Move `spring.rabbitmq.listener.*` to `spring.rabbitmq.listener.simple.*`
in preparation for Spring AMQP 2.0 that supports different container
types.
Closes gh-9108
See gh-9055
We intend to remove this limitation in a future milestone but as the
Lettuce auto-configuration requires `commons-pool2`, this commit updates
the How-To to mention that.
See gh-9091
The testing support in the CLI has proven to be more trouble than
it's worth. Our recommendation is that, once an app gets to the stage
of requiring a test suite, it should be converted to a Maven or
Gradle project. This makes it easy to version, publish, deploy etc
using the vast ecosystems of the two build systems.
As part of this change, the dependency management for Spock has been
moved into spring-boot-parent, thereby making it "private". This
allows it to continue to manage the test-only Spock dependency in
spring-boot-test without also managing the version of Spring that is
used by a user's application.
Closes gh-9087
Fixes gh-9043
The new spring-jcl module is now a dependency of spring-core. This
commit updates the CLI's dependency resolution-related tests to
account for the new dependency.
This commit provides an auto-configuration for reactive Redis and a
starter that provides Lettuce as Jedis doesn't support reactive
operations.
There are no support for reactive redis repositories at the moment so
only a `ReactiveRedisTemplate` is auto-configured if necessary.
Closes gh-8053
This fix for gh-1928 to address a problemw with Log4j was to set
a system property that forces JBoss Logging to use SLF4J. This has the
unwanted side-effect of crashing Glassfish. We no longer support Log4j
so gh-1928 is no longer a concern. However, you can shoot yourself in
the foot by having Logback and Log4j2 on the classpath in that
particular order, but the latter has to be without its
log4j-slf4j-impl module. At this point you don't get any warnings or
error output at startup, but JBoss Logging uses Log4j2 while
everything else uses SLF4J and Logback. As a result, any logging
that's performed through JBoss Logging is silently lost.
Given the problems caused by setting the system property, and the
contrived nature of the arrangement that's necessary to cause a
problem when the property isn't set, this commit updates
LogbackLoggingSystem so that the system property is no longer set.
Closes gh-8669
Stephane Nicoll
Madhura Bhave
Phillip Webb
Andy Wilkinson
Vedran Pavic
Eddú Meléndez