Basically the acceptance test should be that, for every action which should be logged, is it actually appearing in the log? You have a user story ("Log stuff") which needs to be broken into individual user stories ("Log this activity," "Log that activity."), and each child story needs a test.
On Thu, Jul 9, 2009 at 10:03 AM, Tomas Pollak <tomas_pollak@...> wrote:
Hello all,
I'm working on an agile project and we have been writing acceptance tests for every story so far.
We are using Log4J for logging, and so far the developers have been using it at their own will. But recently some problem that wasn't logged correctly has raised our customer's awareness on this, and he included a story for the next sprint saying that we should define some logging criteria to prevent these problems in the future, and apply it uniformly throughout the code.
My question is: how do you write acceptance tests for such a requirement? Our goal is that whatever tests we write on this sprint will also be useful as the code continues to evolve and more features (and their logging) are added.
Do you have some experience where you tested that the logging was correct that you can share? Any useful links?
Thanks,
Tomás
--
John Overbaugh
blog: http://thoughtsonqa.blogspot.com