[swift-evolution] Pitch: Modify the meaning of IBOutlet to remove the ! from the type

[T.J. Usiyan]

The attribute wouldn't change the behavior. The reference is still
implicitly unwrapped. IB Outlets are, at least currently, IUO by necessity.
I think that this proposal looks to make it less redundant. I didn't fully
spot the use of `let` on the first or second reading but there is another
proposal being pitched for allowing computer properties to be `let`. This
is similar in that regard. From the discussion there, it sounds like that
part might not fit with the core team's vision for the optimizer?

On Wed, May 18, 2016 at 10:36 AM, Tony Allevato via swift-evolution <
swift-evolution at swift.org> wrote:

> Similarly, a weak reference cannot be non-optional. I would be
> uncomfortable with an attribute changing that behavior. Having a
> non-optional field become nil or crash behind the scenes discards the
> safety of the Swift type system.
>
> There are legitimate situations where a class might have unused outlets as
> well, where you would want an optional (`?`) type; for example, you could
> have an iPad nib that has additional views that aren't present in an iPhone
> nib, but both share the same view controller. In that case, you would want
> to have some of those outlets still be optional. (Your pitch may not
> prevent that behavior, but it's worth pointing out as a use case.)
>
>
> On Wed, May 18, 2016 at 6:42 AM Vladimir.S via swift-evolution <
> swift-evolution at swift.org> wrote:
>
>> Correct me if I'm wrong, but `weak` reference can not be `let` just by
>> definition: at some point in time such reference *can* become nil.
>>
>> On 18.05.2016 16:22, Daniel Steinberg via swift-evolution wrote:
>> > I would like to propose we modify the meaning of @IBOutlet in light of
>> the accepted proposal SE-0054 Abolish ImplicitlyUnwrappedOptional type. I
>> think this fits in with the current Swift 3 goals.
>> >
>> > Currently we use “var” and “!” when we declare an outlet like this:
>> >
>> > @IBOutlet weak var myLabel: UILabel!
>> >
>> > The “!” in the declaration allows us to use the outlet like this
>> without unwrapping it
>> >
>> > myLabel.text = “Hello"
>> >
>> > We use “var” and “UILabel!" because myLabel starts its life out as nil
>> and does not have a value until the connection is made. i.e. myLabel must
>> be an optional if it accepts nil and the type is UILabel! instead of
>> UILabel? so that we don’t have to unwrap it each time we use it.
>> >
>> > If we break the connection to the UILabel instance we crash at runtime
>> for attempting to unwrap nil.
>> >
>> > Given this, I propose that we be able to write
>> >
>> > @IBOutlet weak let myLabel: UILabel
>> >
>> > In this case @IBOutlet has a meaning somewhat similar to lazy - it’s
>> not that myLabel doesn’t exist until we first call it, but @IBOutlet
>> indicates that myLabel should exist before we call it.
>> >
>> > If the connection isn’t made and myLabel doesn’t exist, we should crash
>> as we do now. If the connection is not made in the nib or storyboard, this
>> will crash at development time.
>> >
>> > This removes a case in which we use var - not because we want to change
>> the value of a property but because of a detail in the tooling.
>> >
>> > This change also removes a case in which we use an Optional again for a
>> detail in lifecycle and tooling.
>> >
>> > Best,
>> >
>> > Daniel
>> > _______________________________________________
>> > swift-evolution mailing list
>> > swift-evolution at swift.org
>> > https://lists.swift.org/mailman/listinfo/swift-evolution
>> >
>> _______________________________________________
>> swift-evolution mailing list
>> swift-evolution at swift.org
>> https://lists.swift.org/mailman/listinfo/swift-evolution
>>
>
> _______________________________________________
> swift-evolution mailing list
> swift-evolution at swift.org
> https://lists.swift.org/mailman/listinfo/swift-evolution
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.swift.org/pipermail/swift-evolution/attachments/20160518/416e2338/attachment-0001.html>