AsyncConfigurer implementations are loaded too early

[ipavkovic]

Affects: 5.3.3

We use an AsyncConfigurer that needs some beans. Spring wants to create a BeanPostProcessor asyncAdvisor at
org.springframework.scheduling.annotation.ProxyAsyncConfiguration. This bean post processor uses the instance variables this.executor and this.exceptionHandler

all AsyncConfigurer are autowired to ProxyAsyncConfiguration via the following method

org.springframework.scheduling.annotation.AbstractAsyncConfiguration

	@Autowired(required = false)
	void setConfigurers(Collection<AsyncConfigurer> configurers) {
		if (CollectionUtils.isEmpty(configurers)) {
			return;
		}
		if (configurers.size() > 1) {
			throw new IllegalStateException("Only one AsyncConfigurer may exist");
		}
		AsyncConfigurer configurer = configurers.iterator().next();
		this.executor = configurer::getAsyncExecutor;
		this.exceptionHandler = configurer::getAsyncUncaughtExceptionHandler;
	}

This leads to a prelimary instance creation of all asyncConfigurer and its dependent beans.

IMHO this should be changed to use an ObjectProvider instead:

    protected ObjectProvider<AsyncConfigurer> configurersProvider;

    @Autowired(required = false)
    void setConfigurers(ObjectProvider<AsyncConfigurer> configurers) {
        this.configurersProvider = configurers;
    }

    protected void evaluateAsyncConfigurer() {
        List<AsyncConfigurer> configurers = configurersProvider.orderedStream().collect(Collectors.toList());
        if (CollectionUtils.isEmpty(configurers)) {
            return;
        }
        if (configurers.size() > 1) {
            throw new IllegalStateException("Only one AsyncConfigurer may exist");
        }
        AsyncConfigurer configurer = configurers.iterator().next();
        this.executor = configurer::getAsyncExecutor;
        this.exceptionHandler = configurer::getAsyncUncaughtExceptionHandler;
    }

method evaluateAsyncConfigurer() should be called before access to this.executor and this.exceptionHandler in child classes.

[snicoll]

Thanks for the report, this is similar to #27751 for which I am working on a fix.

[ipavkovic]

@snicoll ah, did not find that issue before, thanks for your feedback 👍