Auditing for where more unit tests needed
Captain's Log: Stardate 77626.5
I finished up the PrecisionSlider unit tests today. They're pretty much exhaustive, and I'm glad I wrote them, because they found a number of bugs. Most of the bugs were small, but one could result in NaNs which I am really, really glad to have caught in a test rather than on some user's machine during a live performance.
I took a little time to audit the rest of my code for where more unit tests are needed. I already have pretty good coverage for the core/important stuff, but there are a few core things that I wasn't sure were permanent so I didn't write tests. At this point most of those things are definitely permanent, so I think I'll probably spend some cycles to write unit tests for them as well.
Anukari is getting fairly complex (by my solo-developer standards) at ~36 kLOC so I really need to start buttoning things down. I'm way past the point where it's fun to see the lines of code rise -- every line of code I write now makes me slightly sad, because it's adding to the accumulating pile of code that might have bugs that I don't know about. I really want to get to the point where I'm not constantly discovering stupid bugs in old code, and more tests will help.