When I run any particular test project by itself, all test projects will pass. But when I run all test projects together, certain ones will fail. I've narrowed these down to test projects that make use of a configuration file, so my guess is when R# runs "all", it's not loading each test project as its own process, separate from the other projects. And in turn, is probably picking the first configuration it finds and applies it to all test projects.
Anyone have a work around for this? (I mean, besides running each test project one at a time?)