adam-submit aborts if ADAM_HOME is set

[pbashyal-nmdp]

If ADAM_HOME is set, adam-submit aborts with the message:

java.lang.ClassNotFoundException: org.bdgenomics.adam.cli.ADAMMain
	at java.net.URLClassLoader.findClass(URLClassLoader.java:381)
	at java.lang.ClassLoader.loadClass(ClassLoader.java:424)
	at java.lang.ClassLoader.loadClass(ClassLoader.java:357)
	at java.lang.Class.forName0(Native Method)
	at java.lang.Class.forName(Class.java:348)

The bin/find-adam-assembly.sh used by adam-submit sourcesbin/find-adam-home. find-adam-home script exits when ADAM_HOME is found. Since it is sourced, the adam-submit also exits.

To reproduce this, set ADAM_HOME and try running adam-submit .

Here's my patch to get it working:

diff --git a/bin/find-adam-home b/bin/find-adam-home
index 10269078..967f62a1 100755
--- a/bin/find-adam-home
+++ b/bin/find-adam-home
@@ -25,7 +25,7 @@ FIND_ADAM_HOME_PYTHON_SCRIPT="$(cd "$(dirname "$0")"; pwd)/find_adam_home.py"

 # Short cirtuit if the user already has this set.
 if [ ! -z "${ADAM_HOME}" ]; then
-   exit 0
+   true
 elif [ ! -f "$FIND_ADAM_HOME_PYTHON_SCRIPT" ]; then
   # If we are not in the same directory as find_adam_home.py we are not pip installed so we don't
   # need to search the different Python directories for a ADAM installation.

[heuermh]

Thanks @pbashyal-nmdp! What is the use case for setting ADAM-HOME?

[pbashyal-nmdp]

Setting ADAM_HOME lets you decouple the installation location of adam from the scripts that call adam-submit. The same script can be run locally for testing and on the cluster.

Having export PATH=$ADAM_HOME/bin:$PATH also lets you run adam-submit from anywhere on the system.

[heuermh]

@fnothaft I'm still a bit shaky on changes to the bin scripts to support Python, is there a reason why find-adam-home short circuits out if ADAM_HOME is set?

[fnothaft]

@fnothaft I'm still a bit shaky on changes to the bin scripts to support Python, is there a reason why find-adam-home short circuits out if ADAM_HOME is set?

It shouldn't. I'd tested the exact use case that @pbashyal-nmdp is running locally (before we merged the PR) and it had worked for me; I'm guessing that I then made a change after testing that locally and broke things. @pbashyal-nmdp do you mind if I commit the patch you provided (after testing it on my end)?

[pbashyal-nmdp]

Yes, please use the patch. Thanks!