John, Travis fails because of too much output during the testing phase... I am a bit puzzled as to why, as this did not seem to be a problem when run on my copy... but I will turn one debug info output and then look at this bugfix...

OK, I think I found the reason :) (See commit on master)

@johnmay OK, your patches changes the API according to a now broken test: SMILESReaderTest.testReadingSmiFile_2() which expects that if there was no name defined in the file, no property is set. Your code now sets it, but as an empty string, causing this fail:

java.lang.AssertionError: expected null, but was:<>
        at org.junit.Assert.fail(Assert.java:88)
        at org.junit.Assert.failNotNull(Assert.java:664)
        at org.junit.Assert.assertNull(Assert.java:646)
        at org.junit.Assert.assertNull(Assert.java:656)
        at org.openscience.cdk.io.SMILESReaderTest.testReadingSmiFile_2(SMILESReaderTest.java:101)

Can you let me know as release manager, what you opt for? Update patch and keep behavior, or keep this patch and change the behavior?

Otherwise looks good.

Hmm I'll update to existing - i'm guessing it's quite rare to have some entries with and without names (e.g. normally either all have names or none) so null makes sense.

Applied and pushed.

travis++, Noel had previously showed me it on OB - good to see it's use.