如何让 sbt-assembly 合并正确？

Question

在我们的 Scala/Scalatra 项目中，我们有针对插件 sbt-assembly 的合并策略：

assemblyMergeStrategy in assembly := {
  case x =>
   val oldStrategy = (assemblyMergeStrategy in assembly).value
   oldStrategy(x)
}

[error] 11 errors were encountered during merge java.lang.RuntimeException: deduplicate: different file contents found in the following: ~/.ivy2/cache/org.scalatra/scalatra_2.11/jars/scalatra_2.11-2.3.1.jar:mime.types ~/.ivy2/cache/com.amazonaws/aws-java-sdk-s3/jars/aws-java-sdk-s3-1.10.1.jar:mime.types

deduplicate: different file contents found in the following:
~/.ivy2/cache/commons-beanutils/commons-beanutils/jars/commons-beanutils-1.8.3.jar:org/apache/commons/collections/ArrayStack.class
~/.ivy2/cache/commons-collections/commons-collections/jars/commons-collections-3.2.1.jar:org/apache/commons/collections/ArrayStack.class
deduplicate: different file contents found in the following:

and the same error for different class names

这里正确的合并逻辑是什么？

版本：

斯卡拉：2.11.7
SBT：0.13.9
sbt-assembly: 0.13.0

Answer 1

我不认为，这是 "merge-strategy" 的问题，但更多的是您正在使用的库及其依赖项。

谁"pulls"这些依赖项？您具体使用了哪些库？

一种限制方法是使用 excludeAll（以及类似的）和依赖声明。（参见 library management in SBT ），例如

excludeAll(
  ExclusionRule("commons-beanutils", "commons-beanutils-core"),
  ExclusionRule("commons-collections", "commons-collections"),
  ExclusionRule("commons-logging", "commons-logging"),
  ExclusionRule("org.slf4j", "slf4j-log4j12"),
  ExclusionRule("org.hamcrest", "hamcrest-core"),
  ExclusionRule("junit", "junit"),
  ExclusionRule("org.jboss.netty", "netty"),
  ExclusionRule("com.esotericsoftware.minlog", "minlog")
  )

Answer 2

我的原始问题已通过以下方式解决：

assemblyMergeStrategy in assembly := {
  case PathList("org", "apache", "commons", "collections", xs @ _*) =>
      MergeStrategy.last
  case PathList("mime.types") =>
      MergeStrategy.last
  case x =>
      val oldStrategy = (assemblyMergeStrategy in assembly).value
      oldStrategy(x)
}

如何让 sbt-assembly 合并正确？

How to get sbt-assembly merge right?

scala

sbt

scalatra