How Does Spring Parse @Component and @Service Annotations? A Deep Dive

Preface

@Component and @Service are commonly used annotations in Spring; how does Spring resolve them?

1. @Component Parsing Process

Find Entry

Since Spring Framework 2.0, an extensible XML programming mechanism maps XML schema namespaces to handlers, configured in /META-INF/spring.handlers on the classpath.

Find Core Method

Browsing ContextNamespaceHandler reveals that the core scanning is performed by ClassPathBeanDefinitionScanner#doScan, which scans and registers bean definitions.

// Actually scan for bean definitions and register them. ClassPathBeanDefinitionScanner scanner = configureScanner(parserContext, element);

The method essentially iterates over candidate components, resolves scopes, generates bean names, processes annotations, and registers the definitions.

protected Set<BeanDefinitionHolder> doScan(String... basePackages) {<br/>    Assert.notEmpty(basePackages, "At least one base package must be specified");<br/>    Set<BeanDefinitionHolder> beanDefinitions = new LinkedHashSet<>();<br/>    for (String basePackage : basePackages) {<br/>        Set<BeanDefinition> candidates = findCandidateComponents(basePackage);<br/>        for (BeanDefinition candidate : candidates) {<br/>            ScopeMetadata scopeMetadata = this.scopeMetadataResolver.resolveScopeMetadata(candidate);<br/>            candidate.setScope(scopeMetadata.getScopeName());<br/>            String beanName = this.beanNameGenerator.generateBeanName(candidate, this.registry);<br/>            if (candidate instanceof AbstractBeanDefinition) {<br/>                postProcessBeanDefinition((AbstractBeanDefinition) candidate, beanName);<br/>            }<br/>            if (candidate instanceof AnnotatedBeanDefinition) {<br/>                AnnotationConfigUtils.processCommonDefinitionAnnotations((AnnotatedBeanDefinition) candidate);<br/>            }<br/>            if (checkCandidate(beanName, candidate)) {<br/>                BeanDefinitionHolder definitionHolder = new BeanDefinitionHolder(candidate, beanName);<br/>                definitionHolder = AnnotationConfigUtils.applyScopedProxyMode(scopeMetadata, definitionHolder, this.registry);<br/>                beanDefinitions.add(definitionHolder);<br/>                registerBeanDefinition(definitionHolder, this.registry);<br/>            }<br/>        }<br/>    }<br/>    return beanDefinitions;<br/>}<br/>

The key method findCandidateComponents resides in ClassPathScanningCandidateComponentProvider and ultimately delegates to scanCandidateComponents which builds a search path, loads resources, and filters candidates.

public Set<BeanDefinition> findCandidateComponents(String basePackage) {<br/>    if (this.componentsIndex != null && indexSupportsIncludeFilters()) {<br/>        return addCandidateComponentsFromIndex(this.componentsIndex, basePackage);<br/>    } else {<br/>        return scanCandidateComponents(basePackage);<br/>    }<br/>}<br/><br/>private Set<BeanDefinition> scanCandidateComponents(String basePackage) {<br/>    Set<BeanDefinition> candidates = new LinkedHashSet<>();<br/>    try {<br/>        String packageSearchPath = ResourcePatternResolver.CLASSPATH_ALL_URL_PREFIX + resolveBasePackage(basePackage) + '/' + this.resourcePattern;<br/>        Resource[] resources = getResourcePatternResolver().getResources(packageSearchPath);<br/>        for (Resource resource : resources) {<br/>            if (resource.isReadable()) {<br/>                MetadataReader metadataReader = getMetadataReaderFactory().getMetadataReader(resource);<br/>                if (isCandidateComponent(metadataReader)) {<br/>                    ScannedGenericBeanDefinition sbd = new ScannedGenericBeanDefinition(metadataReader);<br/>                    sbd.setSource(resource);<br/>                    if (isCandidateComponent(sbd)) {<br/>                        candidates.add(sbd);<br/>                    }<br/>                }<br/>            }<br/>        }<br/>    } catch (IOException ex) { /* omitted */ }<br/>    return candidates;<br/>}<br/>

How @Service Is Handled

Spring documentation states that @Component is a generic stereotype and that @Repository, @Service, and @Controller are specializations of it. The source of @Service shows it is meta‑annotated with @Component via @AliasFor.

@Target({ElementType.TYPE})<br/>@Retention(RetentionPolicy.RUNTIME)<br/>@Documented<br/>@Component<br/>public @interface Service {<br/>    @AliasFor(annotation = Component.class)<br/>    String value() default "";<br/>}<br/>

During scanning, AnnotationTypeFilter checks both direct annotations and meta‑annotations via metadata.hasMetaAnnotation. The metadata is built by AnnotationMetadataReadingVisitor, which records meta‑annotation relationships in a metaAnnotationMap while visiting class files.

public AnnotationVisitor visitAnnotation(String desc, boolean visible) {<br/>    String className = Type.getType(desc).getClassName();<br/>    this.annotationSet.add(className);<br/>    return new AnnotationAttributesReadingVisitor(className, this.attributesMap, this.metaAnnotationMap, this.classLoader);<br/>}<br/>

The visitor’s visitEnd method recursively collects meta‑annotations of the visited annotation and stores them in metaAnnotationMap, enabling AnnotationTypeFilter#matchSelf to recognise @Service as a candidate because it inherits from @Component.

Summary

The overall flow for discovering @Component and its derivatives is:

Convert a base package into a class‑loader search path (e.g., classpath*:com/example/**/*.class).

Load all resources matching that path.

For each resource, create a MetadataReader and determine whether it is a candidate component via isCandidateComponent, which relies on AnnotationTypeFilter and meta‑annotation detection.

Add the resulting ScannedGenericBeanDefinition objects to the list returned by findCandidateComponents.

Because @Service is meta‑annotated with @Component, the same scanning logic applies, allowing Spring to treat @Service as a component automatically.