NS caching has been discussed multiple times in the past. The consensus seems to be that Go won't go there: github.com/golang/go/issues/24796
I've seen a few DNS caching solutions for Go (one, two), however, I haven't seen any implementations that allow replacing the net.DefaultResolver?
Package github.com/artyom/dot got me thinking if I could do the same for DNS caching, and also DNS over HTTPS.
So github.com/ncruces/go-dns is my attempt. Replacing net.DefaultResolver with a caching DNS…
I feel like it's a dummy question, but I want to ask it anyway since I do not have much experience with concurrent programming.
In the app I am currently working on, someone wrote a code that should do something (in this case, printing "TICK!") every, let's say 5 minutes:
go func() {
for {
fmt.Println("TICK!")
time.Sleep(5 * time.Minute)
}
}()
However, from what I have read, the recommended approach is:
func CallTickerInBackground(done <-chan bool) {
ticker := time.NewTicker(5 * time.Minute)
go func() {
for {
select {
case <-done:
logrus.Info("Stopping ticker")
ticker.Stop()
return
case <-ticker.C:
fmt.Println("TICK!")
}
}
}()
}
The advantages of the latter are that we have more stable time…
最近做一个小型项目，因为rabbitMQ等专业软件比较重，故团队决定采用redis实现一个轻量的队列。
在这篇文章中，你可以获得：
redigo包的基本用法context包的作用select、channel和goroutine)的利用最终代码量大概265行左右。
redis并不是被设计用来做队列的，事实上它并不是那么适合作为队列载体——官方也不推荐用来做队列，甚至因为使用redis做队列的人太多，而促使antirez(redis的作者)开发了另一个名为Disque1的专业队列库，据说将会加入到redis6中。
尽管如此，redis依旧提供了号称Reliable queue的队列指令2。
我们知道，当生产者在另一端生成消息之后，这一端的消费者就要取出这一消息进行消费动作；而在消费的过程中如果出现任何异常——例如“程序崩溃”等问题，造成进程的退出，消息就会丢失。为此，
…redis官方提供了RPOPLPUSH这一队列指令，在从队列中取出消息的同时又塞进另一个队列中。这样当程序发生异常退出时，我们也可以通过第二个队列来找回丢失的消息。