Changed JD2 behavior to fail and exit when a bad RosettaScript is parsed and fixed 2 integration tests that were not running
The affected tests that were failing silently are:
write_mol_file and features
If you care about these tests, you might want to make sure they work properly now.
If multiple inputs were passed, the prior behavior was to iterate over
all of these inputs, even if a bad RosettaScript was used that wouldn't
work on any of them.
I added a new catch block in JD2 to specifically catch RosettaScripts
option exceptions (already existing) and reraise this exception, instead
of trying to continue as before.
The other problem that this fixes is if an integration test used a bad RosettaScript,
perhaps because the script's parse_my_tag function changed, this would only be picked
up by the initial test server run, and would fail silently after that. The
script parsing error wasn't picked up because JD2 exited with a "0" exit code.
This commit merges branch 'kylebarlow/bugs' into master
Expected test changes:
write_mol_file (wasn't running before, should at least run now)
features (wasn't running before, should at least run now)