RSAAndroid has two impossible padding check tests:

Both of which should be "!= && !=".

Theoretically, there should be a test that passes in bad padding values, which should be triggering Debug.Fails in a later switch statement. (The purpose of the precondition here is to avoid triggering lazy key generation on an argument exception/PNSE)

(Filing an issue because my dev environment is currently offline)

• Verify that a test for this exists. If it doesn't, write one. If there does, we're probably only running it in release builds.
• Fix the conditions.