<div dir="ltr">I have seen a couple of areas where generics seem lacking to me:<div><br></div><div> 1. Declaring associated types with where clauses</div><div> 2. Declaring generic arguments for calculated properties</div><div> 3. Similar to above, declaring generic properties for subscripts</div><div><br></div><div>Examples of 1 and 2:</div><div><br></div><div><p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo"><span style="color:rgb(187,44,162)">protocol</span> IterableCollection {</p>
<p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo;color:rgb(187,44,162)"><span style="color:rgb(0,0,0)"> </span>associated type<span style="color:rgb(0,0,0)"> Element</span></p><p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo;color:rgb(187,44,162)"><span style="color:rgb(0,0,0)"><br></span></p><p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo;color:rgb(187,44,162)"><span style="color:rgb(0,0,0)"> </span><span style="color:rgb(0,132,0)">/// ...</span></p><p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo;color:rgb(187,44,162)"><span style="color:rgb(0,0,0)"><br></span></p>
<p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo;color:rgb(187,44,162)"><span style="color:rgb(0,0,0)"> </span><span style="color:rgb(0,132,0)">/// Would prefer:</span><br></p>
<p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo;color:rgb(0,132,0)"><span style="color:rgb(0,0,0)"> </span>/// `var lazy<L: LazyNextableCollection where L.Element == Element>: L { get }` // Point 2 above</p>
<p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo;color:rgb(0,132,0)"><span style="color:rgb(0,0,0)"> </span>/// Or:</p>
<p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo;color:rgb(0,132,0)"><span style="color:rgb(0,0,0)"> </span>/// `associatedtype L: LazyNextableCollection where L.Element == Element` // Point 1 above</p>
<p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo;color:rgb(0,132,0)"><span style="color:rgb(0,0,0)"> </span>/// `var lazy: L { get }`</p>
<p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo;color:rgb(0,132,0)"><span style="color:rgb(0,0,0)"> </span>/// But nearest possible is a function :(.</p>
<p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo"> <span style="color:rgb(187,44,162)">func</span> lazy<L: <span style="color:rgb(79,129,135)">LazyNextableCollection</span> <span style="color:rgb(187,44,162)">where</span> <span style="color:rgb(79,129,135)">L</span>.Element == Element>() -> <span style="color:rgb(79,129,135)">L</span><span style="color:rgb(0,132,0)"> </span><span style="color:rgb(0,132,0)">// Best I seem to be able to do</span></p>
<p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo">}</p></div><div><br></div><div>Example of 3:</div><div><br></div><div><p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo"><span style="color:rgb(187,44,162)">protocol</span> SubstriptableCollection {</p>
<p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo"> <span style="color:rgb(187,44,162)">associatedtype</span> Index: Rangeable</p>
<p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo;color:rgb(187,44,162)"><span style="color:rgb(0,0,0)"> </span>associatedtype<span style="color:rgb(0,0,0)"> Element</span></p>
<p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo"><span style="color:rgb(0,0,0)"><br></span></p><p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo;color:rgb(187,44,162)"><span style="color:rgb(0,0,0)"> </span><span style="color:rgb(0,132,0)">/// ...</span></p><p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo;color:rgb(187,44,162)"><span style="color:rgb(0,0,0)"><br></span></p><p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo"><span style="color:rgb(0,0,0)"> </span><span style="color:rgb(0,132,0)">/// Ideally would write:</span><br></p>
<p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo;color:rgb(0,132,0)"><span style="color:rgb(0,0,0)"> </span>/// `subscript<S: SubstriptableCollection, R: SubstriptableCollection where S.Element == Element, R.Element == Index>(range: R) -> S { get set }` // Point 3 above</p>
<p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo;color:rgb(0,132,0)"><span style="color:rgb(0,0,0)"> </span>/// However nearest in Swift is seperate get and set methods :(.</p>
<p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo;color:rgb(0,132,0)"><span style="color:rgb(34,34,34)"> </span><span style="color:rgb(187,44,162)">func</span><span style="color:rgb(34,34,34)"> getSubscript<S: </span><span style="color:rgb(79,129,135)">SubstriptableCollection</span><span style="color:rgb(34,34,34)">, R: </span><span style="color:rgb(79,129,135)">SubstriptableCollection</span><span style="color:rgb(34,34,34)"> </span><span style="color:rgb(187,44,162)">where</span><span style="color:rgb(34,34,34)"> </span><span style="color:rgb(79,129,135)">S</span><span style="color:rgb(34,34,34)">.Element == Element, </span><span style="color:rgb(79,129,135)">R</span><span style="color:rgb(34,34,34)">.Element == Index>(range: </span><span style="color:rgb(79,129,135)">R</span><span style="color:rgb(34,34,34)">) -> </span><span style="color:rgb(79,129,135)">S</span><br></p>
<p style="margin:0px;font-size:11px;line-height:normal;font-family:Menlo">}</p></div><div><br></div></div><div class="gmail_extra"><br clear="all"><div><div class="gmail_signature"> -- Howard.<br></div></div>
<br><div class="gmail_quote">On 3 March 2016 at 19:34, Haravikk via swift-evolution <span dir="ltr"><<a href="mailto:swift-evolution@swift.org" target="_blank">swift-evolution@swift.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div style="word-wrap:break-word"><div>Nested generic types are definitely a big +1 from me. In particular if I can use them to fulfil associated type requirements, for example:</div><div><br></div><div><span style="white-space:pre-wrap"> </span>protocol FooType {</div><div><span style="white-space:pre-wrap"> </span>typealias Element</div><div><span style="white-space:pre-wrap"> </span>typealias Index</div><div><span style="white-space:pre-wrap"> </span>}</div><div><br></div><div><span style="white-space:pre-wrap"> </span>struct Foo<E> : FooType {</div><div><span style="white-space:pre-wrap"> </span>typealias Element = E</div><div><span style="white-space:pre-wrap"> </span>struct Index { … }</div><div><span style="white-space:pre-wrap"> </span>}</div><div><br></div><div>The other thing I’d like to see for generics isn’t really a new feature, but I’d like to be able to define protocol generics in the same format as for types, i.e- I could rewrite the above protocol as:</div><div><br></div><div><span style="white-space:pre-wrap"> </span>protocol FooType<Element, Index> {}</div><div><br></div><div>Likewise when placing constraints on methods etc.:</div><div><br></div><div><span style="white-space:pre-wrap"> </span>func myMethod(someFoo:FooType<String, Int>) { … }</div><div><br></div><div>Even if behind the scenes these are still unwrapped into associated types and where clauses, it’s just much, much easier to work with in the majority of cases (where clauses would still exist for the more complex ones).</div><div><br></div><div>The other capabilities you’ve described all seem very useful, but it’s probably going to take a day or two to get my head around all of them!</div><div><br></div><div><blockquote type="cite"><div>On 3 Mar 2016, at 01:22, Douglas Gregor via swift-evolution <<a href="mailto:swift-evolution@swift.org" target="_blank">swift-evolution@swift.org</a>> wrote:</div><br><div><div style="word-wrap:break-word">Hi all,<div><br></div><div><b><font size="4">Introduction</font></b></div><div><b><br></b></div><div>The “Complete Generics” goal for Swift 3 has been fairly ill-defined thus fair, with just this short blurb in the list of goals:</div><div><br></div><div><ul style="padding:0px 0px 0px 2em;margin-top:0px;margin-bottom:16px;color:rgb(51,51,51);font-family:'Helvetica Neue',Helvetica,'Segoe UI',Arial,freesans,sans-serif,'Apple Color Emoji','Segoe UI Emoji','Segoe UI Symbol';font-size:16px;background-color:rgb(255,255,255)"><li><strong>Complete generics</strong>: Generics are used pervasively in a number of Swift libraries, especially the standard library. However, there are a number of generics features the standard library requires to fully realize its vision, including recursive protocol constraints, the ability to make a constrained extension conform to a new protocol (i.e., an array of <code style="font-family:Consolas,'Liberation Mono',Menlo,Courier,monospace;font-size:14px;padding:0.2em 0px;margin:0px;background-color:rgba(0,0,0,0.0392157);border-top-left-radius:3px;border-top-right-radius:3px;border-bottom-right-radius:3px;border-bottom-left-radius:3px">Equatable</code> elements is <code style="font-family:Consolas,'Liberation Mono',Menlo,Courier,monospace;font-size:14px;padding:0.2em 0px;margin:0px;background-color:rgba(0,0,0,0.0392157);border-top-left-radius:3px;border-top-right-radius:3px;border-bottom-right-radius:3px;border-bottom-left-radius:3px">Equatable</code>), and so on. Swift 3.0 should provide those generics features needed by the standard library, because they affect the standard library's ABI.</li></ul><span class=""><div>This message expands upon the notion of “completing generics”. It is not a plan for Swift 3, nor an official core team communication, but it collects the results of numerous discussions among the core team and Swift developers, both of the compiler and the standard library. I hope to achieve several things:</div><div><br></div></span><div><ul><span class=""><li>Communicate a vision for Swift generics, building on the <a href="https://github.com/apple/swift/blob/master/docs/Generics.rst" target="_blank">original generics design document</a>, so we have something concrete and comprehensive to discuss.</li></span><span class=""><li>Establish some terminology that the Swift developers have been using for these features, so our discussions can be more productive (“oh, you’re proposing what we refer to as ‘conditional conformances’; go look over at this thread”).</li></span><div><div class="h5"><li>Engage more of the community in discussions of specific generics features, so we can coalesce around designs for public review. And maybe even get some of them implemented.</li></div></div></ul><div><br></div></div><div><div class="h5"><div>A message like this can easily turn into a <a href="http://www.urbandictionary.com/define.php?term=centithread" target="_blank">centithread</a>. To separate concerns in our discussion, I ask that replies to this specific thread be limited to discussions of the vision as a whole: how the pieces fit together, what pieces are missing, whether this is the right long-term vision for Swift, and so on. For discussions of specific language features, e.g., to work out the syntax and semantics of conditional conformances or discuss the implementation in compiler or use in the standard library, please start a new thread based on the feature names I’m using.</div><div><br></div></div></div></div><div><div class="h5"><div>This message covers a lot of ground; I’ve attempted a rough categorization of the various features, and kept the descriptions brief to limit the overall length. Most of these aren’t my ideas, and any syntax I’m providing is simply a way to express these ideas in code and is subject to change. Not all of these features will happen, either soon or ever, but they are intended to be a fairly complete whole that should mesh together. I’ve put a * next to features that I think are important in the nearer term vs. being interesting “some day”. Mostly, the *’s reflect features that will have a significant impact on the Swift standard library’s design and implementation.</div><div><br></div><div>Enough with the disclaimers; it’s time to talk features.</div><div><br></div><div><b><font size="4">Removing unnecessary restrictions</font></b></div><div><b><br></b></div><div>There are a number of restrictions to the use of generics that fall out of the implementation in the Swift compiler. Removal of these restrictions is a matter of implementation only; one need not introduce new syntax or semantics to realize them. I’m listing them for two reasons: first, it’s an acknowledgment that these features are intended to exist in the model we have today, and, second, we’d love help with the implementation of these features.</div><div><b><br></b></div><div><b><br></b></div><div><span style="font-size:14px"><i>*Recursive protocol constraints</i><br></span></div><div><br></div><div>Currently, an associated type cannot be required to conform to its enclosing protocol (or any protocol that inherits that protocol). For example, in the standard library SubSequence type of a Sequence should itself be a Sequence:</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">protocol Sequence {</font></div><div><font face="Menlo"> associatedtype Iterator : IteratorProtocol</font></div><div><font face="Menlo"> …</font></div><div><font face="Menlo"> </font><span style="font-family:Menlo">associatedtype</span><font face="Menlo"> SubSequence <b>: Sequence </b><i>// currently ill-formed, but should be possible</i></font></div><div><font face="Menlo">}</font></div></blockquote><div><br></div><div>The compiler currently rejects this protocol, which is unfortunate: it effectively pushes the SubSequence-must-be-a-Sequence requirement into every consumer of SubSequence, and does not communicate the intent of this abstraction well.</div><div><br></div><div><i style="font-size:14px">Nested generics</i></div><div><br></div><div>Currently, a generic type cannot be nested within another generic type, e.g.</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">struct X<T> {</font></div><div><font face="Menlo"> struct Y<U> { } <i>// currently ill-formed, but should be possible</i></font></div><div><font face="Menlo">}</font></div></blockquote><div><br></div><div>There isn’t much to say about this: the compiler simply needs to be improved to handle nested generics throughout.</div><div><br></div><div><br></div><div><span style="font-size:14px"><i>Concrete same-type requirements</i><br></span></div><div><br></div><div><div>Currently, a constrained extension cannot use a same-type constraint to make a type parameter equivalent to a concrete type. For example:</div></div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">extension Array <b>where Element == String</b> {</font></div><div><font face="Menlo"> func makeSentence() -> String {</font></div><div><font face="Menlo"> // uppercase first string, concatenate with spaces, add a period, whatever</font></div><div><font face="Menlo"> }</font></div><div><font face="Menlo">}</font></div></blockquote><div><br></div><div>This is a highly-requested feature that fits into the existing syntax and semantics. Note that one could imagine introducing new syntax, e.g., extending “Array<String>”, which gets into new-feature territory: see the section on “Parameterized extensions”.</div><div><br></div><div><b><font size="4">Parameterizing other declarations</font></b></div><div><br></div><div>There are a number of Swift declarations that currently cannot have generic parameters; some of those have fairly natural extensions to generic forms that maintain their current syntax and semantics, but become more powerful when made generic.</div><div><br></div><i style="font-size:14px">Generic typealiases<br></i><br><div>Typealiases could be allowed to carry generic parameters. They would still be aliases (i.e., they would not introduce new types). For example:</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">typealias StringDictionary<Value> = Dictionary<String, Value></font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo">var d1 = StringDictionary<Int>()</font></div><div><font face="Menlo">var d2: </font><span style="font-family:Menlo">Dictionary<String, Int> = d1 // okay: d1 and d2 have the same type, Dictionary<String, Int></span></div></blockquote><div><br></div><div><br></div><div><i style="font-size:14px">Generic subscripts</i><br><br>Subscripts could be allowed to have generic parameters. For example, we could introduce a generic subscript on a Collection that allows us to pull out the values at an arbitrary set of indices:</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">extension Collection {</font></div><div><font face="Menlo"> subscript<b><Indices: Sequence where Indices.Iterator.Element == Index></b>(indices: Indices) -> [Iterator.Element] {</font></div><div><font face="Menlo"> get {</font></div><div><font face="Menlo"> var result = [Iterator.Element]()</font></div><div><font face="Menlo"> for index in indices {</font></div><div><font face="Menlo"> result.append(self[index])</font></div><div><font face="Menlo"> }</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo"> return result</font></div><div><font face="Menlo"> }</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo"> set {</font></div><div><font face="Menlo"> for (index, value) in zip(indices, newValue) {</font></div><div><font face="Menlo"> self[index] = value</font></div><div><font face="Menlo"> }</font></div><div><font face="Menlo"> }</font></div><div><span style="font-family:Menlo"> }</span></div><div><font face="Menlo">}</font></div></blockquote><div><br></div><div><br><i style="font-size:14px">Generic constants</i></div><div><br></div><div>let constants could be allowed to have generic parameters, such that they produce differently-typed values depending on how they are used. For example, this is particularly useful for named literal values, e.g.,</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">let π<T : FloatLiteralConvertible>: T = 3.141592653589793238462643383279502884197169399</font></div></blockquote><div><br></div><div>The Clang importer could make particularly good use of this when importing macros.</div><div><br></div><div><br></div><div><i style="font-size:14px">Parameterized extensions</i></div><div><br></div></div></div><div><div class="h5"><div>Extensions themselves could be parameterized, which would allow some structural pattern matching on types. For example, this would permit one to extend an array of optional values, e.g.,</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">extension<b><T></b> Array <b>where Element == T?</b> {</font></div><div><font face="Menlo"> var someValues: [T] {</font></div><div><font face="Menlo"> var result = [T]()</font></div><div><font face="Menlo"> for opt in self {</font></div><div><font face="Menlo"> if let value = opt { result.append(value) }</font></div><div><font face="Menlo"> }</font></div><div><font face="Menlo"> return result</font></div><div><font face="Menlo"> }</font></div><div><font face="Menlo">}</font></div></blockquote><div><br></div><div><div>We can generalize this to a protocol extensions:</div><div><br></div></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><div><div><font face="Menlo">extension<b><T></b> Sequence <b>where Element == T?</b> {</font></div></div></div><div><div><font face="Menlo"> var someValues: [T] {</font></div></div><div><div><font face="Menlo"> var result = [T]()</font></div></div><div><div><font face="Menlo"> for opt in self {</font></div></div><div><div><font face="Menlo"> if let value = opt { result.append(value) }</font></div></div><div><div><font face="Menlo"> }</font></div></div><div><div><font face="Menlo"> return result</font></div></div><div><div><font face="Menlo"> }</font></div></div><div><div><font face="Menlo">}</font></div></div></blockquote><div><font face="Menlo"><br></font></div><div><div>Note that when one is extending nominal types, we could simplify the syntax somewhat to make the same-type constraint implicit in the syntax:</div><div><br></div><div><blockquote style="margin:0px 0px 0px 40px;border:none;padding:0px"><div><font face="Menlo">extension<b><T></b> Array<b><T?></b> {</font></div><div><font face="Menlo"> var someValues: [T] {</font></div><div><font face="Menlo"> var result = [T]()</font></div><div><font face="Menlo"> for opt in self {</font></div><div><font face="Menlo"> if let value = opt { result.append(value) }</font></div><div><font face="Menlo"> }</font></div><div><font face="Menlo"> return result</font></div><div><font face="Menlo"> }</font></div><div><font face="Menlo">}</font></div></blockquote></div></div><div><br></div><div>When we’re working with concrete types, we can use that syntax to improve the extension of concrete versions of generic types (per “Concrete same-type requirements”, above), e.g.,</div><div><br></div><div><blockquote style="margin:0px 0px 0px 40px;border:none;padding:0px"><div><font face="Menlo">extension Array<b><String></b> {</font></div><div><font face="Menlo"> func makeSentence() -> String {</font></div><div><font face="Menlo"> // uppercase first string, concatenate with spaces, add a period, whatever</font></div><div><font face="Menlo"> }</font></div><div><font face="Menlo">}</font></div><div><br></div><div><br></div></blockquote></div><div><font size="4"><b>Minor extensions</b><br></font><br></div><div>There are a number of minor extensions we can make to the generics system that don’t fundamentally change what one can express in Swift, but which can improve its expressivity.</div><div><br></div><div><span style="font-size:14px"><i>*Arbitrary requirements in protocols<br></i></span></div><div><br></div></div></div><span class=""><div>Currently, a new protocol can inherit from other protocols, introduce new associated types, and add new conformance constraints to associated types (by redeclaring an associated type from an inherited protocol). However, one cannot express more general constraints. Building on the example from “Recursive protocol constraints”, we really want the element type of a Sequence’s SubSequence to be the same as the element type of the Sequence, e.g.,</div><div><br></div><div><blockquote style="margin:0px 0px 0px 40px;border:none;padding:0px"><div><font face="Menlo">protocol Sequence {</font></div><div><font face="Menlo"> associatedtype Iterator : IteratorProtocol</font></div><div><span style="font-family:Menlo"> …</span></div><div><font face="Menlo"> </font><span style="font-family:Menlo">associatedtype</span><font face="Menlo"> SubSequence :<b> </b>Sequence<b> where SubSequence.Iterator.Element == Iterator.Element</b></font></div><div><font face="Menlo">}</font></div></blockquote></div><div><br></div><div>Hanging the where clause off the associated type is protocol not ideal, but that’s a discussion for another thread.</div><div><br></div><div><br></div><div><span style="font-size:14px"><i>*Typealiases in protocols and protocol extensions<br></i></span></div><div><br></div><div><div>Now that associated types have their own keyword (thanks!), it’s reasonable to bring back “typealias”. Again with the Sequence protocol:</div><div><br></div></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><div><div><font face="Menlo">protocol Sequence {</font></div></div></div><div><div><div><font face="Menlo"> associatedtype Iterator : IteratorProtocol</font></div></div></div><div><font face="Menlo"> typealias Element = Iterator.Element // rejoice! now we can refer to SomeSequence.Element rather than SomeSequence.Iterator.Element</font></div><div><div><font face="Menlo">}</font></div></div></blockquote></span><div><div><br></div><div><br></div><div><span style="font-size:14px"><i>Default generic arguments </i></span></div></div><span class=""><div><br></div><div>Generic parameters could be given the ability to provide default arguments, which would be used in cases where the type argument is not specified and type inference could not determine the type argument. For example:</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><div><font face="Menlo">public final class Promise<Value, Reason=Error> { … }</font></div></div><div><font face="Menlo"><br></font></div><div><font face="Menlo">func getRandomPromise() -> Promise<Int, ErrorProtocol> { … }</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo">var p1: Promise<Int> = …</font></div><div><font face="Menlo">var p2: Promise<Int, Error> = p1 <i>// okay: p1 and p2 have the same type Promise<Int, Error></i></font></div><div><font face="Menlo">var p3: Promise = getRandomPromise() <i>// p3 has type </i></font><span style="font-family:Menlo"><i>Promise<Int, ErrorProtocol> due to type inference</i></span></div></blockquote><div><br></div><div><br></div><div><i style="font-size:14px">Generalized “class” constraints</i></div><div><br></div><div>The “class” constraint can currently only be used for defining protocols. We could generalize it to associated type and type parameter declarations, e.g.,</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><div><font face="Menlo">protocol P {</font></div></div><div><div><font face="Menlo"> associatedtype A : class</font></div></div><div><div><font face="Menlo">}</font></div></div><div><div><font face="Menlo"><br></font></div></div><div><div><font face="Menlo">func foo<T : class>(t: T) { }</font></div></div></blockquote><div><br></div><div>As part of this, the magical <font face="Menlo">AnyObject</font> protocol could be replaced with an existential with a class bound, so that it becomes a typealias:</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">typealias AnyObject = protocol<class></font></div></blockquote><div><br></div><div>See the “Existentials” section, particularly “Generalized existentials”, for more information.</div><div><br></div><div><br></div></span><span class=""><div><span style="font-size:14px"><i>*Allowing subclasses to override requirements satisfied by defaults</i></span></div><div><span style="font-size:14px"><i><br></i></span></div><div>When a superclass conforms to a protocol and has one of the protocol’s requirements satisfied by a member of a protocol extension, that member currently cannot be overridden by a subclass. For example:</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">protocol P {</font></div><div><font face="Menlo"> func foo()</font></div><div><font face="Menlo">}</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo">extension P {</font></div><div><font face="Menlo"> func foo() { print(“P”) }</font></div><div><font face="Menlo">}</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo">class C : P {</font></div><div><font face="Menlo"> // gets the protocol extension’s </font></div><div><font face="Menlo">}</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo">class D : C {</font></div><div><font face="Menlo"> /*override not allowed!*/ func foo() { print(“D”) }</font></div><div><font face="Menlo">}</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo">let p: P = D()</font></div><div><font face="Menlo">p.foo() // gotcha: prints “P” rather than “D”!</font></div></blockquote><div><br></div><div>D.foo should be required to specify “override” and should be called dynamically.</div><div><br></div><div><br></div><div><b><font size="4">Major extensions to the generics model</font></b></div><div><b><br></b></div><div>Unlike the minor extensions, major extensions to the generics model provide more expressivity in the Swift generics system and, generally, have a much more significant design and implementation cost.</div><div><br></div><div><br></div><div><i style="font-size:14px">*Conditional conformances<br></i></div><div><br></div><div>Conditional conformances express the notion that a generic type will conform to a particular protocol only under certain circumstances. For example, Array is Equatable only when its elements are Equatable:</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">extension Array <b>: Equatable where Element : Equatable</b> { }</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo">func ==<T : Equatable>(lhs: Array<T>, rhs: Array<T>) -> Bool { … }</font></div></blockquote><div><br></div><div>Conditional conformances are a potentially very powerful feature. One important aspect of this feature is how deal with or avoid overlapping conformances. For example, imagine an adaptor over a Sequence that has conditional conformances to Collection and MutableCollection:</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">struct SequenceAdaptor<S: Sequence> : Sequence { }</font></div><div><font face="Menlo">extension SequenceAdaptor : Collection where S: Collection { … }</font></div><div><font face="Menlo">extension SequenceAdaptor : MutableCollection where S: MutableCollection { }</font></div></blockquote><div><br></div><div>This should almost certainly be permitted, but we need to cope with or reject “overlapping” conformances:</div><div><br></div></span><div><blockquote style="margin:0px 0px 0px 40px;border:none;padding:0px"><font face="Menlo">extension SequenceAdaptor : Collection where S: SomeOtherProtocolSimilarToCollection { } <i>// trouble: two ways for SequenceAdaptor to conform to Collection</i></font></blockquote></div><span class=""><div><font face="Menlo"><br></font></div><div>See the section on “Private conformances” for more about the issues with having the same type conform to the same protocol multiple times.</div><div><br></div><div><br></div><div><span style="font-size:14px"><i>Variadic generics</i></span></div><div><br></div><div><div>Currently, a generic parameter list contains a fixed number of generic parameters. If one has a type that could generalize to any number of generic parameters, the only real way to deal with it today involves creating a set of types. For example, consider the standard library’s “zip” function. It returns one of these when provided with two arguments to zip together:</div><div><br></div></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><div><div><font face="Menlo">public struct Zip2Sequence<Sequence1 : Sequence,</font></div><div><font face="Menlo"> Sequence2 : Sequence> : Sequence { … }</font></div></div></div><div><div><font face="Menlo"><br></font></div></div><div><div><font face="Menlo">public func zip<Sequence1 : Sequence, Sequence2 : Sequence>(</font></div><div><font face="Menlo"> sequence1: Sequence1, _ sequence2: Sequence2)</font></div><div><span style="font-family:Menlo"> -> Zip2Sequence<Sequence1, Sequence2> { … }</span></div></div></blockquote></span><div><span class=""><div><br></div><div>Supporting three arguments would require copy-paste of those of those:</div><div><br></div><div><blockquote style="margin:0px 0px 0px 40px;border:none;padding:0px"><div><font face="Menlo">public struct Zip3Sequence<Sequence1 : Sequence,</font></div><div><font face="Menlo"> Sequence2 : Sequence,</font></div><span style="font-family:Menlo"> Sequence3 : Sequence</span><span style="font-family:Menlo">> : Sequence { … }</span><div><font face="Menlo"><br></font></div><div><div><font face="Menlo">public func zip<Sequence1 : Sequence, Sequence2 : Sequence</font><span style="font-family:Menlo">, Sequence3 : Sequence</span><span style="font-family:Menlo">>(</span></div><div><span style="font-family:Menlo"> sequence1: Sequence1, _ sequence2: Sequence2, _ sequence3: sequence3)</span></div><div><font face="Menlo"> -> Zip3Sequence<Sequence1, Sequence2, Sequence3> { … }</font></div></div></blockquote><div></div></div><div><br></div></span><div>Variadic generics would allow us to abstract over a set of generic parameters. The syntax below is hopelessly influenced by <a href="http://www.jot.fm/issues/issue_2008_02/article2/" target="_blank">C++11 variadic templates</a> (sorry), where putting an ellipsis (“…”) to the left of a declaration makes it a “parameter pack” containing zero or more parameters and putting an ellipsis to the right of a type/expression/etc. expands the parameter packs within that type/expression into separate arguments. The important part is that we be able to meaningfully abstract over zero or more generic parameters, e.g.:</div><div><br></div></div><div><div class="h5"><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><div><font face="Menlo">public struct ZipIterator<... <b>Iterators</b> : IteratorProtocol</font><span style="font-family:Menlo">> : Iterator { <i>// zero or more type parameters, each of which conforms to IteratorProtocol</i></span></div></div><div><span style="font-family:Menlo"> public typealias Element = (<b>Iterators.Element...</b>) <i>// a tuple containing the element types of each iterator in Iterators</i></span></div><div><span style="font-family:Menlo"><br></span></div><div><span style="font-family:Menlo"> var (<b>...iterators</b>): (<b>Iterators...</b>) <i>// zero or more stored properties, one for each type in Iterators</i></span><span style="font-family:Menlo"> </span></div><font face="Menlo"> var reachedEnd: Bool = false</font></blockquote><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><font face="Menlo"><br></font></blockquote><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><font face="Menlo"> public mutating func next() -> Element? {</font></blockquote><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><font face="Menlo"> if reachedEnd { return nil }</font></blockquote><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><font face="Menlo"><br></font></blockquote><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><font face="Menlo"> guard let values = (<b>iterators.next()...</b>) { <i>// call “next” on each of the iterators, put the results into a tuple named “values"</i></font></blockquote><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><font face="Menlo"> reachedEnd = true</font></blockquote><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><font face="Menlo"> return nil</font></blockquote><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><font face="Menlo"> }</font></blockquote><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><font face="Menlo"><br></font></blockquote><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><font face="Menlo"> return values</font></blockquote><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><font face="Menlo"> }<br></font><div><span style="font-family:Menlo">}</span></div><div><span style="font-family:Menlo"><br></span></div></blockquote><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><div><div><div><font face="Menlo">public struct ZipSequence<<b>...Sequences</b> : Sequence</font><span style="font-family:Menlo">> : Sequence {</span></div><div><span style="font-family:Menlo"> public typealias Iterator = ZipIterator<<b>Sequences.Iterator...</b>> <i>// get the zip iterator with the iterator types of our Sequences</i></span></div><div><span style="font-family:Menlo"><i><br></i></span></div><div><div><font face="Menlo"> var (...</font><b><font face="Menlo">sequences</font></b><font face="Menlo">): (<b>Sequences</b></font><b style="font-family:Menlo">...</b><font face="Menlo">) </font><i style="font-family:Menlo">// zero or more stored properties, one for each type in Sequences</i><span style="font-family:Menlo"> </span></div><div><span style="font-family:Menlo"><br></span></div><font face="Menlo"></font></div><div><span style="font-family:Menlo"> <i>// details ...</i></span></div><div><span style="font-family:Menlo">}</span></div></div></div></div><div><div><div><br></div></div></div></blockquote><div><div>Such a design could also work for function parameters, so we can pack together multiple function arguments with different types, e.g.,</div><div><br></div><div><blockquote style="margin:0px 0px 0px 40px;border:none;padding:0px"><div><font face="Menlo">public func zip<<b>... Sequences : SequenceType</b>>(<b>... sequences: Sequences...</b>) </font></div><div><span style="font-family:Menlo"> -> ZipSequence<<b>Sequences...</b>> {</span></div><div><span style="font-family:Menlo"> return ZipSequence(<b>sequences...</b>)</span></div><div><span style="font-family:Menlo">}</span></div></blockquote><div></div></div><div><br></div><div>Finally, this could tie into the discussions about a tuple “splat” operator. For example:</div><div><br></div></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><div><font face="Menlo">func apply<... Args, Result>(fn: (Args...) -> Result, <i>// function taking some number of arguments and producing Result</i></font></div><div><font face="Menlo"> args: (Args...)) -> Result { <i>// tuple of arguments</i></font></div></div><div><div><font face="Menlo"> return fn(<b>args...</b>) // expand the arguments in the tuple “args” into separate arguments</font></div><div><font face="Menlo">}</font></div></div></blockquote><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo"><br></font></div><div><font face="Menlo"><br></font></div></blockquote><div><div><span style="font-size:14px"><i>Extensions of structural types</i></span></div></div><div><br></div><div><div>Currently, only nominal types (classes, structs, enums, protocols) can be extended. One could imagine extending structural types—particularly tuple types—to allow them to, e.g., conform to protocols. For example, pulling together variadic generics, parameterized extensions, and conditional conformances, one could express “a tuple type is Equatable if all of its element types are Equatable”:</div></div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">extension<...Elements : Equatable> <b>(Elements...)</b> : Equatable { <i>// extending the tuple type “(Elements…)” to be Equatable</i></font></div><div><font face="Menlo">}</font></div></blockquote><div><br></div><div>There are some natural bounds here: one would need to have actual structural types. One would not be able to extend every type:</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">extension<T> T { <i>// error: neither a structural nor a nominal type</i></font></div><div><font face="Menlo">}</font></div></blockquote><div><br></div><div>And before you think you’re cleverly making it possible to have a conditional conformance that makes every type T that conforms to protocol P also conform to protocol Q, see the section "Conditional conformances via protocol extensions”, below:</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">extension<T : P> T : Q { <i>// error: neither a structural nor a nominal type</i></font></div><div><font face="Menlo">}</font></div></blockquote><div><br></div><div><br></div><div><b><font size="4">Syntactic improvements</font></b></div><div><b><br></b></div><div>There are a number of potential improvements we could make to the generics syntax. Such a list could go on for a very long time, so I’ll only highlight some obvious ones that have been discussed by the Swift developers.</div><div><br></div><div><div><span style="font-size:14px"><i>*Default implementations in protocols</i></span></div></div><div><br></div><div><div>Currently, protocol members can never have implementations. We could allow one to provide such implementations to be used as the default if a conforming type does not supply an implementation, e.g.,</div><div><br></div></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><div><font face="Menlo">protocol Bag {</font></div></div><div><font face="Menlo"> associatedtype Element : Equatable</font></div><div><font face="Menlo"> func contains(element: Element) -> Bool</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo"> func containsAll<S: Sequence where Sequence.Iterator.Element == Element>(elements: S) -> Bool {</font></div><div><font face="Menlo"> for x in elements {</font></div><div><font face="Menlo"> if contains(x) { return true }</font></div><div><font face="Menlo"> }</font></div><div><font face="Menlo"> return false</font></div><div><font face="Menlo"> }</font></div><div><div><span style="font-family:Menlo">}</span></div></div><div><div><font face="Menlo"><br></font></div></div><div><div><font face="Menlo">struct IntBag : Bag {</font></div><div><font face="Menlo"> typealias Element = Int</font></div><div><font face="Menlo"> func contains(element: Int) -> Bool { ... }</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo"> // okay: containsAll requirement is satisfied by Bag’s default implementation</font></div><div><font face="Menlo">}</font></div></div><div><font face="Menlo"><br></font></div></blockquote><div><div>One can get this effect with protocol extensions today, hence the classification of this feature as a (mostly) syntactic improvement:</div><div><br></div></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><div><div><font face="Menlo">protocol Bag {</font></div></div></div><div><div><div><font face="Menlo"> associatedtype Element : Equatable</font></div></div></div><div><div><div><font face="Menlo"> func contains(element: Element) -> Bool</font></div></div></div><div><div><div><font face="Menlo"><br></font></div></div></div><div><div><div><font face="Menlo"> func containsAll<S: Sequence where Sequence.Iterator.Element == Element>(elements: S) -> Bool</font></div></div></div><div><div><div><span style="font-family:Menlo">}</span></div></div></div><div><span style="font-family:Menlo"><br></span></div><div><span style="font-family:Menlo">extension Bag {</span></div><div><div><font face="Menlo"> func containsAll<S: Sequence where Sequence.Iterator.Element == Element>(elements: S) -> Bool {</font></div><div><font face="Menlo"> for x in elements {</font></div><div><font face="Menlo"> if contains(x) { return true }</font></div><div><font face="Menlo"> }</font></div><div><font face="Menlo"> return false</font></div><div><font face="Menlo"> }</font></div><div><div></div></div></div><div><span style="font-family:Menlo">}</span></div></blockquote><div><div><br></div><div><br></div><div><span style="font-size:14px"><i>*Moving the where clause outside of the angle brackets<br></i></span></div></div><div><br></div><div>The “where” clause of generic functions comes very early in the declaration, although it is generally of much less concern to the client than the function parameters and result type that follow it. This is one of the things that contributes to “angle bracket blindness”. For example, consider the containsAll signature above:</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><span style="font-family:Menlo">func containsAll<S: Sequence where Sequence.Iterator.Element == Element>(elements: S) -> Bool</span></div></blockquote><div><br></div><div>One could move the “where” clause to the end of the signature, so that the most important parts—name, generic parameter, parameters, result type—precede it:</div><div><br></div><div><blockquote style="margin:0px 0px 0px 40px;border:none;padding:0px"><span style="font-family:Menlo">func containsAll<S: Sequence>(elements: S) -> Bool </span></blockquote><blockquote style="margin:0px 0px 0px 40px;border:none;padding:0px"><span style="font-family:Menlo"> where Sequence.Iterator.Element == Element</span></blockquote></div><div><span style="font-family:Menlo"><br></span></div><div><span style="font-family:Menlo"><br></span></div><div><span style="font-size:14px"><i>*Renaming “protocol<…>” to “Any<…>”.</i></span></div><div><br></div><div>The “protocol<…>” syntax is a bit of an oddity in Swift. It is used to compose protocols together, mostly to create values of existential type, e.g.,</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">var x: protocol<NSCoding, NSCopying></font></div></blockquote><div><br></div><div>It’s weird that it’s a type name that starts with a lowercase letter, and most Swift developers probably never deal with this feature unless they happen to look at the definition of Any:</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">typealias Any = protocol<></font></div></blockquote><div><br></div><div>“Any” might be a better name for this functionality. “Any” without brackets could be a keyword for “any type”, and “Any” followed by brackets could take the role of “protocol<>” today:</div><div><br></div><div><blockquote style="margin:0px 0px 0px 40px;border:none;padding:0px"><font face="Menlo">var x: Any<NSCoding, NSCopying></font></blockquote></div><div><br></div><div>That reads much better: “Any type that conforms to NSCoding and NSCopying”. See the section "Generalized existentials” for additional features in this space.</div><div><br></div><div><b><font size="4">Maybe</font></b></div><div><b><br></b></div><div>There are a number of features that get discussed from time-to-time, while they could fit into Swift’s generics system, it’s not clear that they belong in Swift at all. The important question for any feature in this category is not “can it be done” or “are there cool things we can express”, but “how can everyday Swift developers benefit from the addition of such a feature?”. Without strong motivating examples, none of these “maybes” will move further along.</div><div><b><br></b></div><div><div><span style="font-size:14px"><i>Dynamic dispatch for members of protocol extensions</i></span></div><div><br></div><div>Only the requirements of protocols currently use dynamic dispatch, which can lead to surprises:</div></div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">protocol P {</font></div><div><font face="Menlo"> func foo()</font></div><div><font face="Menlo">}</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo">extension P {</font></div><div><font face="Menlo"> func foo() { print(“P.foo()”)</font></div><div><font face="Menlo"> func bar() { print(“P.bar()”)</font></div><div><font face="Menlo">}</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo">struct X : P {</font></div><div><font face="Menlo"> func foo() { print(“X.foo()”)</font></div><div><font face="Menlo"> func bar() { print(“X.bar()”)</font></div><div><font face="Menlo">}</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo">let x = X()</font></div><div><font face="Menlo">x.foo() // X.foo()</font></div><div><font face="Menlo">x.bar() // X.bar()</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo">let p: P = X()</font></div><div><div><font face="Menlo">p.foo() // X.foo()</font></div></div><div><div><font face="Menlo">p.bar() // P.bar()</font></div></div></blockquote><div><br></div><div>Swift could adopt a model where members of protocol extensions are dynamically dispatched.</div><div><br></div><div><div><span style="font-size:14px"><i>Named generic parameters<br></i></span></div></div><div><br></div><div>When specifying generic arguments for a generic type, the arguments are always positional: Dictionary<String, Int> is a Dictionary whose Key type is String and whose Value type is Int, by convention. One could permit the arguments to be labeled, e.g.,</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">var d: Dictionary<<b>Key:</b> String, <b>Value:</b> Int></font></div></blockquote><div><br></div><div>Such a feature makes more sense if Swift gains default generic arguments, because generic argument labels would allow one to skip defaulted arguments.</div><div><br></div><div><div><span style="font-size:14px"><i>Generic value parameters<br></i></span></div><div><br></div><div>Currently, Swift’s generic parameters are always types. One could imagine allowing generic parameters that are values, e.g.,</div><div><br></div><blockquote style="margin:0px 0px 0px 40px;border:none;padding:0px"><div><font face="Menlo">struct MultiArray<T,<b> let Dimensions: Int</b>> { <i>// specify the number of dimensions to the array</i></font></div><div><font face="Menlo"> subscript (indices: Int...) -> T {</font></div><div><font face="Menlo"> get {</font></div><div><font face="Menlo"> require(indices.count == <b>Dimensions</b>)</font></div><div><font face="Menlo"> // ...</font></div><div><font face="Menlo"> }</font></div><div><font face="Menlo">}</font></div></blockquote><div><br></div><div>A suitably general feature might allow us to express fixed-length array or vector types as a standard library component, and perhaps also allow one to implement a useful dimensional analysis library. Tackling this feature potentially means determining what it is for an expression to be a “constant expression” and diving into dependent-typing, hence the “maybe”.</div><div><br></div></div><div><span style="font-size:14px"><i>Higher-kinded types<br></i></span></div><div><br></div><div>Higher-kinded types allow one to express the relationship between two different specializations of the same nominal type within a protocol. For example, if we think of the Self type in a protocol as really being “Self<T>”, it allows us to talk about the relationship between “Self<T>” and “Self<U>” for some other type U. For example, it could allow the “map” operation on a collection to return a collection of the same kind but with a different operation, e.g.,</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">let intArray: Array<Int> = …</font></div><div><font face="Menlo">intArray.map { String($0) } <i>// produces Array<String></i></font></div><div><font face="Menlo">let intSet: Set<Int> = …</font></div><div><font face="Menlo">intSet.map { String($0) } <i>// produces Set<String></i></font></div></blockquote><div><br></div><div><br></div><div>Potential syntax borrowed from <a href="https://lists.swift.org/pipermail/swift-evolution/Week-of-Mon-20151214/002736.html" target="_blank">one thread on higher-kinded types</a> uses ~= as a “similarity” constraint to describe a Functor protocol:</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">protocol Functor {</font></div><div><font face="Menlo"> associatedtype A</font></div><div><font face="Menlo"> func fmap<FB where <b>FB ~= Self</b>>(f: A -> FB.A) -> FB</font></div><div><font face="Menlo">}</font></div></blockquote><div><br></div><div><br></div><div><div><span style="font-size:14px"><i>Specifying type arguments for uses of generic functions</i></span></div></div><div><br></div><div>The type arguments of a generic function are always determined via type inference. For example, given:</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">func f<T>(t: T)</font></div></blockquote><div><br></div><div>one cannot directly specify T: either one calls “f” (and T is determined via the argument’s type) or one uses “f” in a context where it is given a particular function type (e.g., “let x: (Int) -> Void = f” would infer T = Int). We could permit explicit specialization here, e.g.,</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">let x = f<Int> // x has type (Int) -> Void</font></div></blockquote><div><br></div><div><br></div><div><b><font size="4">Unlikely</font></b></div><div><br></div><div>Features in this category have been requested at various times, but they don’t fit well with Swift’s generics system because they cause some part of the model to become overly complicated, have unacceptable implementation limitations, or overlap significantly with existing features.</div><div><br></div><div><div><span style="font-size:14px"><i>Generic protocols</i></span></div><div><br></div><div>One of the most commonly requested features is the ability to parameterize protocols themselves. For example, a protocol that indicates that the Self type can be constructed from some specified type T:</div><div><br></div></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><div><font face="Menlo">protocol ConstructibleFromValue<b><T></b> {</font></div></div><div><font face="Menlo"> init(_ value: T)</font></div><div><div><font face="Menlo">}</font></div></div></blockquote><div><div><br></div><div>Implicit in this feature is the ability for a given type to conform to the protocol in two different ways. A “Real” type might be constructible from both Float and Double, e.g.,</div><div><br></div></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><div><font face="Menlo">struct Real { … }</font></div></div><div><div><font face="Menlo">extension Real : ConstructibleFrom<Float> {</font></div></div><div><div><font face="Menlo"> init(_ value: Float) { … }</font></div></div><div><div><font face="Menlo">}</font></div></div><div><div><div><font face="Menlo">extension Real : ConstructibleFrom<Double> {</font></div></div></div><div><div><div><font face="Menlo"> init(_ value: Double) { … }</font></div></div></div><div><div><div><font face="Menlo">}</font></div></div></div></blockquote><div><div><br></div><div>Most of the requests for this feature actually want a different feature. They tend to use a parameterized Sequence as an example, e.g.,</div><div><br></div></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><div><font face="Menlo">protocol Sequence<Element> { … }</font></div></div><div><div><font face="Menlo"><br></font></div></div><div><div><font face="Menlo">func foo(strings: Sequence<String>) { /// works on any sequence containing Strings</font></div></div><div><div><font face="Menlo"> // ...</font></div></div><div><div><font face="Menlo">}</font></div></div></blockquote><div><div><br></div><div>The actual requested feature here <span style="white-space:pre-wrap"> </span>is the ability to say “Any type that conforms to Sequence whose Element type is String”, which is covered by the section on “Generalized existentials”, below.</div><div><br></div><div>More importantly, modeling Sequence with generic parameters rather than associated types is tantalizing but wrong: you don’t want a type conforming to Sequence in multiple ways, or (among other things) your for..in loops stop working, and you lose the ability to dynamically cast down to an existential “Sequence” without binding the Element type (again, see “Generalized existentials”). Use cases similar to the ConstructibleFromValue protocol above seem too few to justify the potential for confusion between associated types and generic parameters of protocols; we’re better off not having the latter.</div><div><br></div><div><br></div><div><span style="font-size:14px"><i>Private conformances </i></span></div><div><br></div><div>Right now, a protocol conformance can be no less visible than the minimum of the conforming type’s access and the protocol’s access. Therefore, a public type conforming to a public protocol must provide the conformance publicly. One could imagine removing that restriction, so that one could introduce a private conformance:</div></div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><div><font face="Menlo">public protocol P { }</font></div><div><font face="Menlo">public struct X { }</font></div></div><div><div><font face="Menlo">extension X : <b>internal P</b> { … } // X conforms to P, but only within this module</font></div></div></blockquote><div><div><br></div><div>The main problem with private conformances is the interaction with dynamic casting. If I have this code:</div><div><br></div></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><div><font face="Menlo">func foo(value: Any) {</font></div></div><div><div><font face="Menlo"> if let x = value as? P { print(“P”) }</font></div></div><div><div><font face="Menlo">}</font></div></div><div><div><font face="Menlo"><br></font></div></div><div><div><font face="Menlo">foo(X())</font></div></div></blockquote><div><div><br></div><div>Under what circumstances should it print “P”? If foo() is defined within the same module as the conformance of X to P? If the call is defined within the same module as the conformance of X to P? Never? Either of the first two answers requires significant complications in the dynamic casting infrastructure to take into account the module in which a particular dynamic cast occurred (the first option) or where an existential was formed (the second option), while the third answer breaks the link between the static and dynamic type systems—none of which is an acceptable result.</div><div><br></div><span style="font-size:14px"><i>Conditional conformances via protocol extensions<br></i></span></div><div><br></div><div><div>We often get requests to make a protocol conform to another protocol. This is, effectively, the expansion of the notion of “Conditional conformances” to protocol extensions. For example:</div></div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">protocol P {</font></div><div><font face="Menlo"> func foo()</font></div><div><font face="Menlo">}</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo">protocol Q {</font></div><div><font face="Menlo"> func bar()</font></div><div><font face="Menlo">}</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo">extension <b>Q : P</b> { <i>// every type that conforms to Q also conforms to P</i></font></div><div><font face="Menlo"> func foo() { <i>// implement “foo” requirement in terms of “bar"</i></font></div><div><font face="Menlo"> bar()</font></div><div><font face="Menlo"> }</font></div><div><font face="Menlo">}</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo">func f<T: P>(t: T) { … }</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo">struct X : Q {</font></div><div><font face="Menlo"> func bar() { … }</font></div><div><font face="Menlo">}</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo">f(X()) // okay: X conforms to P through the conformance of Q to P</font></div></blockquote><div><br></div><div>This is an extremely powerful feature: is allows one to map the abstractions of one domain into another domain (e.g., every Matrix is a Graph). However, similar to private conformances, it puts a major burden on the dynamic-casting runtime to chase down arbitrarily long and potentially cyclic chains of conformances, which makes efficient implementation nearly impossible.</div><div><br></div><div><div><b><font size="4">Potential removals</font></b></div><div><b><br></b></div><div>The generics system doesn’t seem like a good candidate for a reduction in scope; most of its features do get used fairly pervasively in the standard library, and few feel overly anachronistic. However...</div><div><br></div><div><span style="font-size:14px"><i>Associated type inference</i></span></div></div><div><br></div><div>Associated type inference is the process by which we infer the type bindings for associated types from other requirements. For example:</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">protocol IteratorProtocol {</font></div><div><font face="Menlo"> associatedtype Element</font></div><div><font face="Menlo"> mutating func next() -> Element?</font></div><div><font face="Menlo">}</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo">struct IntIterator : IteratorProtocol {</font></div><div><font face="Menlo"> mutating func next() -> Int? { … } // use this to infer Element = Int</font></div><div><font face="Menlo">}</font></div></blockquote><div><br></div><div>Associated type inference is a useful feature. It’s used throughout the standard library, and it helps keep associated types less visible to types that simply want to conform to a protocol. On the other hand, associated type inference is the only place in Swift where we have a <b>global</b> type inference problem: it has historically been a major source of bugs, and implementing it fully and correctly requires a drastically different architecture to the type checker. Is the value of this feature worth keeping global type inference in the Swift language, when we have deliberatively avoided global type inference elsewhere in the language?</div><div><br></div><div><br></div><div><b><font size="4">Existentials</font></b><br><br>Existentials aren’t really generics per se, but the two systems are closely intertwined due to their mutable dependence on protocols.<br><br><span style="font-size:14px"><i>*Generalized existentials</i></span></div><div><br></div><div>The restrictions on existential types came from an implementation limitation, but it is reasonable to allow a value of protocol type even when the protocol has Self constraints or associated types. For example, consider IteratorProtocol again and how it could be used as an existential:</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><div><font face="Menlo">protocol IteratorProtocol {</font></div></div><div><div><font face="Menlo"> associatedtype Element</font></div></div><div><div><font face="Menlo"> mutating func next() -> Element?</font></div></div><div><div><font face="Menlo">}</font></div></div><div><font face="Menlo"><br></font></div><div><font face="Menlo">let it: IteratorProtocol = …</font></div><div><font face="Menlo">it.next() // if this is permitted, it could return an “Any?”, i.e., the existential that wraps the actual element</font></div></blockquote><div><font face="Menlo"><br></font></div><div>Additionally, it is reasonable to want to constrain the associated types of an existential, e.g., “a Sequence whose element type is String” could be expressed by putting a where clause into “protocol<…>” or “Any<…>” (per “Renaming protocol<…> to Any<…>”):</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">let strings: Any<Sequence<b> where .Iterator.Element == String</b>> = [“a”, “b”, “c”]</font></div></blockquote><div><br></div><div>The leading “.” indicates that we’re talking about the dynamic type, i.e., the “Self” type that’s conforming to the Sequence protocol. There’s no reason why we cannot support arbitrary “where” clauses within the “Any<…>”. This very-general syntax is a bit unwieldy, but common cases can easily be wrapped up in a generic typealias (see the section “Generic typealiases” above):</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">typealias AnySequence<Element> = <b>Any<Sequence where .Iterator.Element == Element></b></font></div><div><font face="Menlo">let strings: AnySequence<String> = [“a”, “b”, “c”]</font></div></blockquote><div><br></div><div><br></div><div><i style="font-size:14px">Opening existentials</i></div><div><br></div><div>Generalized existentials as described above will still have trouble with protocol requirements that involve Self or associated types in function parameters. For example, let’s try to use Equatable as an existential:</div><div><br></div><blockquote style="margin:0 0 0 40px;border:none;padding:0px"><div><font face="Menlo">protocol Equatable {</font></div><div><font face="Menlo"> func ==(lhs: Self, rhs: Self) -> Bool</font></div><div><font face="Menlo"> func !=(lhs: Self, rhs: Self) -> Bool</font></div><div><font face="Menlo">}</font></div><div><font face="Menlo"><br></font></div><div><font face="Menlo">let e1: Equatable = …</font></div><div><font face="Menlo">let e2: Equatable = …</font></div><div><font face="Menlo">if e1 == e2 { … } <i>// <b>error</b>:</i> e1 and e2 don’t necessarily have the same dynamic type</font></div></blockquote><div><br></div><div>One explicit way to allow such operations in a type-safe manner is to introduce an “open existential” operation of some sort, which extracts and gives a name to the dynamic type stored inside an existential. For example:</div><div><br></div><div><span style="white-space:pre-wrap"> </span> </div><div><blockquote style="margin:0px 0px 0px 40px;border:none;padding:0px"><font face="Menlo">if let storedInE1 = e1 openas T { // T is a the type of storedInE1, a copy of the value stored in e1</font></blockquote><blockquote style="margin:0px 0px 0px 40px;border:none;padding:0px"><font face="Menlo"> if let storedInE2 = e2 as? T { // is e2 also a T?</font></blockquote><blockquote style="margin:0px 0px 0px 40px;border:none;padding:0px"><font face="Menlo"> if storedInE1 == storedInE2 { … } // okay: storedInT1 and storedInE2 are both of type T, which we know is Equatable</font></blockquote><blockquote style="margin:0px 0px 0px 40px;border:none;padding:0px"><font face="Menlo"> }</font></blockquote><blockquote style="margin:0px 0px 0px 40px;border:none;padding:0px"><font face="Menlo">}</font></blockquote></div><div><br></div><div>Thoughts?</div><div><br></div></div></div><div><span style="white-space:pre-wrap"> </span>- Doug</div><div><br></div></div><span class="">_______________________________________________<br>swift-evolution mailing list<br><a href="mailto:swift-evolution@swift.org" target="_blank">swift-evolution@swift.org</a><br><a href="https://lists.swift.org/mailman/listinfo/swift-evolution" target="_blank">https://lists.swift.org/mailman/listinfo/swift-evolution</a><br></span></div></blockquote></div><br></div><br>_______________________________________________<br>
swift-evolution mailing list<br>
<a href="mailto:swift-evolution@swift.org">swift-evolution@swift.org</a><br>
<a href="https://lists.swift.org/mailman/listinfo/swift-evolution" rel="noreferrer" target="_blank">https://lists.swift.org/mailman/listinfo/swift-evolution</a><br>
<br></blockquote></div><br></div>