如何使用 sbt-assembly 排除测试依赖项答案

【问题标题】：How to exclude test dependencies with sbt-assembly如何使用 sbt-assembly 排除测试依赖项
【发布时间】：2019-08-23 11:55:06
【问题描述】：

我有一个 sbt 项目，我正在尝试使用 sbt-assembly 插件将其构建到 jar 中。

build.sbt:

      name := "project-name"

      version := "0.1"

      scalaVersion := "2.11.12"

      val sparkVersion = "2.4.0"

      libraryDependencies ++= Seq(
        "org.scalatest" %% "scalatest" % "3.0.5" % "test",
        "org.apache.spark" %% "spark-core" % sparkVersion % "provided",
        "org.apache.spark" %% "spark-sql" % sparkVersion % "provided",
        "org.apache.spark" %% "spark-streaming" % sparkVersion % "provided",
        "com.holdenkarau" %% "spark-testing-base" % "2.3.1_0.10.0" % "test",
        // spark-hive dependencies for DataFrameSuiteBase. https://github.com/holdenk/spark-testing-base/issues/143
        "org.apache.spark" %% "spark-hive" % sparkVersion % "provided",
        "com.amazonaws" % "aws-java-sdk" % "1.11.513" % "provided",
        "com.amazonaws" % "aws-java-sdk-sqs" % "1.11.513" % "provided",
        "com.amazonaws" % "aws-java-sdk-s3" % "1.11.513" % "provided",
        //"org.apache.hadoop" % "hadoop-aws" % "3.1.1"
        "org.json" % "json" % "20180813"
      )

      assemblyOption in assembly := (assemblyOption in assembly).value.copy(includeScala = false)
      assemblyMergeStrategy in assembly := {
       case PathList("META-INF", xs @ _*) => MergeStrategy.discard
       case x => MergeStrategy.first
      }
      test in assembly := {}

      // https://github.com/holdenk/spark-testing-base
      fork in Test := true
      javaOptions ++= Seq("-Xms512M", "-Xmx2048M", "-XX:MaxPermSize=2048M", "-XX:+CMSClassUnloadingEnabled")
      parallelExecution in Test := false

当我使用 sbt 程序集构建项目时，生成的 jar 包含 /org/junit/... 和 /org/opentest4j/... 文件

有什么办法可以不将这些测试相关的文件包含在最终的jar中？

我已尝试替换该行：

    "org.scalatest" %% "scalatest" % "3.0.5" % "test"

与：

    "org.scalatest" %% "scalatest" % "3.0.5" % "provided"

我还想知道这些文件是如何包含在 jar 中的，因为 build.sbt 中没有引用 junit（但是项目中有 junit 测试）？

更新：

    name := "project-name"

    version := "0.1"

    scalaVersion := "2.11.12"

    val sparkVersion = "2.4.0"

    val excludeJUnitBinding = ExclusionRule(organization = "junit")

    libraryDependencies ++= Seq(
      // Provided
      "org.apache.spark" %% "spark-core" % sparkVersion % "provided" excludeAll(excludeJUnitBinding),
      "org.apache.spark" %% "spark-sql" % sparkVersion % "provided" excludeAll(excludeJUnitBinding),
      "org.apache.spark" %% "spark-streaming" % sparkVersion % "provided",
      "com.holdenkarau" %% "spark-testing-base" % "2.3.1_0.10.0" % "provided" excludeAll(excludeJUnitBinding),
      "org.apache.spark" %% "spark-hive" % sparkVersion % "provided",
      "com.amazonaws" % "aws-java-sdk" % "1.11.513" % "provided",
      "com.amazonaws" % "aws-java-sdk-sqs" % "1.11.513" % "provided",
      "com.amazonaws" % "aws-java-sdk-s3" % "1.11.513" % "provided",

      // Test
      "org.scalatest" %% "scalatest" % "3.0.5" % "test",

      // Necessary
      "org.json" % "json" % "20180813"
    )

    excludeDependencies += excludeJUnitBinding

    // https://stackoverflow.com/questions/25144484/sbt-assembly-deduplication-found-error
    assemblyOption in assembly := (assemblyOption in assembly).value.copy(includeScala = false)
    assemblyMergeStrategy in assembly := {
     case PathList("META-INF", xs @ _*) => MergeStrategy.discard
     case x => MergeStrategy.first
    }


    // https://github.com/holdenk/spark-testing-base
    fork in Test := true
    javaOptions ++= Seq("-Xms512M", "-Xmx2048M", "-XX:MaxPermSize=2048M", "-XX:+CMSClassUnloadingEnabled")
    parallelExecution in Test := false

【问题讨论】：

默认情况下，sbt-assembly 不包含测试 jars。当我包含自己的依赖项（错误地）将测试框架列为运行时依赖项时，我遇到了这个问题。你知道junit里拉的是哪个包吗？
我不确定，如果我在每个依赖项后面加上“必需”，测试文件仍然包括在内。这是否意味着它没有任何包含的依赖项在运行时将它们拉入？

标签： scala sbt sbt-assembly

【解决方案1】：

要排除依赖项的某些传递依赖项，请使用excludeAll 或exclude 方法。

为项目发布 pom 时，应使用 exclude 方法。它要求排除组织和模块名称。

例如：

libraryDependencies += 
  "log4j" % "log4j" % "1.2.15" exclude("javax.jms", "jms")

excludeAll 方法更灵活，但是因为它不能在 pom.xml 中表示，所以应该只在不需要生成 pom 时使用。

例如，

libraryDependencies +=
  "log4j" % "log4j" % "1.2.15" excludeAll(
    ExclusionRule(organization = "com.sun.jdmk"),
    ExclusionRule(organization = "com.sun.jmx"),
    ExclusionRule(organization = "javax.jms")
  )

在某些情况下，应从所有依赖项中排除传递依赖项。这可以通过在 excludeDependencies 中设置 ExclusionRules 来实现（对于 sbt 0.13.8 及更高版本）。

excludeDependencies ++= Seq(
  ExclusionRule("commons-logging", "commons-logging")
)

JUnit jar 文件作为以下依赖项的一部分下载。

"org.apache.spark" %% "spark-core" % sparkVersion % "provided" //(junit)
"org.apache.spark" %% "spark-sql" % sparkVersion % "provided"// (junit)
"com.holdenkarau" %% "spark-testing-base" % "2.3.1_0.10.0" % "test" //(org.junit)

要排除 junit 文件，请更新您的依赖项，如下所示。

val excludeJUnitBinding = ExclusionRule(organization = "junit")

  "org.scalatest" %% "scalatest" % "3.0.5" % "test",
  "org.apache.spark" %% "spark-core" % sparkVersion % "provided" excludeAll(excludeJUnitBinding),
  "org.apache.spark" %% "spark-sql" % sparkVersion % "provided" excludeAll(excludeJUnitBinding),
  "org.apache.spark" %% "spark-streaming" % sparkVersion % "provided",
  "com.holdenkarau" %% "spark-testing-base" % "2.3.1_0.10.0" % "test" excludeAll(excludeJUnitBinding)

更新： 请按如下方式更新您的 build.abt。

resolvers += Resolver.url("bintray-sbt-plugins",
  url("https://dl.bintray.com/eed3si9n/sbt-plugins/"))(Resolver.ivyStylePatterns)

val excludeJUnitBinding = ExclusionRule(organization = "junit")

libraryDependencies ++= Seq(
  // Provided
  "org.apache.spark" %% "spark-core" % sparkVersion % "provided" excludeAll(excludeJUnitBinding),
  "org.apache.spark" %% "spark-sql" % sparkVersion % "provided" excludeAll(excludeJUnitBinding),
  "org.apache.spark" %% "spark-streaming" % sparkVersion % "provided",
  "com.holdenkarau" %% "spark-testing-base" % "2.3.1_0.10.0" % "provided" excludeAll(excludeJUnitBinding),
  "org.apache.spark" %% "spark-hive" % sparkVersion % "provided",
  //"com.amazonaws" % "aws-java-sdk" % "1.11.513" % "provided",
  //"com.amazonaws" % "aws-java-sdk-sqs" % "1.11.513" % "provided",
  //"com.amazonaws" % "aws-java-sdk-s3" % "1.11.513" % "provided",

  // Test
  "org.scalatest" %% "scalatest" % "3.0.5" % "test",

  // Necessary
  "org.json" % "json" % "20180813"
)

assemblyOption in assembly := (assemblyOption in assembly).value.copy(includeScala = false)
assemblyMergeStrategy in assembly := {
  case PathList("META-INF", xs @ _*) => MergeStrategy.discard
  case x => MergeStrategy.first
}

fork in Test := true
javaOptions ++= Seq("-Xms512M", "-Xmx2048M", "-XX:MaxPermSize=2048M", "-XX:+CMSClassUnloadingEnabled")
parallelExecution in Test := false

plugin.sbt

addSbtPlugin("com.eed3si9n" % "sbt-assembly" % "0.13.0")

我试过了，没有下载junit jar 文件。

【讨论】：

我尝试将唯一未提供的或测试导入更改为 org.json" % "json" % "20180813" exclude("org.junit", "junit")，仍然是 org/junit ...文件存在于 jar 中。我错过了什么吗？
org.junit 不会作为 org.json 依赖项的一部分下载。你知道junit模块的包名吗？
我的理解是，因为 org.json 是唯一没有标记为“提供”或“测试”的库，所以只有它的依赖项才会被拉入。所以从这个库中排除 org.junit 应该使确保文件没有导入到 jar 中。这有意义吗？
@AlexShapovalov junit:junit:jar 文件正在作为"org.apache.spark" %% "spark-sql" % sparkVersion % "provided" 和"com.holdenkarau" %% "spark-testing-base" % "2.3.1_0.10.0" % "test" 的一部分下载。请检查我更新的评论以排除 Junit jar 文件。
如果您遇到同样的问题，请告诉我。