Any mistakes in below code? Multiple directory serving is not working from the below code. When I access the localhost:9090/ide, the server will return 404 error.
package main
import (
"log"
"net/http"
)
func serveIDE(w http.ResponseWriter, r *http.Request) {
http.FileServer(http.Dir("/home/user/ide")).ServeHTTP(w, r)
}
func serveConsole(w http.ResponseWriter, r *http.Request) {
http.FileServer(http.Dir("/home/user/console")).ServeHTTP(w, r)
}
func main() {
http.HandleFunc("/ide", serveIDE)
http.HandleFunc("/console", serveConsole)
err := http.ListenAndServe(":9090", nil)
if err != nil {
log.Fatal("ListenAndServe: ", err)
}
}
When I change the code like this,
http.HandleFunc("/", serveIDE)
It will work as I expected.
One of the issues with using http.FileServer is that the request path is used to build the file name, so if you're serving from anywhere but the root you need to strip the route prefix to that handler.
The standard library includes a helpful tool for that http.StripPrefix, but that only works on http.Handlers, not http.HandleFuncs, so to use it you need to adapt your HandleFunc to a Handler.
Here is a working version that should do what you want. Note that wHandler is just an adapter from your HttpFunc methods to Hander interface:
package main
import (
"log"
"net/http"
)
func serveIDE(w http.ResponseWriter, r *http.Request) {
http.FileServer(http.Dir("/home/user/ide")).ServeHTTP(w, r)
}
func serveConsole(w http.ResponseWriter, r *http.Request) {
http.FileServer(http.Dir("/home/user/console")).ServeHTTP(w, r)
}
type wHandler struct {
fn http.HandlerFunc
}
func (h *wHandler) ServeHTTP(w http.ResponseWriter, r *http.Request) {
log.Printf("Handle request: %s %s", r.Method, r.RequestURI)
defer log.Printf("Done with request: %s %s", r.Method, r.RequestURI)
h.fn(w, r)
}
func main() {
http.Handle("/ide", http.StripPrefix("/ide", &wHandler{fn: serveIDE}))
http.Handle("/console", http.StripPrefix("/console", &wHandler{fn: serveConsole}))
err := http.ListenAndServe(":9090", nil)
if err != nil {
log.Fatal("ListenAndServe: ", err)
}
}
Related
What’s the equivalent to middleware handlers in Google Cloud Functions?
In standard approach, normally I do:
router.Handle("/receive", middlewares.ErrorHandler(MyReceiveHandler))
And then, in the middleware:
type ErrorHandler func(http.ResponseWriter, *http.Request) error
func (fn ErrorHandler) ServeHTTP(w http.ResponseWriter, r *http.Request) {
err := fn(w, r)
if err == nil {
return
}
log.Printf("An error accured: %v", err)
clientError, ok := err.(errors.BaseError)
if !ok {
w.WriteHeader(500)
return
}
w.WriteHeader(clientError.GetStatusCode())
w.Write([]byte(clientError.Error()))
}
In AWS Lambda, I can achieve the same thing using, for example:
func main() {
lambda.Start(
middlewares.Authentication(Handler),
)
}
But I could not find a way to do this in GCP Functions.
How would it work?
Can you help me?
Let's say you start with the following server code in your development environment:
package main
import (
"fmt"
"log"
"net/http"
)
func main() {
http.Handle("/", MiddlewareFinalMsg(" Goodbye!", http.HandlerFunc(HelloWorld)))
if err := http.ListenAndServe(":8080", nil); err != nil {
log.Fatal(err)
}
}
func MiddlewareFinalMsg(msg string, h http.Handler) http.Handler {
return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
h.ServeHTTP(w, r)
fmt.Fprint(w, msg)
})
}
func HelloWorld(w http.ResponseWriter, r *http.Request) {
w.Header().Set("Content-Type", "text/plain")
fmt.Fprint(w, "Hello, World!")
}
As far as I can tell, GCF requires its entry point to be an exported identifier of type func(http.ResponseWriter, *http.Request) (not http.HandlerFunc, not http.Handler); therefore, if you have a http.Handler, you'll need to select its ServeHTTP method explicitly to obtain a function of the expected type. However, that identifier can be a package-level function, a method, or a variable.
Here is how you can adapt the code above for GCF:
package p
import (
"fmt"
"net/http"
)
// use F as your GCF's entry point
var F = MiddlewareFinalMsg(" Goodbye!", http.HandlerFunc(HelloWorld)).ServeHTTP
func MiddlewareFinalMsg(msg string, h http.Handler) http.Handler {
return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
h.ServeHTTP(w, r)
fmt.Fprint(w, msg)
})
}
func HelloWorld(w http.ResponseWriter, r *http.Request) {
w.Header().Set("Content-Type", "text/plain")
fmt.Fprint(w, "Hello, World!")
}
I am trying to have a reverse proxy in Golang, but I am unable to stop target ip from showing in browser ie. it is simply redirecting to wikipedia (target) rather than showing doing reverseproxy.
Can anyone tell what I am doing wrong?
package main
import (
"net/http"
"net/http/httputil"
"net/url"
"github.com/gorilla/mux"
)
func main() {
target := "https://www.wikipedia.org"
remote, err := url.Parse(target)
if err != nil {
panic(err)
}
proxy := httputil.NewSingleHostReverseProxy(remote)
r := mux.NewRouter()
r.HandleFunc("/forward/{rest:.*}", handler(remote, proxy))
http.ListenAndServe(":8080", r)
}
func handler(ur *url.URL, p *httputil.ReverseProxy) func(http.ResponseWriter, *http.Request) {
return func(w http.ResponseWriter, r *http.Request) {
r.URL.Host = ur.Host
r.URL.Scheme = ur.Scheme
r.Host = ur.Host
r.URL.Path = mux.Vars(r)["rest"]
p.ServeHTTP(w, r)
}
}
I'm working on a website using Go. The server constraints require that I use CGI. When I test the following code locally using http.ListenAndServe() (commented out below), the various handlers are called correctly depending on the address requested. However, if I use cgi.Serve() instead, the default router is executed for all addresses (i.e., the handler for "/" is always executed). I'd appreciate any clues as to how to fix the issue.
EDIT: Here is the simplest test case I can think of to show the problem:
//=============SIMPLIFIED CODE================//
package main
import (
"fmt"
"net/http"
"net/http/cgi"
)
func defaultHandler(w http.ResponseWriter, r *http.Request) {
fmt.Fprintf(w, "Default")
}
func otherHandler(w http.ResponseWriter, r *http.Request) {
fmt.Fprintf(w, "Other")
}
func main() {
http.HandleFunc("/other", otherHandler)
http.HandleFunc("/", defaultHandler)
/*
//Works fine
err := http.ListenAndServe(":8090", nil)
if err != nil {
panic(err)
}
*/
//Always fires defaultHandler no matter the address requested
err := cgi.Serve(nil)
if err != nil {
panic(err)
}
}
//=============CODE FROM ORIGINAL POST===================//
package main
import (
"fmt"
"net/http"
"net/http/cgi"
"net/url"
"os"
"github.com/go-cas/cas"
)
func logoutHandler(w http.ResponseWriter, r *http.Request) {
cas.RedirectToLogout(w, r)
}
func calendarHandler(w http.ResponseWriter, r *http.Request) {
if !cas.IsAuthenticated(r) {
cas.RedirectToLogin(w, r)
}
fmt.Fprintf(w, "Calendar for %s", cas.Username(r))
}
func defaultHandler(w http.ResponseWriter, r *http.Request) {
if !cas.IsAuthenticated(r) {
cas.RedirectToLogin(w, r)
}
fmt.Fprintf(w, "Hi there %s!", cas.Username(r))
}
func main() {
u, _ := url.Parse("https://www.examplecasserver.com")
client := cas.NewClient(&cas.Options{
URL: u,
})
http.Handle("/logout", client.HandleFunc(logoutHandler))
http.Handle("/calendar", client.HandleFunc(calendarHandler))
http.Handle("/", client.HandleFunc(defaultHandler))
/*
err := http.ListenAndServe(":8080", nil)
if err != nil {
log.Fatal("ListenAndServe: ", err)
}
*/
err := cgi.Serve(nil)
if err != nil {
panic(err)
}
}
The CGI program expects some variables to be set in order to build the request.
Probably there is some issue with the configuration of your web server in which the variables are either not set correctly or not named correctly.
To verify this:
1) Add this before calling cgi.Serve and you'll see how the right handler is called (otherHandler)
os.Setenv("REQUEST_METHOD", "get")
os.Setenv("SERVER_PROTOCOL", "HTTP/1.1")
os.Setenv("SCRIPT_NAME", "/other")
2) Add this at the beginning of the main to check how the variables are being set by the web server:
fmt.Println(os.Environ())
In that output, look for the CGI meta variables defined in the spec:
http://www.ietf.org/rfc/rfc3875
Look for the section "Request Meta-Variables" in that page, you are probably looking for the SCRIPT_NAME or PATH_INFO variables.
EDIT
From the variable values you pasted below, it seems the issue is that the REQUEST_URI contains an additional path component:
REQUEST_URI=/main.cgi/other
So the easiest fix would be for you to map the routes accordingly:
http.HandleFunc("/main.cgi/other", otherHandler)
http.HandleFunc("/", defaultHandler) // or maybe /main.cgi
I want to create global err handler to send it by email.
package main
import (
"github.com/gorilla/mux"
"log"
"net/http"
)
func main() {
rtr := mux.NewRouter()
rtr.HandleFunc("/", withPanic).Methods("GET")
http.Handle("/", rtr)
log.Println("Listening...")
http.ListenAndServe(":3001", http.DefaultServeMux)
}
func withPanic(w http.ResponseWriter, r *http.Request) {
panic("somewhere here will be panic, but I don't know where exactly")
}
How to make it global. It would be easy if I know where error will occur
if err != nil {
sendMeMail(err)
}
But what to do in cases when I don't know exactly where an error will occur? So I should add a global recoverish handler. But how to do it exactly I don't know.
Update
I added defer recover to beginning of main but it never executes on requesting http://localhost:3001. So panic is not emailed.
package main
import (
"errors"
"fmt"
"github.com/gorilla/mux"
"log"
"net/http"
)
func main() {
defer func() {
if r := recover(); r != nil {
fmt.Println("Recovered in f", r)
// find out exactly what the error was and set err
var err error
switch x := r.(type) {
case string:
err = errors.New(x)
case error:
err = x
default:
err = errors.New("Unknown panic")
}
if err != nil {
// sendMeMail(err)
fmt.Println("sendMeMail")
}
}
}()
rtr := mux.NewRouter()
rtr.HandleFunc("/", withPanic).Methods("GET")
http.Handle("/", rtr)
log.Println("Listening...")
http.ListenAndServe(":3001", http.DefaultServeMux)
}
func withPanic(w http.ResponseWriter, r *http.Request) {
panic("somewhere here will be panic, but I don't know where exactly")
}
You can wrap your handlers in a recovery middleware
package main
import (
"errors"
"github.com/gorilla/mux"
"log"
"net/http"
)
func main() {
m := mux.NewRouter()
m.Handle("/", RecoverWrap(http.HandlerFunc(handler))).Methods("GET")
http.Handle("/", m)
log.Println("Listening...")
http.ListenAndServe(":3001", nil)
}
func handler(w http.ResponseWriter, r *http.Request) {
panic(errors.New("panicing from error"))
}
func RecoverWrap(h http.Handler) http.Handler {
return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
defer func() {
r := recover()
if r != nil {
var err error
switch t := r.(type) {
case string:
err = errors.New(t)
case error:
err = t
default:
err = errors.New("Unknown error")
}
sendMeMail(err)
http.Error(w, err.Error(), http.StatusInternalServerError)
}
}()
h.ServeHTTP(w, r)
})
}
func sendMeMail(err error) {
// send mail
}
You can take a a look at codahale recovery handler or negroni middleware for more details.
I believe that is what the gorilla recovery handler is for
I want to know if code 1 manage internal goutines and can use all the cores of one CPU when the requets will increase(dozens) or if per each handler I have to put the key word go that indicate that the funcion handler will be manage by one gorotine like is show it in code 2, and so can use all the cores of the server.
code 1
package main
import (
"fmt"
"net/http"
)
func HandlerOne(w http.ResponseWriter, req *http.Request) {
fmt.Println("message one")
}
func HandlerTwo(w http.ResponseWriter, req *http.Request) {
fmt.Println("message two")
}
func main() {
http.HandleFunc("/R1", HandlerOne)
http.HandleFunc("/R2", HandlerTwo)
err := http.ListenAndServe(":9998", nil)
if err != nil {
fmt.Printf("Server failed: ", err.Error())
}
}
code 2
package main
import (
"fmt"
"net/http"
)
func HandlerOne(w http.ResponseWriter, req *http.Request) {
fmt.Println("message one")
}
func HandlerTwo(w http.ResponseWriter, req *http.Request) {
fmt.Println("message two")
}
func main() {
go http.HandleFunc("/R1", HandlerOne)
go http.HandleFunc("/R2", HandlerTwo)
err := http.ListenAndServe(":9998", nil)
if err != nil {
fmt.Printf("Server failed: ", err.Error())
}
}
Note: Both run without problems and you be able to test it with
curl -l http://localhost:9998/R1
or
curl -l http://localhost:9998/R2
Version 2 is wrong. Package http handles all this stuff for you.
Just make sure you invoke your program with an appropriate GOMAXPROCS, e.g. like GOMAXPROCS=4 ./main