Fixed java options in Travis causing test failure

[sonofmun]

Fix for #108

[coveralls]

Coverage increased (+0.7%) to 94.597% when pulling 2394bcc7b6018a33ee52b24a73569d7b286652d9 on java_options into 85295c558b616e2fbef6f1993268fc7f8f0d3535 on master.

[coveralls]

Coverage increased (+0.7%) to 94.597% when pulling 2394bcc7b6018a33ee52b24a73569d7b286652d9 on java_options into 85295c558b616e2fbef6f1993268fc7f8f0d3535 on master.

[coveralls]

Coverage increased (+0.7%) to 94.592% when pulling 8e39bb8fde2a4457bc9862bf75efa0cc4d1247eb on java_options into 85295c558b616e2fbef6f1993268fc7f8f0d3535 on master.

[coveralls]

Coverage increased (+0.7%) to 94.592% when pulling 8e39bb8fde2a4457bc9862bf75efa0cc4d1247eb on java_options into 85295c558b616e2fbef6f1993268fc7f8f0d3535 on master.

[PonteIneptique]

@sonofmun up to you !

[sonofmun]

I like the way you re-organized the code. There is still no test, however, for the case on Travis where they have a _JAVA_OPTIONS environment variable that prints either (most likely) to stdout or (less likely) to stderr. We can see that it works since the files now pass (e.g., https://travis-ci.org/sonofmun/canonical-lat-test/builds/277242587?utm_source=github_status&utm_medium=notification). But we are not testing whether this actually removes the lines that it is supposed to. Perhaps that is no big deal, I don't know. Just saying.

[PonteIneptique]

You're completely right. I think it could just be left as an issue for later, issue that would not need a new release, unlike the need for this quickfix

[sonofmun]

Good. I will add that to the issue description (#108) and merge this.

[sonofmun]

Ach, Mensch. No update of the CHANGES.txt or setup.py. I am not a very good administrator. Do you want to do that or should I?

[PonteIneptique]

I'd prefer that you do it ;)