I understand the distinction between the two ways of writing the conditional. The point is that the documented way is surprising, and as far as I can tell every

I understand the distinction between the two ways of writing the conditional. The point is that the documented way is surprising, and as far as I can tell everyone agrees that it is wrong. It's really not a matter of making the original poster's code work, it's a matter of making the language constructs have the expected behavior. Everyone seems to agree that code that used this behavior in the documented way would be bad code. That being the case, I see no way in which changing the behavior to the expected behavior would make anything worse.