Ok, so basically there's two options for handling errors. The old-school approach (which I consider to be technically superior) is "if (function returns nonzero indicating error) { do error handling }". The exception approach is "try { stuff } catch ( ... ) { handle failure }".

The problem is, what happens when an error is not handled?

With the return-error approach, things explode immediately. You're using variables which weren't initialized, trying to read from a file which wasn't opened, et cetera. That sort of failure gets found fast -- often just by looking at the code and saying "wait, you're calling a function which can fail and not checking if it failed".

With the exception approach, the fact that a failure occurred bubbles up until it gets caught. At some point you catch the exception and you don't do anything which would explode immediately... but you quietly missed a layer of cleanup calls because you didn't catch the exception in the right place.

If you're really careful you make sure that every single function call is wrapped with try/catch... but at that point you've basically reimplemented return-error on top of exceptions, except with extra verbosity.