Github user JoshRosen commented on a diff in the pull request:
https://github.com/apache/spark/pull/4215#discussion_r23984864
--- Diff: core/src/main/scala/org/apache/spark/deploy/SparkSubmit.scala ---
@@ -475,6 +501,195 @@ object SparkSubmit {
}
}
+/** Provides utility functions to be used inside SparkSubmit. */
+private[spark] object SparkSubmitUtils extends Logging {
+
+ /**
+ * Represents a Maven Coordinate
+ * @param groupId the groupId of the coordinate
+ * @param artifactId the artifactId of the coordinate
+ * @param version the version of the coordinate
+ */
+ private[spark] case class MavenCoordinate(groupId: String, artifactId:
String, version: String)
+
+/**
+ * Extracts maven coordinates from a comma-delimited string
+ * @param coordinates Comma-delimited string of maven coordinates
+ * @return Sequence of Maven coordinates
+ */
+ private[spark] def extractMavenCoordinates(coordinates: String):
Seq[MavenCoordinate] = {
+ coordinates.split(",").map { p =>
+ val splits = p.split(":")
+ require(splits.length == 3, s"Provided Maven Coordinates must be in
the form " +
+ s"'groupId:artifactId:version'. The coordinate provided is: $p")
+ require(splits(0) != null && splits(0).trim.nonEmpty, s"The groupId
cannot be null or " +
+ s"be whitespace. The groupId provided is: ${splits(0)}")
+ require(splits(1) != null && splits(1).trim.nonEmpty, s"The
artifactId cannot be null or " +
+ s"be whitespace. The artifactId provided is: ${splits(1)}")
+ require(splits(2) != null && splits(2).trim.nonEmpty, s"The version
cannot be null or " +
+ s"be whitespace. The version provided is: ${splits(2)}")
+ new MavenCoordinate(splits(0), splits(1), splits(2))
+ }
+ }
+
+ /**
+ * Extracts maven coordinates from a comma-delimited string
+ * @param remoteRepos Comma-delimited string of remote repositories
+ * @return A ChainResolver used by Ivy to search for and resolve
dependencies.
+ */
+ private[spark] def createRepoResolvers(remoteRepos: Option[String]):
ChainResolver = {
+ // We need a chain resolver if we want to check multiple repositories
+ val cr = new ChainResolver
+ cr.setName("list")
+
+ // the biblio resolver resolves POM declared dependencies
+ val br: IBiblioResolver = new IBiblioResolver
+ br.setM2compatible(true)
+ br.setUsepoms(true)
+ br.setName("central")
+ cr.add(br)
+
+ val repositoryList = remoteRepos.getOrElse("")
+ // add any other remote repositories other than maven central
+ if (repositoryList.trim.nonEmpty) {
+ repositoryList.split(",").zipWithIndex.foreach { case (repo, i) =>
+ val brr: IBiblioResolver = new IBiblioResolver
+ brr.setM2compatible(true)
+ brr.setUsepoms(true)
+ brr.setRoot(repo)
+ brr.setName(s"repo-${i + 1}")
+ cr.add(brr)
+ logInfo(s"$repo added as a remote repository with the name:
${brr.getName}")
+ }
+ }
+ cr
+ }
+
+ /**
+ * Output a comma-delimited list of paths for the downloaded jars to be
added to the classpath
+ * (will append to jars in SparkSubmit). The name of the jar is given
+ * after a '!' by Ivy. It also sometimes contains '(bundle)' after
'.jar'. Remove that as well.
+ * @param artifacts Sequence of dependencies that were resolved and
retrieved
+ * @param cacheDirectory directory where jars are cached
+ * @return a comma-delimited list of paths for the dependencies
+ */
+ private[spark] def resolveDependencyPaths(
+ artifacts: Array[AnyRef],
+ cacheDirectory: File): String = {
+ artifacts.map { artifactInfo =>
+ val artifactString = artifactInfo.toString
+ val jarName = artifactString.drop(artifactString.lastIndexOf("!") +
1)
+ cacheDirectory.getAbsolutePath + "/" + jarName.substring(0,
jarName.lastIndexOf(".jar") + 4)
--- End diff --
Hardcoding `/` as the file separator character will probably break things
on Windows; I think we should use `File.separator` instead.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
