Correct publish location (apache-spark-on-k8s#268)

Author: robert3005

dev/publish.sh

@@ -20,10 +20,10 @@ make_dist() {
   build_flags="$1"
   shift 1
   hadoop_name="hadoop-palantir"
-  artifact_name="spark-dist_2.11-${hadoop_name}-${version}.tgz"
+  artifact_name="spark-dist_2.11-${hadoop_name}"
   file_name="spark-dist-${version}-${hadoop_name}.tgz"
   ./dev/make-distribution.sh --name "hadoop-palantir" --tgz "$@" $build_flags
-  curl -u $BINTRAY_USERNAME:$BINTRAY_PASSWORD -T $file_name "https://api.bintray.com/content/palantir/releases/spark/${version}/org/apache/spark/${artifact_name}/${version}/${artifact_name}"
+  curl -u $BINTRAY_USERNAME:$BINTRAY_PASSWORD -T $file_name "https://api.bintray.com/content/palantir/releases/spark/${version}/org/apache/spark/${artifact_name}/${version}/${artifact_name}-${version}.tgz"
 }
 
 publish_artifacts | tee -a "$CIRCLE_ARTIFACTS/publish.log"

dists/hadoop-palantir/pom.xml

@@ -110,11 +110,6 @@
         <artifactId>spark-dist_${scala.binary.version}-hadoop-palantir</artifactId>
         <version>${project.version}</version>
       </dependency>
-      <dependency>
-        <groupId>org.apache.spark</groupId>
-        <artifactId>spark-dist</artifactId>
-        <version>${project.version}-hadoop-${hadoop.version}</version>
-      </dependency>
     <!--
       Propagate these from spark-parent hadoop-palantir profile
     -->