advhttp

advhttp

Advhttp is a go utility library to help out with some common http needs that aren't in the standard library. The library can be broken up into the following functionality categories.

• Handlers
• Logging
• OAuth2
• Cross Origin Resource Sharing
• Reverse Proxy

Handlers

The included handlers try to build a handler stack to alleviate boilerplate code in go server applications. A common setup might look like:

ch := advhttp.NewDefaultCorsHandler(http.DefaultServeMux)
ph := advhttp.NewPanicRecoveryHandler(ch)
lh := advhttp.NewLoggingHandler(ph, os.Stdout)

http.ListenAndServe(":http", lh)

This creates a request pipeline that will eventually end up using the default serve mux provided in the http library. But you'll have Common Log Format logs printed to std out, panic recovery, and cors headers to allow cross origin resouce sharing.

Logging

You can also bypass the handlers and get a little bit more out of the library on your own. I provide a custom ResponseWriter that tracks bytes written and status code so that you can have those after you write. These are typically included in logs.

trw := advhttp.NewResponseWriter(w)
fmt.Fprintf(os.Stderr, "Status: %v\n", trw.Status())
fmt.Fprintf(os.Stderr, "Bytes Written: %v\n", trw.Length())
fmt.Fprint(os.Stdout, trw.LogCommonExtended(r))

OAuth2

The OAuth2 portion of the library is primarily for applications which may need to communicate with other apis via oauth2. The library will cache a token so that it may be re-used accross multiple calls and reduce latency.

tracker, err := advhttp.NewClientCredentialsTokenTracker(tokenEndpoint, tokenInfoEndpoint, client_id, client_secret, scope)
token, err := tracker.GetToken() //Get a new token or a cached token
token, err := tracker.GetNewToken() //Force getting a new token
ti, err := tracker.GetTokenInfo() //Return an object with the result from the tokenInfoEndpoint

//You can also just use the apis yourself
token, tokenExpires, refreshToken, err := advhttp.GetPasswordToken(tokenEndpoint, client_id, client_secret, username, password, scope)
token, tokenExpires, err := advhttp.GetRefreshToken(tokenEndpoint, client_id, client_secret, refresh_token, scope)
token, tokenExpires, err := advhttp.GetClientCredentialsToken(tokenEndpoint, client_id, client_secret string, scope []string)
ti, err := advhttp.GetTokenInformation(tokenInfoEndpoint, token string)

Cross Origin Resource Sharing

The cors library allows you to set up cors how you'd like, and then serve based on that. advhttp comes with a DefaultCors object that contains the default cors settings. advhttp.DefaultCors. A cors object includes:

• AllowOrigin (string) Default "" meaning to mirror the Origin header
• AllowHeaders ([]string) Default advhttp.CorsDefaultAllowHeaders
• AllowMethods ([]string) Default advhttp.CorsDefaultAllowMethods
• ExposeHeaders ([]string) Default advhttp.CorsDefaultExposeHeaders
• MaxAge (int64) Default advhttp.CorsDefaultMaxAge
• AllowCredentials (bool) Default advhttp.CorsDefaultAllowCredentials

You can use the default cors handler:

advhttp.ProcessCors(w,r)

Or you can create your own:

cors := new(Cors)
cors.AllowOrigin = "*"
cors.AllowHeaders = []string{"Location", "Content-Type", "ETag", "Accept-Patch"}
cors.AllowMethods = []string{"OPTIONS", "HEAD", "GET", "POST", "PUT", "PATCH", "DELETE"}
cors.ExposeHeaders = []string{"Location", "Content-Type", "ETag", "Accept-Patch"}
cors.MaxAge = int64(1728000)
cors.AllowCredentials = true
cors.ProcessCors(w,r)

Or you can use the included handler:

ch := advhttp.NewDefaultCorsHandler(http.DefaultServeMux)
ch.ServeHTTP(w,r)

Reverse Proxy

The reverse proxy adds to the httputils and allows you to reverse proxy to different hosts, mutate the path on it's way there, (and also in the Location header in responses). It is helpful in gateway projects.

googleURL, _ := url.Parse("http://www.google.com/")
downloads := advhttp.NewGatewayReverseProxy(googleURL, true, "/google/")
http.Handle("/google/", someHandlerFunc)

In the above example, calls to the /google/ endpoint on your server http://yourserver.com/google/ will result in your server then calling google http://www.google.com/ but stripping off the /google/ path from the original url. If the response from google is a redirect or includes the Location header, then advhttp will rewrite the header to include the original /google/ path. This results in browser being able to (almost) transparently use the other server through your server.