Optimize DefaultLifecycleProcessor::startBeans

[chenqimiao]

Use lambda for minor refactoring

[wind57]

documentation of computeIfAbsent says:

return the current (existing or computed) value...

so you could do:

phases.computeIfAbsent(
						phase,
						p -> new LifecycleGroup(phase, timeoutPerShutdownPhase, lifecycleBeans, autoStartupOnly)
				).add(beanName, bean);

But looking at the implementation, phases is really just a sorted by Keys, which mean it's really supposed to be a TreeMap? So that code could be written as :

	private void startBeans(boolean autoStartupOnly) {
		Map<String, Lifecycle> lifecycleBeans = getLifecycleBeans();
		Map<Integer, LifecycleGroup> phases = new TreeMap<>();

		lifecycleBeans.forEach((beanName, bean) -> {
			if (!autoStartupOnly || (bean instanceof SmartLifecycle && ((SmartLifecycle) bean).isAutoStartup())) {

				int phase = getPhase(bean);
				phases.computeIfAbsent(
						phase,
						p -> new LifecycleGroup(phase, timeoutPerShutdownPhase, lifecycleBeans, autoStartupOnly)
				).add(beanName, bean);
			}
		});

		phases.forEach((key, value) -> value.start());
	}

The use of forEach is safe since it is documented as :

Unless otherwise specified by the implementing class, actions are performed in the order of entry set iteration

and TreeMap::entrySet is documented as :

The set's iterator returns the entries in ascending key order...

What do you think?

[chenqimiao]

I think what you said makes sense, but I added a judgment to reduce the execution of the iterator automatic code.
Thanks !

[wind57]

I'm sorry, but can you explain "but I added a judgment to reduce the execution of the iterator automatic code" - what this means? I don't get it. thank you

[sbrannen]

This has been merged into master and slightly revised in 1f35cc5.

Thanks