Closed rhwood closed 7 years ago
In this case, I think the percent covered decreased only because the amount of coverable code decreased.
What I find interesting is that the coveralls check appears, but the Travis one isn't, and it should be. We know it ran, because that is how the coveralls check was triggered.
I'm going to close and re-open to see if that appears.
@pabender Thanks; I hadn't clued into that.
Ok, tests area appearing now. This is the first real commit since the test coverage was instituted. I'm going to commit this, and hen I need to adjust the threshholds for failure on coveralls.
@rhwood, you snuck another change in here.
I want to get this committed before @balazsracz fixes the commit history, which is going to force us to rebase our local repositories.
@pabender Go ahead and merge. I put a change that the first Travis CI build caught. Nothing more to add.
To the original topic of this commit: after we do logging, will the errors appear on the JMRI console?
As for the rebase: I'm ready with the script; I can do whenever you guys tell me you're ready.
They will reliably appear, they will be controllable, so if someone wants debug level output, that can be turned on only for that user without a source code change. Dumping to STDOUT/ERR does not ensure the message is logged, as not all uses of the JMRI libraries capture that in output in logs.
Ok, merging.
@balazsracz go ahead and run the script when you can.
Use a java.util.logging.Logger in place of most uses of ``System.(err|out).println``` in the sources (but not in the tests) so that consuming applications (mostly JMRI) can manage logging. See JMRI/JMRI#3063