<div dir="ltr"><div>An even better explanation of my suggestion: instead of <b>sealing</b> <font face="monospace, monospace">T.Type</font>, I suggest to <b>rename</b> it to <font face="monospace, monospace">Type<T></font>, formally turning it into a struct, and then add size and align properties to it.</div><div>And regardless of how you interpret it, I suggest to remove <font face="monospace, monospace">.Type</font> notation.<br></div><div><br></div><div><span style="font-size:12.8px">> In Swift, only class instances and metatypes have unique identities. There is no notion of identity for structs, enums, functions, or tuples.</span></div><div><span style="font-size:12.8px">So <font face="monospace, monospace">Type<T></font> will lose identity. That's fine. They will contain unique integer identifiers (assigned to types during compilation) to do its work.</span></div><div><br></div><div><span style="font-size:12.8px">> </span><span style="font-size:12.8px">There is no special compiler magic needed</span></div><div><span style="font-size:12.8px">Maybe. At least, this behaviour will be needed in construction from type literals:</span></div><div><br></div><div><font face="monospace, monospace">func foo(_ x: Type<BaseClass>)</font></div><div><font face="monospace, monospace">foo(DerivedClass) // ok</font></div><div><font face="monospace, monospace">foo(Int) // compilation error</font></div><div><br></div><div><font face="arial, helvetica, sans-serif">With </font><font face="monospace, monospace">Type<T></font>, I can see reflection coming! For example, <font face="monospace, monospace">Type<T></font> can have the following property (although we shouldn't right now):</div><div><br></div><div><font face="monospace, monospace">var fields: [String: (Type<Any>, (T) -> Any)] { get }</font></div><div><br></div><div>I suggest you to prepare a formal proposal, or I will try tomorrow. I would take the direction on removal of <font face="monospace, monospace">T.Type</font> notation.</div></div><div class="gmail_extra"><br><div class="gmail_quote">2016-07-13 22:48 GMT+03:00 Adrian Zubarev via swift-evolution <span dir="ltr"><<a href="mailto:swift-evolution@swift.org" target="_blank">swift-evolution@swift.org</a>></span>:<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><p>Okay I get it now. You meant the size for a metatype <code>sizeof(T.Type.self)</code>. This is indeed 8 bytes, at least not for optimized <code>Bool</code> (as you showed).</p>
<blockquote>
<p>Internally, Type<u></u> contains an Int, i.e. identifier of a type. For every type, compiler must choose a different identifier<u></u></p>
</blockquote>
<p>We can already implement this with <code>Hashable</code> protocol. </p>
<blockquote>
<p>ObjectIdentifier: A unique identifier for a class instance or metatype.</p>
<p>In Swift, only class instances and metatypes have unique identities. There
is no notion of identity for structs, enums, functions, or tuples.</p>
</blockquote>
<pre><code>// version 1:
public let hashValue: Int = ObjectIdentifier(T.self).hashValue
// version 2 (uses ObjectIdentifier calculation without
// constructing an instance of ObjectIdentifier):
init() {
// calculate the hashValue only once
// I'd rename `.self` to `.metatype`
let rawPointerMetatype = unsafeBitCast(T.self, to: Builtin.RawPointer.self)
self.hashValue = Int(Builtin.ptrtoint_Word(rawPointerMetatype))
}
public let hashValue: Int
</code></pre>
<blockquote>
<p>API of Type<u></u> is defined so that it can only contain identifiers of subtypes of T<u></u></p>
<p>For example, when you get a variable of Type<u></u>, it can correspond to BaseClass or DerivedClass<u></u></p>
</blockquote>
<p>I did a quick test and I feel like this falls under the part of tweaking dynamic casts to work with <code>Type<T></code>. There is no special compiler magic needed, <code>unsafeBitCast</code> should do the trick. Or we should teach dynamic casts to work with the inner type <code>T</code> instead of the whole <code>Type<T></code>.</p>
<pre><code>public struct Type<T> : Hashable {
public let metatype: T.Type = T.self
public let hashValue: Int = ObjectIdentifier(T.self).hashValue
}
public func ==<T, U>(lhs: Type<T>, rhs: Type<U>) -> Bool {
return lhs.hashValue == rhs.hashValue
}
public func asOptionalCast<U, T>(type: Type<T>) -> Type<U>? {
guard (type.metatype as? U.Type) != nil else {
return nil
}
return unsafeBitCast(type, to: Type<U>.self)
}
class A {}
class B: A {}
let typeB = Type<B>()
// downcast Type<B> to Type<A>
let downcast: Type<A>? = asOptionalCast(type: typeB)
(downcast! == typeB) == true
// cast Type<A> (which here is Type<B>) back to Type<B>
let upcast: Type<B>? = asOptionalCast(type: downcast!)
(upcast! == Type<B>()) == true
</code></pre>
<p>The good part here that the hash value of the casted type won’t change and testing against a new instance of the same dynamic type will be always true.</p>
<p></p></div><div><span class=""><div style="font-family:Helvetica,Arial;font-size:13px;color:rgba(0,0,0,1.0);margin:0px;line-height:auto"><br></div> <br> <div><div style="font-family:helvetica,arial;font-size:13px">-- <br>Adrian Zubarev<br>Sent with Airmail</div></div> <br></span><div><div class="h5"><p>Am 13. Juli 2016 um 20:31:22, Anton Zhilin (<a href="mailto:antonyzhilin@gmail.com" target="_blank">antonyzhilin@gmail.com</a>) schrieb:</p> <blockquote type="cite"><span><div><div></div><div>
<div dir="ltr">
<div class="gmail_extra">
<div class="gmail_quote">2016-07-13 20:19 GMT+03:00 Adrian Zubarev
via swift-evolution <span dir="ltr"><<a href="mailto:swift-evolution@swift.org" target="_blank">swift-evolution@swift.org</a>></span>:
<blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;border-left-color:rgb(204,204,204);padding-left:1ex">
<div style="word-wrap:break-word">
<div>
<p><span>Am 13. Juli 2016 um 18:30:53, Anton Zhilin
(<a href="mailto:antonyzhilin@gmail.com" target="_blank">antonyzhilin@gmail.com</a>) schrieb:</span></p>
<div>
<blockquote type="cite" style="margin:15px 0px;font-family:Helvetica,Arial;font-size:13px;font-style:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px">
<div>
<div class="gmail_quote" style="color:rgb(0,0,0);font-family:"helvetica Neue",helvetica;font-size:14px;font-style:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px">
<div><span>I see model of<span> </span><font face="monospace, monospace">Type<T></font><span> </span>as
follows:</span></div>
<div>
<ol style="margin:15px 0px">
<li style="margin:15px 0px">Values of<span> </span><font face="monospace, monospace">Type<T></font><span> </span>are
identifiers of types (8 bytes, I guess)<br></li>
<li style="margin:15px 0px">All used identifiers are contained
in<span> </span><font face="monospace, monospace">Type<Any></font><br></li>
<li style="margin:15px 0px"><font face="monospace, monospace">Type<T></font><span> </span>contains
a subset of those identifiers</li>
</ol>
</div>
</div>
</div>
</blockquote>
</div>
<p>I can’t follow your though here. Is this a totally new
behavior?</p>
</div>
</div>
</blockquote>
<div>
<div>That's how it works right now:</div>
<div><font face="monospace, monospace"><br></font></div>
<div><font face="monospace, monospace">sizeofValue(Bool.self)
//=> 0
(optimized away)<br></font></div>
<div><font face="monospace, monospace">sizeofValue(Bool.self as
Any.self) //=> 8</font><br></div>
<div><br></div>
<div>I'll put my thoughts another way:</div>
</div>
<div>
<ul>
<li>Internally, <font face="monospace, monospace">Type<T></font> contains an <font face="monospace, monospace">Int</font>, i.e. identifer of a type. For
every type, compiler must choose a different identifier</li>
<li>API of <font face="monospace, monospace">Type<T></font>
is defined so that it can only contain identifiers of subtypes of
<font face="monospace, monospace">T</font></li>
<li>For example, when you get a variable of <font face="monospace, monospace">Type<BaseClass></font>, it can
correspond to <font face="monospace, monospace">BaseClass</font><font face="arial, helvetica, sans-serif"> </font>or<font face="arial, helvetica, sans-serif"> </font><font face="monospace, monospace">DerivedClass</font></li>
</ul>
</div>
<blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;border-left-color:rgb(204,204,204);padding-left:1ex">
<div style="word-wrap:break-word">
<div>
<div>
<div>
<div>
<blockquote type="cite" style="margin:15px 0px;font-family:Helvetica,Arial;font-size:13px;font-style:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px">
<div>
<div class="gmail_quote" style="color:rgb(0,0,0);font-family:"helvetica Neue",helvetica;font-size:14px;font-style:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px">
<div>
<ol style="margin:15px 0px">
<li style="margin:15px 0px"><span>Upcasting uses
constructor<span> </span><font face="monospace, monospace">init<U: T>(upcasting:
Type<U>)</font></span></li>
</ol>
</div>
</div>
</div>
</blockquote>
</div>
<p>It does look neat but I can’t implement it. At least I have no
clue how to solve the problem that `T` is not a protocol or a class
type.</p>
</div>
</div>
</div>
</div>
</blockquote>
<div>We should add implicit convertion of <font face="monospace, monospace">Type<U></font> to <font face="monospace, monospace">Type<T></font><font face="arial, helvetica, sans-serif">, dulicating current behaviour
of</font> <font face="monospace, monospace">`as
T.Type`</font><font face="arial, helvetica, sans-serif">.</font></div>
<div><font face="arial, helvetica, sans-serif">That constructor
still can be useful, but it will be failable and not that
significant in practise.</font></div>
<div><font face="arial, helvetica, sans-serif">I don't like adding
compiler magic, but I guess we really should in this case, because
otherwise</font> <font face="monospace, monospace">Type<T></font> <font face="arial, helvetica, sans-serif">can't replace</font> <font face="monospace, monospace">T.Type</font><font face="arial, helvetica, sans-serif">.</font></div>
<blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;border-left-color:rgb(204,204,204);padding-left:1ex">
<div style="word-wrap:break-word">
<div>
<div>
<div>
<div>
<div>
<div>
<blockquote type="cite" style="margin:15px 0px;font-family:Helvetica,Arial;font-size:13px;font-style:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px">
<div>
<div class="gmail_quote" style="color:rgb(0,0,0);font-family:"helvetica Neue",helvetica;font-size:14px;font-style:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px">
<div>
<ol style="margin:15px 0px">
<li style="margin:15px 0px"><span>Size
of<span> </span><font face="monospace, monospace">Type<Type<T>></font><span> </span>is
8, static size of<span> </span><font face="monospace, monospace">Type<SomeProtocol></font><span> </span>is
0</span></li>
</ol>
</div>
</div>
</div>
</blockquote>
</div>
<p>I feel like you mean something different here than this:</p>
<p>```swift</p>
<p style="margin-top:0px;margin-bottom:0px;font-size:11px;line-height:normal;font-family:Menlo">
<span>protocol</span> <span>SomeProtocol {}</span></p>
<p style="margin-top:0px;margin-bottom:0px;font-size:11px;line-height:normal;font-family:Menlo;min-height:13px">
<br></p>
<p style="margin-top:0px;margin-bottom:0px;font-size:11px;line-height:normal;font-family:Menlo">
<span>sizeof</span><span>(</span><span>SomeProtocol</span><span>.</span><span>self</span><span>)
==</span> <span>40</span></p>
<p style="margin-top:0px;margin-bottom:0px;font-size:11px;line-height:normal;font-family:Menlo">
<span>```</span></p>
</div>
</div>
</div>
</div>
</div>
</div>
</blockquote>
<div><br></div>
<div>That was a mistake. Static size of <font face="monospace, monospace">Type<SomeProtocol></font> will be 40,
and size property of its values can be less or greater than 40,
depending on internal type identifier.<br></div>
</div>
<br></div>
</div>
</div></div></span></blockquote></div></div></div><div><p></p></div></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>