[SPARK-4075] [Deploy] Jar url validation is not enough for Jar file

In deploy.ClientArguments.isValidJarUrl, the url is checked as follows.

    def isValidJarUrl(s: String): Boolean = s.matches("(.+):(.+)jar")

So, it allows like 'hdfs:file.jar' (no authority).

Author: Kousuke Saruta <sarutak@oss.nttdata.co.jp>

Closes apache#2925 from sarutak/uri-syntax-check-improvement and squashes the following commits:

cf06173 [Kousuke Saruta] Improved URI syntax checking

Author: sarutak

core/src/main/scala/org/apache/spark/deploy/ClientArguments.scala

@@ -17,6 +17,8 @@
 
 package org.apache.spark.deploy
 
+import java.net.{URI, URISyntaxException}
+
 import scala.collection.mutable.ListBuffer
 
 import org.apache.log4j.Level
@@ -114,5 +116,12 @@ private[spark] class ClientArguments(args: Array[String]) {
 }
 
 object ClientArguments {
-  def isValidJarUrl(s: String): Boolean = s.matches("(.+):(.+)jar")
+  def isValidJarUrl(s: String): Boolean = {
+    try {
+      val uri = new URI(s)
+      uri.getScheme != null && uri.getAuthority != null && s.endsWith("jar")
+    } catch {
+      case _: URISyntaxException => false
+    }
+  }
 }

core/src/test/scala/org/apache/spark/deploy/ClientSuite.scala

@@ -29,6 +29,12 @@ class ClientSuite extends FunSuite with Matchers {
     ClientArguments.isValidJarUrl("hdfs://someHost:1234/foo") should be (false)
     ClientArguments.isValidJarUrl("/missing/a/protocol/jarfile.jar") should be (false)
     ClientArguments.isValidJarUrl("not-even-a-path.jar") should be (false)
+
+    // No authority
+    ClientArguments.isValidJarUrl("hdfs:someHost:1234/jarfile.jar") should be (false)
+
+    // Invalid syntax
+    ClientArguments.isValidJarUrl("hdfs:") should be (false)
   }
 
 }