Properly take @Unremoveable into account for @ConfigMapping

[geoand]

Before this change, @Unremoveable was only being considered if there was an injection point

[quarkus-bot]

Failing Jobs - Building 42ed095

Status	Name	Step	Failures	Logs	Raw logs
✔️	JVM Tests - JDK 11
✔️	JVM Tests - JDK 17
✖	JVM Tests - JDK 17 MacOS M1	Set up runner	⚠️ Check →	Logs	Raw logs
✔️	JVM Tests - JDK 18

[radcortez]

I think this is not the right fix. We already check for the @Unremoveable annotation and don't remove these beans:

quarkus/extensions/arc/deployment/src/main/java/io/quarkus/arc/deployment/ConfigBuildStep.java

Lines 320 to 322 in 99e0134

	if (configClass.getConfigClass().isAnnotationPresent(Unremovable.class)) {
	bean.unremovable();
	}

quarkus/extensions/arc/deployment/src/main/java/io/quarkus/arc/deployment/ConfigBuildStep.java

Lines 425 to 430 in 99e0134

	for (ConfigClassBuildItem configClass : configMappingTypes.values()) {
	// We don't look in the beans here, because SR Config has an API that can retrieve the mapping without CDI
	if (!arcConfig.shouldEnableBeanRemoval() || configClass.getConfigClass().isAnnotationPresent(Unremovable.class)) {
	toRegister.add(new ConfigMappingBuildItem(configClass.getConfigClass(), configClass.getPrefix()));
	}
	}

The issue is when the bean is marked unremovable with UnremovableBeanBuildItem. Arc may not remove it, but we are not registering these with ConfigMappingBuildItem, so they are missing in SmallRye Config and throw the NoSuchElementException as detailed in #29583.

[geoand]

Okay, then I'll leave this one to you 😎

[radcortez]

Sure :)

[gsmet]

Let's close this one then.