Uh oh!
There was an error while loading. Please reload this page.
Let pyspark execute files even when IPYTHON=1 - #515
Conversation
Also remove IPYTHON_OPTS
AmplabJenkins
commented
Apr 23, 2014
Merged build triggered. |
AmplabJenkins
commented
Apr 23, 2014
Merged build started. |
AmplabJenkins
commented
Apr 23, 2014
Merged build finished. All automated tests passed. |
AmplabJenkins
commented
Apr 23, 2014
All automated tests passed. |
mateiz
commented
Apr 23, 2014
Because this won't work on IPython < 2, we unfortunately can't merge it. You should just file an issue in https://issues.apache.org/jira/browse/SPARK and we will revisit it when IPython 2 has much broader adoption. Unfortunately for a while lots of people will probably still be on 1. |
JoshRosen
commented
Jul 29, 2014
If you can figure out a way to retain backwards-compatibility with IPython < 2, I'd be happy to merge this. Maybe you can do something like parsing |
As per the issue kubernetes/kubernetes#77521 (comment), issue seems to be happen only when KUBE_FASTBUILD=true. Changing the same to check if it fixes the e2e job.
This reverts commit acd5e29.
With the release of IPython 2.0 having the
PYTHONSTARTUPenvironment variable and passing a file at the same time works as expected. So, we can just pass the commandline args straight to IPython.Also remove
IPYTHON_OPTS, as you can now just pass options to IPython directlyThis puts the behavior of
IPYTHON=1in line with the regular Python interpreterNote that this won't work correctly when IPython version < 2.0