I just added code coverage report support to my work project by integrating Cobertura into my project. I'm quite happy with the simplicity, ease, and usefulness of Cobertura. It looks like my project has 55% test code coverage, which I'm rather happy with, since this includes a considerable amount of untested UI code. (I've tried to make use of JSFUnit in the past, but the code-deploy-run cycle is entirely too burdensome--it really needs a run-time interpreted scripting language to make this an effective tool, IMO.). I also think I'm including the test code itself in the coverage result, which is not honest. I'll tweak things a bit, to get some accurate results, and repost my coverage stats. Most importantly, the Java packages that I deem to be most critical have very high test coverage (80-100%).
The vast majority of the time it took to accomplish this was consumed by trying to get Ant tasks to work. Boy, I really dislike Ant, and its lack of well-written documentation. I find the uptodate task docs are rather misleading, making one think you can compare respective files in parallel directories, but this either doesn't work at all, or I'm missing something. The nested
I'm back to Scala, finally. Reading through the Artima book (printed version). Happy to see my name showed up in the acknowledgment section! (thanks to the feedback I provided on an early pre-print version) I hope to start using Scala initially by finding some scripting needs on my work project.
