Github user pwendell commented on a diff in the pull request:
https://github.com/apache/spark/pull/237#discussion_r11324292
--- Diff: bin/compute-classpath.sh ---
@@ -71,6 +57,23 @@ else
CLASSPATH="$CLASSPATH:$ASSEMBLY_JAR"
fi
+# When Hive support is needed, Datanucleus jars must be included on the
classpath.
+# Datanucleus jars do not work if only included in the uberjar as
plugin.xml metadata is lost.
+# Both sbt and maven will populate "lib_managed/jars/" with the
datanucleus jars when Spark is
+# built with Hive, so first check if the datanucleus jars exist, and then
ensure the current Spark
+# assembly is built for Hive, before actually populating the CLASSPATH
with the jars.
+# Note that this check order is faster (by up to half a second) in the
case where Hive is not used.
+num_datanucleus_jars=$(ls "$FWDIR"/lib_managed/jars/ | grep
"datanucleus-.*\\.jar" | wc -l)
+if [ $num_datanucleus_jars -gt 0 ]; then
+ AN_ASSEMBLY_JAR=${ASSEMBLY_JAR:-$DEPS_ASSEMBLY_JAR}
+ num_hive_files=$(jar tvf "$AN_ASSEMBLY_JAR"
org/apache/hadoop/hive/ql/exec 2>/dev/null | wc -l)
--- End diff --
also an extra space after `tvf` unless it's intentional?
---
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.
---
