Done. https://bugs.swift.org/browse/SR-2963.
> On 14 Oct 2016, at 22:42, John McCall <rjmcc...@apple.com> wrote: > >> On Oct 11, 2016, at 12:04 AM, Karl via swift-evolution >> <swift-evolution@swift.org <mailto:swift-evolution@swift.org>> wrote: >>> >>> On 11 Oct 2016, at 08:49, Benjamin Spratling <bspratl...@mac.com >>> <mailto:bspratl...@mac.com>> wrote: >>> >>> Howdy, >>> The error message is not saying that aFunction throws, it says “??" might >>> throw. After all, you supplied a ()rethrows->(Int) to it as its second >>> argument, which is wrapping a ()throws->Int, “bFunction()" >>> ?? and && and || wrap the trailing expression in an @autoclosure. >>> >>> I am a little surprised two “try” are not required. This would be my >>> expectation: >>>> let value = try aFunction() ?? try bFunction() >>> but, using try to the right of a non-assignment operator is not allowed. >>> >>> This, however, is not disallowed: >>> >>> let value = try aFunction() ?? (try bFunction()) >>> >>> The purpose of the @autoclosure is to make developers forget they need to >>> write a closure, and it apparently worked for you. >>> >>> -Ben Spratling >>> >> >> Yeah, I understand the reason for it, but I still think we should change it >> so you write the ‘try’ before the part which actually throws. Perhaps the >> rule should be something more general around rethrowing autoclosures? >> >> After all, I thought that was the entire point of the ‘try’ keyword. The >> compiler doesn't really need it (it already knows what throws and what >> doesn’t), but it helps humans to mark where the throwing stuff happens. > > I agree that you should be able to place the try inside the autoclosure and > have that propagate out. This doesn't need evolution discussion; please file > a bug. > > John.
_______________________________________________ swift-evolution mailing list swift-evolution@swift.org https://lists.swift.org/mailman/listinfo/swift-evolution
