Golang 將改變常見的 closure 地雷

在「Fixing for loops in Go 1.22 (go.dev)」看到的，原文在「Fixing For Loops in Go 1.22」這邊。

算是各種有 closure 的程式語言會遇到的經典問題，下面的 Golang 程式在撰寫時會預期輸出 a、b、c (不保證順序)，但實際上會輸出 c、c、c，因為變數的 scope 設計原因：

func main() {
    done := make(chan bool)

    values := []string{"a", "b", "c"}
    for _, v := range values {
        go func() {
            fmt.Println(v)
            done <- true
        }()
    }

    // wait for all goroutines to complete before exiting
    for _ = range values {
        <-done
    }
}

這在 JavaScript 也是個經典的問題，搜「javascript closure loop」就可以看到「JavaScript closure inside loops – simple practical example」這個例子，對應的答案則是有提到 ES6 的 let 因為改變了變數的 scope，可以用來解決這個問題。

剛好 let 也是 best practice 了，所以現在 JavaScript 這邊遇到這個問題的情況會少一些。

回到 Golang 這邊，Golang 是打算改變在 for loop 時，對應變數的 scope 來解決：

For Go 1.22, we plan to change for loops to make these variables have per-iteration scope instead of per-loop scope. This change will fix the examples above, so that they are no longer buggy Go programs; it will end the production problems caused by such mistakes; and it will remove the need for imprecise tools that prompt users to make unnecessary changes to their code.

另外因為這是 breaking compatibility 的改變 (畢竟有些程式碼是清楚知道這個特性而刻意這樣寫的)，所以會有對應的措施，只有在有指定 Golang 1.22 或是更新的版本才會用新的 scope 編譯：

To ensure backwards compatibility with existing code, the new semantics will only apply in packages contained in modules that declare go 1.22 or later in their go.mod files. This per-module decision provides developer control of a gradual update to the new semantics throughout a codebase. It is also possible to use //go:build lines to control the decision on a per-file basis.