Github user zjffdu commented on the issue:
https://github.com/apache/zeppelin/pull/2592
It would not break the existing interpreter. I introduce `Launcher` concept
to allow customize interpreter to launch. But for now all the interpreter still
use `interpreter.sh` to launch it even for spark. But I have a plan to create
a specific script to launch spark, because there's lots of specific logic for
spark interpreter. ---
