We use a bunch of specific apps/APIs that (unfortunately) differ quite a bit from dev to staging/production. We use tests and continuous integration at each stage, but in dev, the tests fail annoyingly (throwing dialogs, etc – thanks Windows for the 64-bit notification!). I hate to write custom code, but are there some best practices for how to allow a subset of testing in ruby/rails – or for patching out specific tests when you’re running on Windows?
Some specific situations that:
- Identify.exe does not support 64-bit Windows and throws a dialog.
- Sethostname is not supported, and throws an error (at least it’s command line).
You can mock out the code to decouple the dependency on the other apps. Use Mocha to create the mocks dynamically.