前言

You use optionals in situations where a value may be absent. An optional represents two possibilities: Either there is a value, and you can unwrap the optional to access that value, or there isn’t a value at all.

这是摘录自The Swift Programming Language 书中的一句话。讲了optionals的使用场景，大概就是说：当一个值可能是缺省的时候，要么可以访问这个值，要么不是一个值（为null）

一段简单的代码

demo.swift

var code: Int? = 100
code = 200
var code_value: Int = code!
var code_b: Int?
code_b = 2
var code_c = code

编译一下(带上了编译优化的参数)

swiftc -O  -emit-ir ./demo.swift > demo.ll

demo.ll核心的几行代码

%GSqSi_ = type <{ [8 x i8], [1 x i8] }>
%Si = type <{ i64 }>

@_Tv4demo4codeGSqSi_ = hidden local_unnamed_addr global %GSqSi_ zeroinitializer, align 8
@_Tv4demo10code_valueSi = hidden local_unnamed_addr global %Si zeroinitializer, align 8
@_Tv4demo6code_bGSqSi_ = hidden local_unnamed_addr global %GSqSi_ zeroinitializer, align 8
@_Tv4demo6code_cGSqSi_ = hidden local_unnamed_addr global %GSqSi_ zeroinitializer, align 8
@__swift_reflection_version = linkonce_odr hidden constant i16 1
@llvm.used = appending global [1 x i8*] [i8* bitcast (i16* @__swift_reflection_version to i8*)], section "llvm.metadata"

; Function Attrs: norecurse nounwind
define i32 @main(i32, i8** nocapture readnone) local_unnamed_addr #0 {
entry:
  store i64 200, i64* bitcast (%GSqSi_* @_Tv4demo4codeGSqSi_ to i64*), align 8
  store i1 false, i1* bitcast ([1 x i8]* getelementptr inbounds (%GSqSi_, %GSqSi_* @_Tv4demo4codeGSqSi_, i64 0, i32 1) to i1*), align 8
  store i64 200, i64* getelementptr inbounds (%Si, %Si* @_Tv4demo10code_valueSi, i64 0, i32 0), align 8
  store i64 2, i64* bitcast (%GSqSi_* @_Tv4demo6code_bGSqSi_ to i64*), align 8
  store i1 false, i1* bitcast ([1 x i8]* getelementptr inbounds (%GSqSi_, %GSqSi_* @_Tv4demo6code_bGSqSi_, i64 0, i32 1) to i1*), align 8
  store i64 200, i64* bitcast (%GSqSi_* @_Tv4demo6code_cGSqSi_ to i64*), align 8
  store i1 false, i1* bitcast ([1 x i8]* getelementptr inbounds (%GSqSi_, %GSqSi_* @_Tv4demo6code_cGSqSi_, i64 0, i32 1) to i1*), align 8
  ret i32 0
}

分析

代码很简单声明定义了optional int型的变量

定义了两种结构体:

%GSqSi_ = type <{ [8 x i8], [1 x i8] }> //其实就是一个int64 用来存值 一个int8用来存bool值标识是否值存在
%Si = type <{ i64 }> //为了存一个整数

初始化变量
@_Tv4demo4codeGSqSi_ = hidden local_unnamed_addr global %GSqSi_ zeroinitializer, align 8 //code
@_Tv4demo10code_valueSi = hidden local_unnamed_addr global %Si zeroinitializer, align 8  //code_value
@_Tv4demo6code_bGSqSi_ = hidden local_unnamed_addr global %GSqSi_ zeroinitializer, align 8 //code_b
@_Tv4demo6code_cGSqSi_ = hidden local_unnamed_addr global %GSqSi_ zeroinitializer, align 8 //code_c

可以看出 变量code code_b code_c 都被声明成%GSqSi_的类型
变量 code_value 则是%Si

store i64 200, i64* bitcast (%GSqSi_* @_Tv4demo4codeGSqSi_ to i64*), align 8
store i1 false, i1* bitcast ([1 x i8]* getelementptr inbounds (%GSqSi_, %GSqSi_* @_Tv4demo4codeGSqSi_, i64 0, i32 1) to i1*), align 8

很典型的赋值 value为200 同时值为空的标识设置为false 若是空值则为true

再来一段代码

code:Int? = 200
code = nil

define i32 @main(i32, i8** nocapture readnone) local_unnamed_addr #0 {
entry:
  store i64 0, i64* bitcast (%GSqSi_* @_Tv5demo14codeGSqSi_ to i64*), align 8
  store i1 true, i1* bitcast ([1 x i8]* getelementptr inbounds (%GSqSi_, %GSqSi_* @_Tv5demo14codeGSqSi_, i64 0, i32 1) to i1*), align 8
  ret i32 0
}

可以看到 code为nil的时候就是为空标识设置为true了

小结:

由此可得optional本质其实就是用了一个结构体包装了它修饰的类型以及一个为空标识符当值存在则标识为false 不存在则标识为true