/* Copyright 2011 Google Inc. Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. */ package main import ( "bytes" "flag" "fmt" "html/template" "io" "io/ioutil" "log" "net/http" "net/http/cgi" "net/http/httputil" "net/url" "os" "os/exec" "path/filepath" "regexp" "strings" txttemplate "text/template" "time" ) const defaultAddr = ":31798" // default webserver address var h1TitlePattern = regexp.MustCompile(`
go get camlistore.org/cmd/` + subtitle + `
` content = bytes.Replace(content, []byte("
"), []byte(toInsert), 1) } d := struct { Title string Subtitle string Content template.HTML }{ title, subtitle, template.HTML(content), } if err := pageHtml.Execute(w, &d); err != nil { log.Printf("godocHTML.Execute: %s", err) } } func readTemplate(name string) *template.Template { fileName := filepath.Join(*root, "tmpl", name) data, err := ioutil.ReadFile(fileName) if err != nil { log.Fatalf("ReadFile %s: %v", fileName, err) } t, err := template.New(name).Funcs(fmap).Parse(string(data)) if err != nil { log.Fatalf("%s: %v", fileName, err) } return t } func readTemplates() { pageHtml = readTemplate("page.html") errorHtml = readTemplate("error.html") // TODO(mpl): see about not using text template anymore? packageHTML = readTextTemplate("package.html") } func serveError(w http.ResponseWriter, r *http.Request, relpath string, err error) { contents := applyTemplate(errorHtml, "errorHtml", err) // err may contain an absolute path! w.WriteHeader(http.StatusNotFound) servePage(w, "File "+relpath, "", contents) } func mainHandler(rw http.ResponseWriter, req *http.Request) { relPath := req.URL.Path[1:] // serveFile URL paths start with '/' if strings.Contains(relPath, "..") { return } if strings.HasPrefix(relPath, "gw/") { path := relPath[3:] http.Redirect(rw, req, "http://camlistore.org/code/?p=camlistore.git;f="+path+";hb=master", http.StatusFound) return } absPath := filepath.Join(*root, "content", relPath) fi, err := os.Lstat(absPath) if err != nil { log.Print(err) serveError(rw, req, relPath, err) return } if fi.IsDir() { relPath += "/index.html" absPath = filepath.Join(*root, "content", relPath) fi, err = os.Lstat(absPath) if err != nil { log.Print(err) serveError(rw, req, relPath, err) return } } switch { case !fi.IsDir(): serveFile(rw, req, relPath, absPath) } } func serveFile(rw http.ResponseWriter, req *http.Request, relPath, absPath string) { data, err := ioutil.ReadFile(absPath) if err != nil { serveError(rw, req, absPath, err) return } title := "" if m := h1TitlePattern.FindSubmatch(data); len(m) > 1 { title = string(m[1]) } servePage(rw, title, "", data) } func isBot(r *http.Request) bool { agent := r.Header.Get("User-Agent") return strings.Contains(agent, "Baidu") || strings.Contains(agent, "bingbot") || strings.Contains(agent, "Ezooms") || strings.Contains(agent, "Googlebot") } type noWwwHandler struct { Handler http.Handler } func (h *noWwwHandler) ServeHTTP(rw http.ResponseWriter, r *http.Request) { // Some bots (especially Baidu) don't seem to respect robots.txt and swamp gitweb.cgi, // so explicitly protect it from bots. if ru := r.URL.RequestURI(); strings.Contains(ru, "/code/") && strings.Contains(ru, "?") && isBot(r) { http.Error(rw, "bye", http.StatusUnauthorized) log.Printf("bot denied") return } host := strings.ToLower(r.Host) if host == "www.camlistore.org" { http.Redirect(rw, r, "http://camlistore.org"+r.URL.RequestURI(), http.StatusFound) return } h.Handler.ServeHTTP(rw, r) } // runAsChild runs res as a child process and // does not wait for it to finish. func runAsChild(res string) { cmdName, err := exec.LookPath(res) if err != nil { log.Fatalf("Could not find %v in $PATH: %v", res, err) } cmd := exec.Command(cmdName) cmd.Stderr = os.Stderr cmd.Stdout = os.Stdout log.Printf("Running %v", res) if err := cmd.Start(); err != nil { log.Fatal("Program %v failed to start: %v", res, err) } go func() { if err := cmd.Wait(); err != nil { log.Fatalf("Program %s did not end successfully: %v", res, err) } }() } func main() { flag.Parse() if *root == "" { var err error *root, err = os.Getwd() if err != nil { log.Fatalf("Failed to getwd: %v", err) } } readTemplates() mux := http.DefaultServeMux mux.Handle("/favicon.ico", http.FileServer(http.Dir(filepath.Join(*root, "static")))) mux.Handle("/robots.txt", http.FileServer(http.Dir(filepath.Join(*root, "static")))) mux.Handle("/static/", http.StripPrefix("/static/", http.FileServer(http.Dir(filepath.Join(*root, "static"))))) mux.Handle("/talks/", http.StripPrefix("/talks/", http.FileServer(http.Dir(filepath.Join(*root, "talks"))))) mux.Handle(pkgPattern, godocHandler{}) mux.Handle(cmdPattern, godocHandler{}) mux.HandleFunc("/r/", gerritRedirect) mux.HandleFunc("/debugz/ip", ipHandler) testCgi := &cgi.Handler{Path: filepath.Join(*root, "test.cgi"), Root: "/test.cgi", } mux.Handle("/test.cgi", testCgi) mux.Handle("/test.cgi/foo", testCgi) mux.HandleFunc("/issue/", issueRedirect) mux.HandleFunc("/", mainHandler) if *buildbotHost != "" && *buildbotBackend != "" { buildbotUrl, err := url.Parse(*buildbotBackend) if err != nil { log.Fatalf("Failed to parse %v as a URL: %v", *buildbotBackend, err) } buildbotHandler := httputil.NewSingleHostReverseProxy(buildbotUrl) bbhpattern := strings.TrimRight(*buildbotHost, "/") + "/" mux.Handle(bbhpattern, buildbotHandler) } var handler http.Handler = &noWwwHandler{Handler: mux} if *logDir != "" || *logStdout { handler = NewLoggingHandler(handler, *logDir, *logStdout) } if *alsoRun != "" { runAsChild(*alsoRun) } errch := make(chan error) httpServer := &http.Server{ Addr: *httpAddr, Handler: handler, ReadTimeout: 5 * time.Minute, WriteTimeout: 30 * time.Minute, } go func() { errch <- httpServer.ListenAndServe() }() if *httpsAddr != "" { log.Printf("Starting TLS server on %s", *httpsAddr) httpsServer := new(http.Server) *httpsServer = *httpServer httpsServer.Addr = *httpsAddr go func() { errch <- httpsServer.ListenAndServeTLS(*tlsCertFile, *tlsKeyFile) }() } log.Fatalf("Serve error: %v", <-errch) } var issueNum = regexp.MustCompile(`^/issue/(\d+)$`) func issueRedirect(w http.ResponseWriter, r *http.Request) { m := issueNum.FindStringSubmatch(r.URL.Path) if m == nil { http.Error(w, "Bad request", 400) return } http.Redirect(w, r, "https://code.google.com/p/camlistore/issues/detail?id="+m[1], http.StatusFound) } func gerritRedirect(w http.ResponseWriter, r *http.Request) { dest := "https://camlistore-review.googlesource.com/" if len(r.URL.Path) > len("/r/") { dest += r.URL.Path[1:] } http.Redirect(w, r, dest, http.StatusFound) } // Not sure what's making these broken URLs like: // // http://localhost:8080/code/?p=camlistore.git%3Bf=doc/json-signing/json-signing.txt%3Bhb=master // // ... but something is. Maybe Buzz? For now just re-write them // . Doesn't seem to be a bug in the CGI implementation, though, which // is what I'd originally suspected. /* func (fu *fixUpGitwebUrls) ServeHTTP(rw http.ResponseWriter, req *http.Request) { oldUrl := req.URL.String() newUrl := strings.Replace(oldUrl, "%3B", ";", -1) if newUrl == oldUrl { fu.handler.ServeHTTP(rw, req) return } http.Redirect(rw, req, newUrl, http.StatusFound) } */ func ipHandler(w http.ResponseWriter, r *http.Request) { out, _ := exec.Command("ip", "-f", "inet", "addr", "show", "dev", "eth0").Output() str := string(out) pos := strings.Index(str, "inet ") if pos == -1 { return } str = str[pos+5:] pos = strings.Index(str, "/") if pos == -1 { return } str = str[:pos] w.Write([]byte(str)) }