spring-framework | Spring源码编译，附：IntelliJ IDEA2019.3编译教程 | Plugin library

This is a question about the default behavior of Spring. Say I have a singleton bean called BeanA, which has a constructor dependency to a singleton bean called BeanB. BeanB will have to be created before BeanA in order to satisfy that dependency. If both beans implement the DisposableBean interface I would expect the destroy method to be called in the reverse order that the beans were created in, but I can't see it mentioned in the documentation. The best I've found is the documentation for the DependsOn annotation (https://docs.spring.io/spring-framework/docs/current/javadoc-api/org/springframework/context/annotation/DependsOn.html) but it doesn't mention what the behavior is when DependsOn isn't used.

Edit: As I mentioned in a comment below: I've tried this out and in my test it works as expected. BeanA is destroyed before BeanB. I would like some documentation or similar to know that this is always the case though.

We are using Spring 4.3.5 in our application. While we try to uplift the Spring to 5.3.7, we are not able to initialize beans(inside "beanRefFactory.xml") inside the below tag-

I'm following the docs of spring framework and trying to realize it by code, but I'm stuck on AppConfig class. What dependency do I need to import AppConfig class? I've injected spring-beans, spring-web, spring-context, spring-webmvc already and my spring version is 5.2.5.RELEASE.

I googled "AppConfig cannot be resolved to a type" but couldn't get anything about it.

And this is the part of code from https://docs.spring.io/spring-framework/docs/current/reference/html/web.html#spring-web

is there a way to perform insertion of an @Id-column value into a non @Id column when entity is being created/updated to avoid double-save-ing?

I'm trying to write a generic function to do some Webflux operations and I'm getting a class cast exception that I can't figure out

I created WebsocketHandler as it was shown in the Webflux websocket doc.

I am trying to get my head around the Spring CachingConnectionFactory.

ActiveMQ documentation recommends when using JmsTemplate that a Spring CachingConnectionFactory or an ActiveMQ PooledConnectionFactory is used as the connection factory implementation.

I understand this because using the normal ConnectionFactory a connection is created, a session started, and both are closed for EVERY call of the jmsTemplate.send() which is very wasteful.

So I am trying to implement a custom JmsTemplate bean with a CachingConnectionFactory for use where I may have many requests that are A) Persisted to DB B) Enqueued JMS.

Often I work on Spring Boot Java applications (built with Gradle or Maven) that make network calls to external services (e.g. to external REST-like APIs). When running automated (JUnit) tests, I do not want these calls to happen, even by accident. Calling a real endpoint (even a "dev" environment one designed for testing) could affect external state, cause needless server workload, or unintentionally reveal information to the endpoint. It might also cause the test to inadvertently depend on that external resource being up, leading to build failures if the resource ever goes down while building the project, or when building the project offline.

What should typically happen with these calls is that they should either be mocked out to do nothing or should point to a service on the same machine (e.g. a WireMock endpoint on localhost spun up by the test, or a local Docker container spun up by the test using Testcontainers).

What I tend to in these situations is create a test-specific bean definition profile that overrides all HTTP endpoints with something obviously bogus that is guaranteed not to route somewhere, e.g. http://example.invalid/bookSearch. That way, if a developer misses mocking out or changing the endpoint in a test, it won't call a real one and trigger real side effects.

However, this approach is potentially error prone, and oversights in the implementation or with future changes could cause network calls to still be made. Examples include:

1. A new endpoint could be added by a developer without remembering to or knowing that they needed to override the test route
2. An endpoint could be missed being overridden in the first place
3. A third party library could be making a request that the developer didn't know about or wasn't accounted for

This is mostly a concern with Spring integration tests that spin up some or all of the environment (e.g. using @SpringBootTest or @ExtendWith(SpringExtension.class)). However, this could conceivably apply to a unit test if a component used something as a non-required constructor argument that connected to the network, or if a developer passed in a real network-connecting component where they should have used a mock or something connecting to localhost.

It occurs to me that there might be a more bullet-proof solution to this scenario. For instance, perhaps there is a way to tell the JVM or its underlying HTTP/FTP/etc. libraries to block all external network traffic.

Is there a way to prevent nonlocal network access in Java Spring Boot JUnit tests, or otherwise provide a guarantee that external network endpoints will not be called?

Spring Security here and trying to figure out how to use a MethodInvocation instance to get:

1. a list of all the arguments (name and type) passed to the method; and
2. each argument's corresponding value

There is the MethodInvocation#getArguments() : Object[] method but there is absolutely zero Spring Security documentation as to what type(s) can be returned inside the object array.