Github user markhamstra commented on the pull request:
https://github.com/apache/spark/pull/2477#issuecomment-58293963
This certainly works, but I'm not sure that we need to maintain the
complexity of having both a `jars` directory and a `lib` symlink to it. What
we want at a minimum is for the unmodified scripts distributed with Spark (such
as compute-classpath.sh) to work correctly in the package, so that looks like
we now must have the assembly in `lib`. The only reason to retain `jars` is
for backward compatibility and not to break any existing users' scripts that
expect the assembly in `jars`. I'd rather not have both `lib` and `jars`, the
Debian packaging has never really been a supported part of Spark, and anybody
using the package with an installation management system (all that the package
was every really intended for) is already going to be knee-deep in making
changes to various configuration files and scripts, so I don't see the burden
of having just `lib` and not `jars` as too onerous.
In the end, it's a judgment call. @pwendell or @mateiz ?
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
