How to use regular expression matching URL, which does decide to use the corresponding function processing
package main
import(
"fmt"
"net/http"
)
func main() {
http.HandleFunc("/pattern", resolve)
http.ListenAndServe(":8080", nil)
}
func resolve(w http.ResponseWriter, r * http.Request) {
fmt.Println(r.URL.Host)
}
http.HandleFunc()
can not be used to register a pattern to match a regular expression. In short, the pattern specified at HandleFunc()
can match a fixed, rooted path (like /favico.ico
) or rooted subtrees (like /images/
), longer patterns take precedence over shorter ones. You can find more details at the doc of the ServeMux
type.
What you can do is register your handler to a rooted subtree which may be everything with the /
pattern, and inside your handler you can do further regexp matching and routing.
For example:
func main() {
http.HandleFunc("/", route) // Match everything
http.ListenAndServe(":8080", nil)
}
var rNum = regexp.MustCompile(`\d`) // Has digit(s)
var rAbc = regexp.MustCompile(`abc`) // Contains "abc"
func route(w http.ResponseWriter, r *http.Request) {
switch {
case rNum.MatchString(r.URL.Path):
digits(w, r)
case rAbc.MatchString(r.URL.Path):
abc(w, r)
default:
w.Write([]byte("Unknown Pattern"))
}
}
func digits(w http.ResponseWriter, r *http.Request) {
w.Write([]byte("Has digits"))
}
func abc(w http.ResponseWriter, r *http.Request) {
w.Write([]byte("Has abc"))
}
Or use an external library like Gorilla MUX.
I use github.com/gorilla/mux package. So router looks like:
func main() {
r := mux.NewRouter()
r.HandleFunc("/{name:pattern}", handle)
http.ListenAndServe(":8080", r)
}
where {name:pattern}
could be simply {slug}
(without pattern) or {id:[0-9]+}
or combination of them /{category}/{id:[0-9]+}
. And get them in handler func:
func handle(w http.ResponseWriter, r *http.Request) {
params := mux.Vars(r)
// for /{category}/{id:[0-9]+} pattern
category := params["category"]
id := params["id"]
}
run it and try curl http://localhost:8080/whatever/1
Golang doesn't have built-in regex support for URL matching. And it's somewhat complicated to implement from scratch.
Maybe using a framework would be a better choice, such as beego or martin, etc.