[swift-evolution] Optional Assignment Operator

[John McCall]

> On Jan 25, 2017, at 1:48 PM, Xiaodi Wu <xiaodi.wu at gmail.com> wrote:
> Given lack of evidence of harm, is it really important to make such a source-breaking change?

My first instinct is that, no, it's not important.  However, we haven't actually *tried* to find any evidence of harm, so it's a bit conclusory.  If someone wants to make an evidence-based argument that it's harmful and that almost nobody is using it (intentionally/correctly), the balance could swing the other way.  That's for someone else to prove, though, since yes, at this point the bias has to be towards leaving things be.

John.

> 
> On Wed, Jan 25, 2017 at 12:45 John McCall via swift-evolution <swift-evolution at swift.org <mailto:swift-evolution at swift.org>> wrote:
>> On Jan 25, 2017, at 1:35 PM, Jacob Bandes-Storch <jtbandes at gmail.com <mailto:jtbandes at gmail.com>> wrote:
>> 
>> Agreed, IMO it would be quite dangerous for "a ??= b" to mean anything other than "a = a ?? b".
>> 
>> On another note, I don't see the value of "a? = b". I had never realized before that this works. Is this feature actually used in the wild? Should we consider removing it? (I could perhaps see some value if the assignment operator were overloadable, but it's not.)
> 
> The core semantics (that ? on an l-value still produces an l-value) fall out from the ability to call a mutating method with a?.foo().  Once you have that, you have to decide what it means to put such an l-value to the left of an assignment operator, and we decided to make it Just Work™.  I agree that it is not a particularly useful operation in idiomatic Swift, especially with simple assignment (=), and we could consider just disallowing it.
> 
> It also comes up with optional properties, I think, which is something we weren't always certain we were going to ban in native Swift (as opposed to imported ObjC code, where they're a fact of life).
> 
> John.
> 
>> 
>> Jacob
>> 
>> On Wed, Jan 25, 2017 at 10:28 AM, John McCall <rjmccall at apple.com <mailto:rjmccall at apple.com>> wrote:
>>> On Jan 25, 2017, at 12:47 PM, Jacob Bandes-Storch via swift-evolution <swift-evolution at swift.org <mailto:swift-evolution at swift.org>> wrote:
>>> Really? My observation from a quick test is that "a? = b" assigns b to a if a already has a value, or does nothing if it's nil. This is sort of the opposite of what's being proposed, which is that "a ?= b" should assign to a only if it does NOT have a value.
>> 
>> Right.  On the other hand, this does seem like a poor spelling for the operator, given the ease of confusion.
>> 
>> Also, I'm finding it hard to imagine a use for this where the equivalent ?? invocation wouldn't be *much* clearer.  It just feels like you must be doing something backwards — "I've filled in a default value for this variable, now overwrite it if this other value exists".  Wouldn't the reverse generally be better?
>> 
>> John.
>> 
>>> On Wed, Jan 25, 2017 at 9:33 AM Joe Groff via swift-evolution <swift-evolution at swift.org <mailto:swift-evolution at swift.org>> wrote:
>>> 
>>> > On Jan 25, 2017, at 8:40 AM, Nichi Shin via swift-evolution <swift-evolution at swift.org <mailto:swift-evolution at swift.org>> wrote:
>>> >
>>> > I’d like to propose a new operator for optional assignment in Swift.
>>> >
>>> > The idea is that by using this operator (e.g. by doing a ?= b), the optional on the right would be assigned to the variable on the left only when it has something to assign (i.e. when it's not nil).
>>> 
>>> `a? = b` already does this. Maybe we need a fixit to make that more apparent, though.
>>> 
>>> -Joe
>>> 
>>> >
>>> > The implementation could be something as follows:
>>> >
>>> > /// Optional Assignment Operator
>>> > infix operator ?=: AssignmentPrecedence
>>> >
>>> > func ?=<T>(left: inout T, right: T?) {
>>> >     if right != nil {
>>> >         left = right!
>>> >     }
>>> > }
>>> >
>>> > func ?=<T>(left: inout T?, right: T?) {
>>> >     if right != nil {
>>> >         left = right
>>> >     }
>>> > }
>>> >
>>> > I hope you will consider adding this on a future release of this great programming language.
>>> >
>>> > Kind regards,
>>> > N. S.
>>> > _______________________________________________
>>> > swift-evolution mailing list
>>> > swift-evolution at swift.org <mailto:swift-evolution at swift.org>
>>> > https://lists.swift.org/mailman/listinfo/swift-evolution <https://lists.swift.org/mailman/listinfo/swift-evolution>
>>> 
>>> _______________________________________________
>>> swift-evolution mailing list
>>> swift-evolution at swift.org <mailto:swift-evolution at swift.org>
>>> https://lists.swift.org/mailman/listinfo/swift-evolution <https://lists.swift.org/mailman/listinfo/swift-evolution>
>>> _______________________________________________
>>> swift-evolution mailing list
>>> swift-evolution at swift.org <mailto:swift-evolution at swift.org>
>>> https://lists.swift.org/mailman/listinfo/swift-evolution <https://lists.swift.org/mailman/listinfo/swift-evolution>
>> 
>> 
> 
> _______________________________________________
> swift-evolution mailing list
> swift-evolution at swift.org <mailto:swift-evolution at swift.org>
> https://lists.swift.org/mailman/listinfo/swift-evolution <https://lists.swift.org/mailman/listinfo/swift-evolution>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.swift.org/pipermail/swift-evolution/attachments/20170125/f2f45657/attachment.html>