You are testing and returning, and responsible for propagating the error in your example, and have a chance to forget to do that. That is what IMO keeps this from being a general purpose exception. That you could correctly check the error return in one subroutine, but forget to in it's parent is the issue. Each approach has it's benefits and drawbacks. Automatically jumping to a cleanup routine means you can't forget to handle it, but also obfuscates control flow. In any case, I view it as an integral feature of exception handling.