Prevent having to manually kill main.go after running dredd

[kevgilmore]

Is your feature request related to a problem? Please describe.
I keep having to kill the previous go server in system monitor (linux) before re-running dredd test.

Describe the solution you'd like
Auto kill the go server after tests have completed

Describe alternatives you've considered
Manually killing main.go from system monitor

Additional context
Sorry for making an issue, but I can't find the answer in the docs.
My project code is here https://github.com/hex6b/go-heroku

[abtris]

Your code have to support signals to be killed by dredd as normal application. Here is how to do it - https://stackoverflow.com/questions/18106749/golang-catch-signals

Related part in docs

[kevgilmore]

hey, still not able to get it to work after implementing the signals catch solution.

Here's my code

package main

import (
	"fmt"
	"log"
	"net/http"
	"os"
	"os/signal"
	"syscall"
)

func main() {
	port := os.Getenv("PORT")
	fmt.Printf("Started\n")

	if port == "" {
		port = "8080"
	}

	go func() {
		http.HandleFunc("/", HelloServer)
		if err := http.ListenAndServe(":"+port, nil); err != nil {
			log.Fatal(err)
		}
	}()

	signalChannel := make(chan os.Signal, 2)
	signal.Notify(signalChannel, os.Interrupt, syscall.SIGTERM)
	for {
		sig := <-signalChannel
		switch sig {
		case os.Interrupt:
			fmt.Println("sigint")
		case syscall.SIGTERM:
			fmt.Println("sigterm")
                        os.Exit(1)
			return
		}
	}
}


func HelloServer(w http.ResponseWriter, r *http.Request) {
	w.Header().Set("Content-Type", "application/json")
	fmt.Fprintf(w, "{\"message\":\"It works!\"}")
}

and my dredd.yml

color: true
dry-run: null
hookfiles: null
language: nodejs
require: null
server: go run main.go
server-wait: 3
init: false
custom: {}
names: false
only: []
reporter: []
output: []
header: []
sorted: false
user: null
inline-errors: false
details: false
method: []
loglevel: warning
path: []
hooks-worker-timeout: 5000
hooks-worker-connect-timeout: 1500
hooks-worker-connect-retry: 500
hooks-worker-after-connect-wait: 100
hooks-worker-term-timeout: 5000
hooks-worker-term-retry: 500
hooks-worker-handler-host: 127.0.0.1
hooks-worker-handler-port: 61321
config: ./dredd.yml
blueprint: apiary.apib
endpoint: 'http://127.0.0.1:8080'

From what I can see dredd does not seem to trigger SIGTERM after the test have finished. If I trigger SIGTERM another way the code above works.

Any ideas?

[w-vi]

@hex6b I suspect that the SIGTERM is lost somewhere when doing go run main.go can you instead compile the server and use the binary directly instead of the go run main.go? It can be the case that the SIGTERM is consudem by go run process instead of your sever.

[abtris]

The commit refers to golang/go#23716 that explains that this behaviour is not a bug actually. To fully understand the issue I’ve dived into golang repo.

1. go run command kicks off work.Builder with action &work.Action{Mode: "go run", Func: buildRunProgram, ...} inside run.go file.
2. Next the function buildRunProgram gets called and starts the compiled binary. It also sets up the signal handling using the channel Interrupted .
3. When that channel is closed go run process is always exited with 1 status.

Cite from https://medium.com/@ns3777k/go-run-and-exit-codes-7ad1175821c1