Spring cache的源码版本:spring-context-5.0.9.RELEASE.jar
项目demo代码:点我跳转
先讲自定义可以干嘛,再讲解源码:
通过自定义cache config,可以用来设置自定义的过期时间,自定义的序列化方式,自定义前缀等等。@Cacheable
注解不能设置过期时间,这点是由于cache本身是抽象,各种实现过期时间的一些具体缓存框架可能有差异,不过我觉得这是一个非常不爽的点。
所以我们来阅读源代码吧。
AbstractCacheManager
类中有一个cacheMap
变量存储所有的缓存实现,在项目初始化时,由于类中实现了InitializingBean
接口,所有会初始化缓存,代码:
public abstract class AbstractCacheManager implements CacheManager, InitializingBean {
private final ConcurrentMap<String, Cache> cacheMap = new ConcurrentHashMap<>(16);
private volatile Set<String> cacheNames = Collections.emptySet();
@Override
public void afterPropertiesSet() {
initializeCaches();
}
/**
* Initialize the static configuration of caches.
* Triggered on startup through {@link #afterPropertiesSet()};
* can also be called to re-initialize at runtime.
* @since 4.2.2
* @see #loadCaches()
*/
public void initializeCaches() {
// 1⃣️重点在loadCaches方法
Collection<? extends Cache> caches = loadCaches();
synchronized (this.cacheMap) {
this.cacheNames = Collections.emptySet();
this.cacheMap.clear();
Set<String> cacheNames = new LinkedHashSet<>(caches.size());
for (Cache cache : caches) {
String name = cache.getName();
this.cacheMap.put(name, decorateCache(cache));
cacheNames.add(name);
}
this.cacheNames = Collections.unmodifiableSet(cacheNames);
}
}
}
由于loadCaches方法是抽象的,我们实现使用的redis实现,所有直接查看org.springframework.data.redis.cache.RedisCacheManager
类的实现,阅读源代码发现:
public class RedisCacheManager extends AbstractTransactionSupportingCacheManager {
...
private final Map<String, RedisCacheConfiguration> initialCacheConfiguration;
...
@Override
protected Collection<RedisCache> loadCaches() {
//1⃣️可以看到实际上就是取initialCacheConfiguration变量的值
List<RedisCache> caches = new LinkedList<>();
for (Map.Entry<String, RedisCacheConfiguration> entry : initialCacheConfiguration.entrySet()) {
//2⃣️初始化cache
caches.add(createRedisCache(entry.getKey(), entry.getValue()));
}
return caches;
}
protected RedisCache createRedisCache(String name, @Nullable RedisCacheConfiguration cacheConfig) {
return new RedisCache(name, cacheWriter, cacheConfig != null ? cacheConfig : defaultCacheConfig);
}
...
}
通过注入自定义的cacheConfig能够使不同的key拥有不同的cache配置,达到自定义的效果。
回到上面的正题,在cacheManager
初始化完成后,当有请求来到@Cacheable注解处的方法时,会通过aop代理的形式做invoke,顶层是在CacheAspectSupport
的execute方法进行代理,
中间一个步骤省略,它最后会直接通过CacheManager去获取cache,方法为:
public abstract class AbstractCacheManager implements CacheManager, InitializingBean {
...
@Override
@Nullable
public Cache getCache(String name) {
Cache cache = this.cacheMap.get(name);
if (cache != null) {
return cache;
}
else {
// Fully synchronize now for missing cache creation...
synchronized (this.cacheMap) {
cache = this.cacheMap.get(name);
if (cache == null) {
cache = getMissingCache(name);
if (cache != null) {
cache = decorateCache(cache);
this.cacheMap.put(name, cache);
updateCacheNames(name);
}
}
return cache;
}
}
}
...
}
我们查看下RedisCache
内部调用生成缓存的方法来看一下。
public class RedisCache extends AbstractValueAdaptingCache {
@Override
public void put(Object key, @Nullable Object value) {
Object cacheValue = preProcessCacheValue(value);
...
//1⃣️ 过期时间是通过cacheConfig配置进行获取的。
cacheWriter.put(name, createAndConvertCacheKey(key), serializeCacheValue(cacheValue), cacheConfig.getTtl());
}
protected byte[] serializeCacheValue(Object value) {
if (isAllowNullValues() && value instanceof NullValue) {
return BINARY_NULL_VALUE;
}
//2⃣️ value的序列化方式也是通过cacheConfig配置来初始化的
return ByteUtils.getBytes(cacheConfig.getValueSerializationPair().write(value));
}
}
@Bean
public RedisCacheManager redisCacheManager(RedisConnectionFactory connectionFactory) {
RedisCacheManager.RedisCacheManagerBuilder builder = RedisCacheManager.builder(connectionFactory);
builder.withInitialCacheConfigurations(customCacheConfig());
return builder.build();
}
private Map<String, RedisCacheConfiguration> customCacheConfig() {
Map<String, RedisCacheConfiguration> map = new HashMap<>();
RedisCacheConfiguration redisCacheConfiguration = RedisCacheConfiguration.defaultCacheConfig().entryTtl(Duration.ofMinutes(1)).serializeKeysWith(RedisSerializationContext.SerializationPair.fromSerializer(keySerializer()))
PS: 感觉使用Spring cache还是略麻烦,不如自己实现一个基于aop的cache吧。