[swift-evolution] [Review] SE-0024 "Optional Value Setter `??=`"

[David Owens II]

I just want to throw this out in case it changes your opinion - while the book does say that, the Swift library defines both of these:

@warn_unused_result
public func ??<T>(optional: T?, @autoclosure defaultValue: () throws -> T?) rethrows -> T?
@warn_unused_result
public func ??<T>(optional: T?, @autoclosure defaultValue: () throws -> T) rethrows -> T

This has the side-effect that this is legal code already today:

var o: Int? = nil
o = o ?? nil

I agree ??= has a much smaller surface area of potential use cases though.

A lot of times I have a need for something like this is when I'm parsing data into an unstructured type that will then later be realized into a strong type further in the pipeline. 

-David

> On Feb 13, 2016, at 11:52 PM, David Waite via swift-evolution <swift-evolution at swift.org> wrote:
> 
>> What is your evaluation of the proposal?
> -1, on the basis of use being a potential anti-pattern, and the spec focusing on the benefit of a terser syntax over possible uses.
> 
> If there are additional uses outside the proposal, this evaluation may change.
> 
>> Is the problem being addressed significant enough to warrant a change to Swift?
> Not as described. 
> 
> “??” is described as the nil coalescing operator.  It is defined as the following function:
> 
> public func ??<T>(optional: T?, @autoclosure defaultValue: () throws -> T) rethrows -> T
> The purpose of the nil coalescing operator is defined as follows by the Swift Programming Language book
> 
> “The nil coalescing operator (a ?? b) unwraps an optional a if it contains a value, or returns a default value b if a is nil. The expression a is always of an optional type. The expression b must match the type that is stored inside a.”
> 
> This operator has thus two core properties:
> 1. it allows another mechanism to deal with nil values, by specifying an alternate value to use in the nil case
> 2. as either the value or the alternate is non-nil, it casts off optionality
> 
> This operator does not do the second. It allows you to apply a value in the nil case, but the resulting value is still an optional. As such, I fear the usefulness of the “??=“ operator is greatly diminished.
> 
> Note also that unlike other operators with inout equivalents like addition, the type of the left-hand input and the output are not the same in “??” due to the missing safe casting off of optionality.
> 
>> Does this proposal fit well with the feel and direction of Swift?
> The example given by the proposal is as follows:
> 
> really.long.lvalue[expression] = really.long.lvalue[expression] ?? ""
> 
> In this case, you are setting a default value of “” to an optional target within a long expression.
> 
> However, it is unclear why the original expression allowed nil, or why setting this value to “” is an improvement.
> 
> To counter, some existing usage:
> 
> let userDefaults = NSUserDefaults.standardUserDefaults()
> let repeatCount = userDefaults.valueForKey("repeatCount") as? Int ?? 4
> for attempt in 0..<repeatCount {
>     // ...
> }
> 
> The for loop only works because repeatCount is non-optional. Assigning the value back into the expression would mutate the original value, but would still require one to cast off the optionality of the value.
> 
>> If you have used other languages or libraries with a similar feature, how do you feel that this proposal compares to those?
> As mentioned, Ruby has a similar behavior. Other than the “false” and “nil” literal values, everything evaluates as true. The || and && operators preserve the values of the lhs and rhs arguments.
> 
> 1 && 2 # is 2
> 1 || 2    # is 1
> nil || 2  # is 2
> 
> As such, ||= can be used for such a default assignment system.
> 
> However, ruby does not have optionality. All libraries must take into account the possibility of sending nil in for a value. As such, it is not influenced by the limitations of this proposal.
>  
>> How much effort did you put into your review? A glance, a quick reading, or an in-depth study?
> In-between a reading and a study.
> 
> -DW
> 
> _______________________________________________
> 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/20160214/9fa16078/attachment.html>