英文:
Trouble responding to multiple requests at the same time
问题
我正在尝试将我的node.js HTTP服务器转换为Go。这是我想要发生的事情：
我有一个资源，它会间歇性地生成（大约每秒一次），我希望所有对该资源的请求都等待，直到下一次生成。这样，客户端可以轮询并确保只获取最新的资源。我正在使用web.go来简化运行HTTP服务器的复杂性。
这是我代码的简要版本：
package main
import (
"time"
"web"
"fmt"
vector "container/vector"
)
var listeners vector.Vector;
func resource (ctx *web.Context) {
c := make(chan int)
listeners.Push(c)
go func() {
<-c
go func() {
ctx.WriteString("Cool")
fmt.Println("Wrote something")
}()
}()
}
func resourceLoop() {
time.Sleep(5 * 1000 * 1000000) // 睡眠5秒
for ; listeners.Len() > 0 ; {
c := listeners.Pop().(chan int)
c <- 1
}
go resourceLoop()
}
func main() {
web.Get("/resource", resource)
go resourceLoop()
web.Run("localhost:4000")
}
我希望有一个Context.End()或类似的函数，但似乎没有。我阅读了web.go的源代码，但无法找出它在哪里结束响应（web.go:268是我的resource()被调用的地方）。在node.js中，这很简单，你可以调用ServerResponse.end()。
当我在Chrome中运行脚本时关闭服务器，我得到以下输出（似乎是正确的，除了响应没有结束）：
4
Cool
HTTP/1.1 200 OK
Content-Type: text/html; charset=utf-8
Date: Sat, 16 Apr 2011 00:37:58 GMT
Server: web.go
Transfer-Encoding: chunked
0
这是web.go框架的问题还是我做错了什么？如果是框架的问题，我会向开发者提交一个问题。
我对Go还不太熟悉，所以我可能完全弄错了。
英文:
I am trying to convert my node.js HTTP server to Go. Here's what I want to happen:
I have a resource that gets generated intermittently (say every second or so), and I want all requests for this resource to wait until the next time it is generated. This way clients can poll and be guaranteed to get only the up-to-date resource. I am using web.go to remove a lot of the complexity of running an HTTP server.
Here is a brief version of my code:
package main
import (
"time"
"web"
"fmt"
vector "container/vector"
)
var listeners vector.Vector;
func resource (ctx *web.Context) {
c := make(chan int)
listeners.Push(c)
go func() {
<-c
go func() {
ctx.WriteString("Cool")
fmt.Println("Wrote something")
}()
}()
}
func resourceLoop() {
time.Sleep(5 * 1000 * 1000000) // sleep for 5 seconds
for ; listeners.Len() > 0 ; {
c := listeners.Pop().(chan int)
c <- 1
}
go resourceLoop()
}
func main() {
web.Get("/resource", resource)
go resourceLoop()
web.Run("localhost:4000")
}
I would expect there to be a Context.End() or similar function, but it doesn't seem to have one. I read the source for web.go, but I couldn't figure out where it was ending the response (web.go:268 is where my resource() is called). In node.js this is trivial, you can call a ServerResponse.end().
When I kill the server while running the script in Chrome, I get this output (seems to be correct, except that the response isn't ending):
4
Cool
HTTP/1.1 200 OK
Content-Type: text/html; charset=utf-8
Date: Sat, 16 Apr 2011 00:37:58 GMT
Server: web.go
Transfer-Encoding: chunked
0
Is this a problem with the web.go framework or am I doing something wrong? If it's a problem with the framework, I'll file an issue with him.
I'm pretty new to Go, so I could be going about this completely wrong.
答案1
得分: 3
我从未使用过web.go，但是你的示例似乎过于复杂。为什么需要一个goroutine来生成另一个goroutine？我会认为框架本身会处理并发，只需编写如下代码：
func resource(ctx *web.Context, val string) string {
c := make(chan int)
listeners.Push(c)
<-c
return "Cool"
}
否则，它看起来正是你想要的，如果你真的完成了连接，只需关闭连接：
ctx.Close()
英文:
I've never used web.go, but it seems that your example is entirely too complicated. Why do you need a goroutine to spawn a goroutine? I would assume the framework itself would take care of concurrency and just write this:
func resource (ctx *web.Context, val string) string {
c := make(chan int)
listeners.Push(c)
<-c
return "Cool"
}
Otherwise, it looks like it is doing exactly what you want and you just need to close the connection if you're truly done with it:
ctx.Close()
通过集体智慧和协作来改善编程学习和解决问题的方式。致力于成为全球开发者共同参与的知识库，让每个人都能够通过互相帮助和分享经验来进步。
评论