Scott,
When running your integration tests, do you use a separate "test" database instance and populate it with data during the test "SetUp" step? And then remove/undo the data during the "TearDown" step.

Or do you build up an in-memory object graph using a fake/in-memory repository and pass that around for testing?

Hi ksouthworth,

Yes I use a seperate test database, I don't like to use a fake repository as this isn't really testing the application through the layers i.e. I am not testing agains't a real database.

I am working on the source code for my next book and its up on codeplex http://agilenet.codeplex.com/. If you take a look at this project in a couple of weeks you will see how I use a test database and clean it before each feature.

Cheers
Scott

ok, great. Just pre-ordered the book as well, btw