diff options
author | Patrick Wendell <pwendell@gmail.com> | 2014-10-30 20:15:36 -0700 |
---|---|---|
committer | Aaron Davidson <aaron@databricks.com> | 2014-10-30 20:15:36 -0700 |
commit | 0734d09320fe37edd3a02718511cda0bda852478 (patch) | |
tree | bf3bb586abd9c186b2a0a63dbd53ee43a00b885f /python/pyspark/mllib/regression.py | |
parent | 26d31d15fda3f63707a28d1a1115770ad127cf8f (diff) | |
download | spark-0734d09320fe37edd3a02718511cda0bda852478.tar.gz spark-0734d09320fe37edd3a02718511cda0bda852478.tar.bz2 spark-0734d09320fe37edd3a02718511cda0bda852478.zip |
HOTFIX: Clean up build in network module.
This is currently breaking the package build for some people (including me).
This patch does some general clean-up which also fixes the current issue.
- Uses consistent artifact naming
- Adds sbt support for this module
- Changes tests to use scalatest (fixes the original issue[1])
One thing to note, it turns out that scalatest when invoked in the
Maven build doesn't succesfully detect JUnit Java tests. This is
a long standing issue, I noticed it applies to all of our current
test suites as well. I've created SPARK-4159 to fix this.
[1] The original issue is that we need to allocate extra memory
for the tests, happens by default in our scalatest configuration.
Author: Patrick Wendell <pwendell@gmail.com>
Closes #3025 from pwendell/hotfix and squashes the following commits:
faa9053 [Patrick Wendell] HOTFIX: Clean up build in network module.
Diffstat (limited to 'python/pyspark/mllib/regression.py')
0 files changed, 0 insertions, 0 deletions