Unit Test fails in R# and not in nUnit

I'm digging with a very strange problem... I've a bunch of tests that fails using R# integrated test runenr, but are ok using NUnit GUI, Gallio Icarus Runner and TestDriven.Net runner!

I've checked my code with the debugger, and looks like my objects are in an invalid status using R# runner and are ok with other runners: as example a collection that is empty in NUnit is not empty using R# (and in my code actually i create a new empty collection!).

Anyone experiencing problem like this?


We've seen problems when our unit tests had unintended consequences affecting the other tests in the same class (i.e., static collections). R# runs tests in the order they're found in the file whereas almost all other test runners run the tests in alphabetical order. This can be one cause for problems to occur in R#'s test runner and not others.




My tests are "atomic", so order is not important... but I try to verify if this behavour is due to a my mistake


OK was my fault :(
Some tests fails if in a "wrong" order... and the order used by NUnit is different from the order by R#.
Since TestCases needs to be atomic, this is absolutely not a bug of R# but a bug of my mind.
Sorry for the time lost


Please sign in to leave a comment.