplicate build scans when invoking other
maven builds. There is also an empty `.mvn` folder in the
spring-boot-starters project to prevent duplicate build scans as well
since there is no way to pass properties to the maven-javadoc-plugin.
The checkstyle plugin was causing a cache miss with the
`propertyExpansion` because it contains an absolute path. The absolute
path is now ignored and instead the files are added as inputs to the
checkstyle plugin. This only enables the local build cache. The remote
cache is not yet enabled.
On my local machine:
./mvnw clean install build times go from about 30 minutes to about 10 minutes.
./mvnw clean install -Pfull build times go from about 60 minutes to about 13 minutes.
See gh-22089
Update `EnvironmentPostProcessorApplicationListener` so that it can
either use values from `spring.factories` or use a factory interface.
Closes gh-22529
Add support for volume mounted directories where the filename becomes
the property key and the file contents becomes the value.
Support is provided via a dedicated `VolumeMountDirectoryPropertySource`
class which can either be used directly, or via a "volumemount:/..."
`spring.config.import` location.
Closes gh-19990
Co-authored-by: Phillip Webb <pwebb@vmware.com>
Deprecate `ConfigFileApplicationListener` and provide a replacement
mechanism that supports arbitrary config data imports.
This commit updates the following areas:
- Extract `EnvironmentPostProcessor` invocation logic from the
`ConfigFileApplicationListener` to new dedicated listener. Also
providing support for `Log` injection.
- Extract `RandomPropertySource` adding logic from the
`ConfigFileApplicationListener` to a dedicated class.
- Migrate to the recently introduced `DefaultPropertiesPropertySource`
class when moving the defaultProperties `PropertySource`
- Replace processing logic with a phased approach to ensure that
profile enablement happens in a distinct phase and that profiles
can no longer be activated on an ad-hoc basis.
- Provide a more predictable and logical import order for processing
`application.properties` and `application.yml` files.
- Add support for a `spring.config.import` property which can be used
to import additional config data. Also provide a pluggable API
allowing third-parties to resolve and load locations themselves.
- Add `spring.config.activate.on-profile` support which replaces the
existing `spring.profiles` property.
- Add `spring.config.activate.on-cloud-platform` support which allows
a config data document to be active only on a given cloud platform.
- Support a `spring.config.use-legacy-processing` property allowing the
previous processing logic to be used.
Closes gh-22497
Co-authored-by: Madhura Bhave <mbhave@vmware.com>
Update `CloudPlatform` with a new `isEnforced` method that's backed
by a `Binder` rather than the `Environment`. We'll require this when
we overhaul our external config data processing logic.
Closes gh-22498
Extract `DefaultPropertiesPropertySource` from `SpringApplication` so
that logic can be easily accessed by other classes. Specifically the
property source name and logic to move the source to the end of the
list needs to be called from several places.
Closes gh-22520
Add a new `DeferredLogFactory` interface and `DeferredLogs`
implementation that can be used when a `DeferredLog` instance is needed
but the `switchOver` method should be handled elsewhere.
This interface has primarily been added so `EnvironmentPostProcessor`
classes will no longer need to implement `ApplicationEventListener`
just to switch over their logs.
Closes gh-22496
Update `OriginTrackedPropertiesLoader` so that it can support
multi-document properties files. These are similar to multi-document
YAML files but use `#---` as the separator.
Closes gh-22495
Co-authored-by: Phillip Webb <pwebb@vmware.com>
Update `OriginTrackedYamlLoader` so that empty documents are filtered
from the result. Prior to this commit, our origin wrapper would confuse
the YAML processor and cause empty documents to be included in the Map
with a key of "document" and no value.
Closes gh-22493
Update value object binder support so that parameters can be annotated
with `@Name` if a specific property name should be used. Prior to this
commit is was not possible to use Java reserved words as property names.
Closes gh-22492
Prior to this commit, the application availability infrastructure
would mix the `AvailabilityState`, the `HealthIndicator` and the
`HealthGroup` concepts and would not align with the rest.
This commit auto-configures the livenessState and readinessState
health indicators with the relevant configuration properties.
Unlike other indicators, they are not enabled by default but might
be in future versions.
This also moves the `management.health.probes.enabled` property
to `management.endpoint.health.probes.enabled` since "probes" here
is not a health indicator but rather a configuration flag for the
health endpoint.
Finally, the probes auto-configuration is refined to automatically
add liveness and readiness indicators for the probes group if
they're not already present.
Closes gh-22107
BackgroundPreinitializer is not relevant in native images so it is
better to ignore it automatically in order to reduce the footprint
in such environment.
See gh-22392
This commit adds a check to the support code for the Gradle plugin
bootBuildImage task to ensure that the jar file that will be passed
to a builder is readable and has a valid directory. This prevents a
situation where the jar file cannot be read because it is prepended
with a launch script, and the builder does not receive any files to
process.
Notes have also been added to the Gradle plugin documentation to warn
against using a bootJar launchScript configuration and bootBuildImage
together, as well as caveats about launchScript that match the Maven
plugin documentation.
Fixes gh-22223
Prior to this commit, OnClassCondition started a thread even if the number of
passed autoconfiguration class candidates never exceeded 1. This commit only
starts a thread if there is actually work to split in half.
See gh-22294
Previously, regular bean definitions for configuration properties classes
that attempt to use constructor binding were detected in a bean factory
post-processor, ConfigurationPropertiesBeanDefinitionValidator. This
validation examined every standard bean definition and failed if it
encountered one for a class that should have used constructor binding.
There were two downsides to this approach:
1. Reflection used to identify if the bean should be using constructor
binding triggered class loading that could prevent subsequent
instrumentation by a load-time weaver.
2. The cost of the validation was incurred when there was no
misconfiguration to report.
This commit replaces ConfigurationPropertiesBeanDefinitionValidator
with a failure analyzer. This failure analyzer only runs once a failure
has occurred and the application context is not going to complete
refresh. This avoids causing problems with subsequent instrumentation
and also avoids the cost of validation and error reporting unless there
is an error.
Fixes gh-20798
As of Spring Framework 5.3, it is now possible to use `PathPatternParser`
to parse and match request mapping path patterns, as an alternative to
the current default `AntPathMatcher`.
This new implementation has been used for a while in Spring WebFlux and
it’s been designed for consistency and performance.
This commit introduces a new configuration property for opting-in this
new variant:
`spring.mvc.pathmatch.matching-strategy=path_pattern_parser`
The default option is still `ant_path_matcher` for now, but we might
change the default in future versions since Spring Framework considers
it the best choice for modern applications.
There are several behavior differences with this new variant:
* double wildcards `"**"` are rejected when used in the middle patterns,
this is only allowed as the last matching segment in a pattern.
* it is incompatible with some path matching options, like
suffix-pattern, registered-suffix-pattern or configuring a Servlet
prefix on the `DispatcherServlet` (`spring.mvc.servlet.path=/test`)
This commit introduces two `FailureAnalyzer` implementations to guide
developers when facing those issues.
Closes gh-21694
This commit upgrade to Infinispan 11.0.1.Final and update dependency
management to use the bom now that it declares all the modules that
the project ships. See gh-19791
Closes gh-22286
Previously, the tests used absolute values to verify that the work
had completed sufficiently quickly. This led to flaky tests in
environments where the performance can be variable such as CI.
This commit tries to make the tests more robust by comparing the
performance to a baseline and requiring it to be twice as fast.
Closes gh-22137
Previously, spring-boot-docs used the java-base-plugin and then added
configuration on top. This has proven to be error prone, with the most
recent problem being that the tests were not being compiled and run.
This commit changes approach and applies the java plugin to the project
instead of the java-base plugin. Now, rather than adding the necessary
configuration to the base, the unwanted pieces of the java plugin's
configuration – specifically the jar and javadoc tasks – are disabled
instead. The DeployedPlugin has also been updated so that it does not
create a publication from the java component for projects that have a
disabled jar task.
Closes gh-22284
This commit improves the validation of URLs provided in the property
'spring.redis.url' used to auto-configure a Spring Data Redis
connection. In particular, only the URL schemes 'redis://' and
'rediss://' are allowed, and any other scheme will result in a
configuration error. A failure analyzer is also provided to improve
diagnostics for common mis-configurations detected by this validation.
Fixes gh-21999
Previously, the import selector for `@ImportAutoConfiguration` did not
consider the spring.autoconfigure.exclude property when determining
which auto-configurations to exclude. This meant that tests using a
slice that included a particular auto-configuration would include it
even if the application's configuration excluded it via
spring.autoconfigure.exclude. Confusingly, this could result in a
sliced test using an auto-configuration that would be excluded in a
broader `@SpringBootTest`.
This commit updates the ImportAutoConfigurationImportSelector to
consider the spring.autoconfigure.exclude property so that sliced tests
will use a subset of the auto-configurations that a `@SpringBootTest`
would use.
Fixes gh-21736
Since Spring Framework 4.3.0.RC2, a default embedded value resolver
has been registered with the bean factory when one is not otherwise
configured. This meant that placeholders in `@Value` would be resolved
with or without PropertyPlaceholderAutoConfiguration defining a
PropertySourcesPlaceholderConfigurer bean. However, placeholders in
bean definitions would only be resolved if a
PropertySourcesPlaceholderConfigurer was defined.
This commit updates PropertyPlaceholderAutoConfigurationTests to align
with this change in Framework. We now test that placeholders are
resolved in `@Value` annotations with or without the auto-configuration
and that placeholders in bean definitions are only resolved with the
auto-configured.
Closes gh-22230
Prior to this commit, configuring a reactive Elasticsearch client would
auto-configure an Actuator Health check using a synchronous client, with
the default configuration properties (so tarting localhost:9200).
This would lead to false reports of unhealthy Elasticsearch clusters
when using reactive clients.
This commit reproduces the logic for MongoDB repositories: if a reactive
variant is available, it is selected for the health check
infrastructure.
See gh-21042
Previously, if TomcatMetricsBinder destroy() was called before it had
received an ApplicationStartedEvent an NPE would be thrown due to
TomcatMetrics being null. This NPE was then caught and logged at
warning level by the disposable bean adapter.
This prevents the NPE by checking that the TomcatMetrics instance is
null before calling close() on it.
See gh-22141
This commit improves the validation performed on the user
input provided to the layertools jarmode to provide more
clear error messages when the input is not correct and
reduce the chance of ambiguity.
Fixes gh-22042
This commit allows to set the XmlBeanDefinitionReader field from
BeanDefinitionLoader to null in a way that allows the GraalVM native
compiler to remove it from the native image when the spring.xml.ignore
flag introduced by spring-projects/spring-framework#25151
is set to true.
The purpose of this change is to allow smaller footprint on native
images without requiring to use GraalVM native substitutions
which are unmaintainable by nature and also to increase the consistency
between JVM and native images.
In order to effective, this optimization requires BeanDefinitionLoader
class to be initialized at build time.
See gh-22093
Previously, the thread dump endpoint's response could exceed
WebClient's in-memory buffer limt when there were a large number of
threads or the threads had large stacks.
This commit disables WebClient's in-memory buffer size limit so that
the test passing is not dependent on the number of active threads and
their stack sizes.
Closes gh-22101
This commit changes the information provided by
RedisReactiveHealthIndicator to include cluster details when Spring
Data Redis detects that Redis is running in a clustered configuration.
This brings the reactive and non-reactive Redis health indicators
into alignment.
Fixes gh-21514
Prior to Spring Data Redis version 2.2.8, the contents of the
Properties object returned from the
ReactiveRedisConnection.ServerCommands.info API were the same
for clustered and non-clustered Redis configurations, containing a set
of key/value pairs. This allowed ReactiveRedisHealthIndicator to get
a version property using a well-known key. Starting with Spring Data
Redis 2.2.8, the info property keys contain a host:port prefix in a
clustered Redis configuration. This prevented
ReactiveRedisHealthIndicator from getting the version property as
before and resulted in the health always being reported as DOWN.
This commit adjusts ReactiveRedisHealthIndicator to detect the
clustered configuration from Spring Data Redis and find the version
property for one of the reported cluster nodes.
Fixes gh-22061
Update `Group` properties so that the `showDetails` value does not
inherit `Show.NEVER`. Prior to this commit, the `Group` properties
would not correctly inherit a `showDetails` value from the main
`management.endpoint.health.show-details` property.
See gh-22022
Rename `ImageReferenceParser` to `Regex` and remove state. The regular
expressions are now used directly by the `ImageName` and
`ImageReference` classes with the values accessed directly from the
`Matcher`.
See gh-21495
Previously, LifecycleAutoConfiguration would check the current context
and all of its ancestors for a lifecycle processor bean, only
configuring a custom processor if one was not found. Every context
has a lifecycle processor so this check meant that lifecycle processing
timeout could not be customized in any context with a parent.
This commit updates the auto-configuration to only check the current
context.
Closes gh-22014
Prior to this commit, an image name or run image name derived from
the project name or provided by the user would be passed to the CNB
builder without validation by the Maven plugin build-image goal or
Gradle plugin bootBuildImage task. This could lead to error messages
from the plugins that are difficult to understand and diagnose.
This commit makes parsing of the image names more strict, based on
the grammar implemented by the Docker go library. This provides
validation of the image names before passing them to the builder,
with a more descriptive error message when parsing and validation
fails.
Fixes gh-21495
Previously, the configuration property table generation was reading
both the main metadata and the additional metadata from each project.
This was unnecessary as the annotation processor will have already
merged the additional metadata into the main metadata before writing
it to disk. Processing both the main and additional metadata led to
a clash as the metadata overlapped. When the entry in the additional
metadata won the clash, the resulting entry in the configuration
property table would lose any details that aren't contained in the
additional metadata.
This commit updates the property table generation code to only use
the main metadata files.
Fixes gh-21131
Previously, when file-based logging was enabled, Logback would output
the following during logging system initialization:
LOGBACK: No context given for c.q.l.core.rolling.SizeAndTimeBasedRollingPolicy
This commit updates the default logback configuration to set the
context on the SizeAndTimeBasedRollingPolicy that it creates.
Fixes gh-21056
Following the fix for gh-21989, spring-boot-starter-parent no longer
contains an <issueManagement> element. As a result the additional
content was no longer being added to the pom. This commit updates
the additions so that they are now added after the <scm> element
that is still present.
See gh-21989
Previously, Spring Boot's modules published Gradle Module Metadata
(GMM) the declared a platform dependency on spring-boot-dependencies.
This provided versions for each module's own dependencies but also had
they unwanted side-effect of pulling in spring-boot-dependencies
constraints which would influence the version of other dependencies
declared in the same configuration. This was undesirable as users
should be able to opt in to this level of dependency management, either
by using the dependency management plugin or by using Gradle's built-in
support via a platform dependency on spring-boot-dependencies.
This commit reworks how Spring Boot's build uses
spring-boot-dependencies and spring-boot-parent to provide its own
dependency management. Configurations that aren't seen by consumers are
configured to extend a dependencyManagement configuration that has an
enforced platform dependency on spring-boot-parent. This enforces
spring-boot-parent's version constraints on Spring Boot's build without
making them visible to consumers. To ensure that the versions that
Spring Boot has been built against are visible to consumers, the
Maven publication that produces pom files and GMM for the published
modules is configured to use the resolved versions from the module's
runtime classpath.
Fixes gh-21911
This commit adds a runImage property to the Maven plugin build-image
goal and the Gradle bootBuildImage task. The property allows the user
to override the run image reference provided in the builder metadata
with an alternate run image. The runImage property can be specified
in the build file or on the command line.
Fixes gh-21534
Prior to this commit, the welcome page support implemented in gh-9785
would override existing index views in both annotation and functional
variants.
This comes from the fact that the feature was implemented as a
`RouterFunction` configured in the main `RouterFunctionMapping` bean.
Due to ordering issues between mappings, this would override existing
application mappings in some cases.
This commit ensures that the welcome page `RouterFunction` is
contributed to the context in its own handler mapping, ordered after the
application ones.
Fixes gh-21909
In JDK 15 the concept of hidden classes was introduced, which also
affects Lambdas in so far that Class.getCanonicalName() will return null
for those. This commit uses Class.getName() as a fallback when no
canonical name is available.
See gh-21713
This commit provides a CassandraDriverHealthIndicator and
CassandraDriverReactiveHealthIndicator that do not require Spring Data.
As a result, a health indicator for Cassandra is provided even if the
application does not use Spring Data.
See gh-20887
Update `AvailabilityChangeEvent` to be a `PayloadEvent` and ensure
that the `getResolvableType` method returns a generic compatible
result.
Prior to this commit, a ClassCastExeption would be thrown if the
following event listener was declared:
@EventListener
void onEvent(AvailabilityChangeEvent<ReadinessState> event) {
...
}
Closes gh-21898
This commit changes the NamedPipeSocket used for communication with
a local Docker daemon to use a non-blocking AsynchronousByteChannel
instead of a blocking RandomAccessFile, modeled after a similar
change to the docker-java project. This eliminates the potential for
a blocking call to hang indefinitely.
Fixes gh-21672
Pick up the workaround for the AsciidoctorJ bug that prevents the
configprops macro's attributes from being parsed. The lack of parsing
meant that the envvar format was being lost and properties were being
written in the canonical form instead.
Closes gh-21794
This commit aligns log4j2's behavior with logback such that loggers
with a null configuredLevel are also returned by the actuator endpoint.
Fixes gh-20037
This commit updates the logic for handling binding exceptions in the
management context when it is separate from the application context.
The changes allow the exception details to be visible to
DefaultErrorAttributes without causing the servlet container to
detect an error condition.
Fixes gh-21036
Previously, HttpMessageConvertersAutoConfiguration registered
ServerProperties. When this happened in a parent context, any child
contexts would skip registering ServerProperties due to its presence
in the parent context. This prevents the child contexts from
configuring their own server properties.
This commit updates HttpMessageConvertersAutoConfiguration to
bind server.servlet.encoding directly instead of enabling
ServerProperties. With ServerProperties no longer enabled in a parent
context, child contexts are now able to configure their own server
properties.
Fixes gh-21789
This commit makes sure that CassandraReactiveDataAutoConfiguration does
not create a reactiveCassandraSessionFactory bean if the user has
provided their own.
See gh-21769
Andy Wilkinson
Stephane Nicoll
Nelson Osacky
izeye
bono007
dreis2211
Dave Syer
Phillip Webb
Madhura Bhave
Brian Clozel
Sébastien Deleuze
Julien Eyraud
Sebastiaan Fernandez
Sebastiaan Fernandez
Scott Frederick
Filip Hanik
Rodolpho Couto
Vlad Kisel
Clayton Walker
XenoAmess
Julio Gomez Diaz
anshlykov
Aleksander Lech
Eddú Meléndez
Filip Hrisafov
Gaurav Pareek
im47cn
Madhura Bhave
May
Lee Dobryden
Kedar Joshi
Leo Li
Konrado85
Marco Pens
Tomek Szmytka
Radek Koubsky
Alexandre Dutra
lltx