#781 qsarmolecular regressions

[John May]

Resolves the final two failing tests in qsar-molecular:
1. the ALOGP test was testing the wrong molecule, the mistaken was made when SMILES was inlined into cdk source code (e.g. new Atom(), new Atom()) in 2007. The commit lists the git reference.
2. the WHIMDescriptorTest had a test that 'Assert.fail("not tested")' - this is better done with coverage reports (http://wwwdev.ebi.ac.uk/research/steinbeck/jenkins/job/cdk/lastBuild/jacoco/).

https://github.com/johnmay/cdk/compare/patch/qsarmolecuar-regressions

[Egon Willighagen]

Applied and pushed.