In doing some google searches on FakeTable and schemabinding I see that the tSQLt knows of the issue and in one case had it on the backlog to do. I was wondering what that status of that issue was (I don't see it in bugs or feature requests right now).
I've been using tSQLt for over a year in a large DB development project where we have a couple hundred schemabound views. I wrote my own proc to "PrepareTableForFaking" that has been working (the views, for the most part are not indexed and so I'm just recreating them without schemabinding for the test). I just tried the latest tSQLt release and it doesn't work without the preparation phase yet.