JoshRosen commented on code in PR #49212: URL: https://github.com/apache/spark/pull/49212#discussion_r1904751056
########## core/src/main/scala/org/apache/spark/util/BestEffortLazyVal.scala: ########## @@ -0,0 +1,58 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +package org.apache.spark.util + +import java.util.concurrent.atomic.AtomicReference + +/** + * A util class to lazily initialize a variable, re-computation is allowed if multiple + * threads are trying to initialize it concurrently. Review Comment: Following the change to restore the use of `AtomicReference` and the `compareAndSet`, this now offers a slightly stronger guarantee than the text currently implies: If there are concurrent initializations then it may evaluate `compute` more than once, but (due to the use of `compareAndSet`) callers and readers will see the same value. For example, if we have a lazily-initialized mutable object then it might be very important to ensure that readers see the _same_ value. Given this, I think we might consider re-wording the Scaladoc to be slightly more precise, e.g. perhaps something like ```scala /** * A lock-free implementation of a lazily-initialized variable. * If there are concurrent initializations then the `compute()` function may be invoked * multiple times. However, only a single `compute()` result will be stored and all readers * will receive the same result object instance. * ``` -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
