> On Sept. 19, 2014, 1:10 a.m., Lenni Kuff wrote:
> > bin/config_tool, line 49
> > <https://reviews.apache.org/r/25799/diff/2/?file=694227#file694227line49>
> >
> >     hmmm.. Since this now relies on _CMD_JAR being exported, I don't think 
> > this script can run on its own anymore. Consider moving the classpath setup 
> > stuff back in here or adding a check to ensure _CMD_JAR has been print a 
> > relevant error message. Either one seems reasonable.

Even before this change, the config_tool was not intended to be a standalone 
script. It was not setting SENTRY_HOME, hadoop path etc. The idea is to keep 
the sentry script as the top level interface and not to change the existing 
behavior.
Perhaps we can export RUN_CONFIG_TOOL and verify that setting to make sure that 
the script is not directly invoked.


- Prasad


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25799/#review53914
-----------------------------------------------------------


On Sept. 19, 2014, 1:07 a.m., Prasad Mujumdar wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/25799/
> -----------------------------------------------------------
> 
> (Updated Sept. 19, 2014, 1:07 a.m.)
> 
> 
> Review request for sentry, Lenni Kuff and Sravya Tirukkovalur.
> 
> 
> Bugs: SENTRY-430
>     https://issues.apache.org/jira/browse/SENTRY-430
> 
> 
> Repository: sentry
> 
> 
> Description
> -------
> 
> Separate out the config-tool into a separate script. This way the core sentry 
> options don't need any dependancy on Hive.
> 
> 
> Diffs
> -----
> 
>   bin/config_tool PRE-CREATION 
>   bin/sentry 2990ef4 
> 
> Diff: https://reviews.apache.org/r/25799/diff/
> 
> 
> Testing
> -------
> 
> Manually tested various Sentry script option with the patch
> 
> 
> Thanks,
> 
> Prasad Mujumdar
> 
>

Reply via email to