Someone designed code that relied on full data; the XML always had every element. The data source is now sending sparse XML; if it would have been empty before, it's missing now. So, it's time to refactor while fixing bugs.
There's 100+ lines of code like this:
functionDoSomething(foo, bar, getRoot().getChild("1").getChild("A").
Except now, getChild("A") might return null. Or any of the getChild(xxx) methods might.
As one additional twist, instead of getChild(), there are actually four separate methods, which can only occur in certain orders. Someone suggested a varargs call, which isn't a bad idea, but won't work as cleanly as I might like.
What's the quickest way to clean this one up? The best? "try/catch" around every line was suggested, but man, that's ugly. Breaking out the third argument to the above method into it's own function could work... but that would entail 100+ new methods, which feels ugly, albeit less so.
The number of getChild(xxx) calls is somewhere between six and ten per line, with no fixed depth. There's also no possible way to get a correct DTD for this; things will be added later without a prior heads up, and where I'd prefer a warning in the logs when that happens, extra lines in the XML need to be handled gracefully.
getChild() is a convenience method, actually. The cleanest way I have in mind is to have the convenience methods return a valid Child object, but have that "empty" Child's getContent() always return "".
以上就是Null pointers everywhere because data is suddenly sparse.的详细内容，更多请关注web前端其它相关文章！