在Scala参数化(泛型)类型错误中使用Guava的简单缓存管理器

时间:2012-01-18 16:09:07

标签: generics scala guava parameterized

我对Scala很新,但仍然不了解泛型。因此,我无法弄清楚为什么编译器会因类型不匹配错误而讨厌我。

我使用Google的Guava库创建一些简单的缓存,表示为Scala的ConcurrentMap。我想跟踪创建的缓存使用另一个缓存名称Map到ConcurrentMap(缓存)。这是我到目前为止编译但是缺少对缓存的跟踪(我已经注释掉了失败的位):

import scala.collection.mutable.ConcurrentMap

trait CacheManager {

    def getCache[V <: AnyRef](
            cacheName: String,
            cacheListener: Option[CacheListener] = None): ConcurrentMap[String, V]

}


import scala.collection.JavaConversions._
import com.google.common.collect.MapMaker
import java.util.concurrent.{ConcurrentMap => JConcurrentMap, TimeUnit}
import org.slf4j.LoggerFactory
import com.google.common.cache.{RemovalNotification, RemovalListener, CacheBuilder}
import scala.collection.mutable.ConcurrentMap

class LocalCacheManager extends CacheManager {

    private val logger = LoggerFactory.getLogger(classOf[LocalCacheManager])


    private val caches /*: ConcurrentMap[String, ConcurrentMap[String, _ <: AnyRef]]*/ =
            asScalaConcurrentMap[String, ConcurrentMap[String, _ <: AnyRef]](
                new MapMaker().concurrencyLevel(4).makeMap[String, ConcurrentMap[String, _ <: AnyRef]]())

    def getCache[V <: AnyRef](cacheName: String, cacheListener: Option[CacheListener] = None) = {
//        caches.getOrElseUpdate(cacheName, {
            val cache = CacheBuilder.newBuilder()
                        .concurrencyLevel(4)
                        .softValues()
                        .expireAfterAccess(30, TimeUnit.MINUTES)
                        .build[String, V]()
            asScalaConcurrentMap[String, V](cache.asMap())
//        })
    }
}

基本上,如果我尝试将Guava缓存添加到缓存(通过注释掉的caches.getOrElseUpdate),那么编译器会抱怨以下内容:

error: type mismatch;
found   : scala.collection.mutable.ConcurrentMap[String,_$1] where type _$1 <: AnyRef
required: scala.collection.mutable.ConcurrentMap[String,V]
caches.getOrElseUpdate(cacheName, {

1 个答案:

答案 0 :(得分:3)

由于您在检索缓存时提供类型信息,因此无需尝试并维护通配符类型。将值键入AnyRef然后在结尾处将V形式转换为V要简单得多。以下编译并应该有所帮助。此外,没有必要直接调用asScalaConcurrentMap,因为它很好......隐式。

import scala.collection.JavaConversions._
import com.google.common.collect.MapMaker
import java.util.concurrent.TimeUnit
import com.google.common.cache.CacheBuilder
import scala.collection.mutable.ConcurrentMap

trait CacheListener // not sure what this is doing yet.

trait CacheManager {

    def getCache[V <: AnyRef](
            cacheName: String,
            cacheListener: Option[CacheListener] = None): ConcurrentMap[String, V]

}

class LocalCacheManager extends CacheManager {
    private val caches: ConcurrentMap[String, ConcurrentMap[String, AnyRef]] =
                new MapMaker().concurrencyLevel(4).makeMap[String, ConcurrentMap[String, AnyRef]]()
    def getCache[V <: AnyRef](cacheName: String, cacheListener: Option[CacheListener] = None) = 
      caches.getOrElseUpdate(cacheName, {
              CacheBuilder.newBuilder()
                          .concurrencyLevel(4)
                          .softValues()
                          .expireAfterAccess(30, TimeUnit.MINUTES)
                          .asInstanceOf[CacheBuilder[String, AnyRef ]]
                          .build[String, AnyRef ]()
                          .asMap()
              }).asInstanceOf[ConcurrentMap[String, V]]
}