On Tue, Jul 11, 2017 at 2:11 PM, Niall Douglas via Boost < boost@lists.boost.org> wrote:
Outcome does not solve the most important problem that needs solving by an error handling library.
Does Expected?
Nope, but the original outcome<T> was closer. I thought, perhaps it is possible to make it able to transport arbitrary error types, and not by exception_ptr. My attempt at solving that problem lead me to TLS but maybe there are other ways. On the other hand, I don't see a problem with TLS, it seems the perfect solution. And I don't necessarily mean this from performance point of view, it's really a philosophical disagreement. The error object is a communication channel between the error-detecting code and the error-handling code; everyone else should keep quiet except maybe to add relevant contextual data. If you would accept this analogy, if I send you a package, it seems wrong for the postal service to open it up and replace the contents with something they think is better, but I'm fine with them slapping a label or two on the outside of the box. :)