SPARK-1802. Audit dependency graph when Spark is built with -Phive

This initial commit resolves the conflicts in the Hive profiles as noted in https://issues.apache.org/jira/browse/SPARK-1802 .

Most of the fix was to note that Hive drags in Avro, and so if the hive module depends on Spark's version of the `avro-*` dependencies, it will pull in our exclusions as needed too. But I found we need to copy some exclusions between the two Avro dependencies to get this right. And then had to squash some commons-logging intrusions.

This turned up another annoying find, that `hive-exec` is basically an "assembly" artifact that _also_ packages all of its transitive dependencies. This means the final assembly shows lots of collisions between itself and its dependencies, and even other project dependencies. I have a TODO to examine whether that is going to be a deal-breaker or not.

In the meantime I'm going to tack on a second commit to this PR that will also fix some similar, last collisions in the YARN profile.

Author: Sean Owen <sowen@cloudera.com>

Closes #744 from srowen/SPARK-1802 and squashes the following commits:

a856604 [Sean Owen] Resolve JAR version conflicts specific to Hive profile
(cherry picked from commit 8586bf564fe010dfc19ef26874472a6f85e355fb)

Signed-off-by: Patrick Wendell <pwendell@gmail.com>

0 files changed, 0 insertions, 0 deletions