X-Git-Url: https://git.arvados.org/arvados.git/blobdiff_plain/c8f07fc798c13ad8a45fb322eda4a922f664bb4c..1b8106da867aafbe6e4653485d5e644faee4c35d:/services/keep-web/handler.go diff --git a/services/keep-web/handler.go b/services/keep-web/handler.go index c5d439a399..8dee88d485 100644 --- a/services/keep-web/handler.go +++ b/services/keep-web/handler.go @@ -1,7 +1,6 @@ package main import ( - "flag" "fmt" "html" "io" @@ -9,7 +8,10 @@ import ( "net/http" "net/url" "os" + "path" + "strconv" "strings" + "sync" "time" "git.curoverse.com/arvados.git/sdk/go/arvadosclient" @@ -18,29 +20,22 @@ import ( "git.curoverse.com/arvados.git/sdk/go/keepclient" ) -type handler struct{} - -var ( - clientPool = arvadosclient.MakeClientPool() - trustAllContent = false - anonymousTokens []string -) - -func init() { - flag.BoolVar(&trustAllContent, "trust-all-content", false, - "Serve non-public content from a single origin. Dangerous: read docs before using!") +type handler struct { + Config *Config + clientPool *arvadosclient.ClientPool + setupOnce sync.Once } -// return a UUID or PDH if s begins with a UUID or URL-encoded PDH; -// otherwise return "". -func parseCollectionIdFromDNSName(s string) string { +// parseCollectionIDFromDNSName returns a UUID or PDH if s begins with +// a UUID or URL-encoded PDH; otherwise "". +func parseCollectionIDFromDNSName(s string) string { // Strip domain. if i := strings.IndexRune(s, '.'); i >= 0 { s = s[:i] } - // Names like {uuid}--dl.example.com serve the same purpose as - // {uuid}.dl.example.com but can reduce cost/effort of using - // [additional] wildcard certificates. + // Names like {uuid}--collections.example.com serve the same + // purpose as {uuid}.collections.example.com but can reduce + // cost/effort of using [additional] wildcard certificates. if i := strings.Index(s, "--"); i >= 0 { s = s[:i] } @@ -55,9 +50,10 @@ func parseCollectionIdFromDNSName(s string) string { var urlPDHDecoder = strings.NewReplacer(" ", "+", "-", "+") -// return a UUID or PDH if s is a UUID or a PDH (even if it is a PDH -// with "+" replaced by " " or "-"); otherwise return "". -func parseCollectionIdFromURL(s string) string { +// parseCollectionIDFromURL returns a UUID or PDH if s is a UUID or a +// PDH (even if it is a PDH with "+" replaced by " " or "-"); +// otherwise "". +func parseCollectionIDFromURL(s string) string { if arvadosclient.UUIDMatch(s) { return s } @@ -67,7 +63,14 @@ func parseCollectionIdFromURL(s string) string { return "" } +func (h *handler) setup() { + h.clientPool = arvadosclient.MakeClientPool() +} + +// ServeHTTP implements http.Handler. func (h *handler) ServeHTTP(wOrig http.ResponseWriter, r *http.Request) { + h.setupOnce.Do(h.setup) + var statusCode = 0 var statusText string @@ -97,48 +100,82 @@ func (h *handler) ServeHTTP(wOrig http.ResponseWriter, r *http.Request) { return } - arv := clientPool.Get() + if r.Header.Get("Origin") != "" { + // Allow simple cross-origin requests without user + // credentials ("user credentials" as defined by CORS, + // i.e., cookies, HTTP authentication, and client-side + // SSL certificates. See + // http://www.w3.org/TR/cors/#user-credentials). + w.Header().Set("Access-Control-Allow-Origin", "*") + } + + arv := h.clientPool.Get() if arv == nil { - statusCode, statusText = http.StatusInternalServerError, "Pool failed: "+clientPool.Err().Error() + statusCode, statusText = http.StatusInternalServerError, "Pool failed: "+h.clientPool.Err().Error() return } - defer clientPool.Put(arv) + defer h.clientPool.Put(arv) pathParts := strings.Split(r.URL.Path[1:], "/") - var targetId string + var targetID string var targetPath []string var tokens []string var reqTokens []string var pathToken bool - credentialsOK := trustAllContent + var attachment bool + credentialsOK := h.Config.TrustAllContent + + if r.Host != "" && r.Host == h.Config.AttachmentOnlyHost { + credentialsOK = true + attachment = true + } else if r.FormValue("disposition") == "attachment" { + attachment = true + } - if targetId = parseCollectionIdFromDNSName(r.Host); targetId != "" { - // http://ID.dl.example/PATH... + if targetID = parseCollectionIDFromDNSName(r.Host); targetID != "" { + // http://ID.collections.example/PATH... credentialsOK = true targetPath = pathParts } else if len(pathParts) >= 2 && strings.HasPrefix(pathParts[0], "c=") { // /c=ID/PATH... - targetId = parseCollectionIdFromURL(pathParts[0][2:]) + targetID = parseCollectionIDFromURL(pathParts[0][2:]) targetPath = pathParts[1:] } else if len(pathParts) >= 3 && pathParts[0] == "collections" { if len(pathParts) >= 5 && pathParts[1] == "download" { // /collections/download/ID/TOKEN/PATH... - targetId = pathParts[2] + targetID = pathParts[2] tokens = []string{pathParts[3]} targetPath = pathParts[4:] pathToken = true } else { // /collections/ID/PATH... - targetId = pathParts[1] - tokens = anonymousTokens + targetID = pathParts[1] + tokens = h.Config.AnonymousTokens targetPath = pathParts[2:] } } else { statusCode = http.StatusNotFound return } - if t := r.FormValue("api_token"); t != "" { + + formToken := r.FormValue("api_token") + if formToken != "" && r.Header.Get("Origin") != "" && attachment && r.URL.Query().Get("api_token") == "" { + // The client provided an explicit token in the POST + // body. The Origin header indicates this *might* be + // an AJAX request, in which case redirect-with-cookie + // won't work: we should just serve the content in the + // POST response. This is safe because: + // + // * We're supplying an attachment, not inline + // content, so we don't need to convert the POST to + // a GET and avoid the "really resubmit form?" + // problem. + // + // * The token isn't embedded in the URL, so we don't + // need to worry about bookmarks and copy/paste. + tokens = append(tokens, formToken) + } else if formToken != "" { // The client provided an explicit token in the query // string, or a form in POST body. We must put the // token in an HttpOnly cookie, and redirect to the @@ -149,7 +186,7 @@ func (h *handler) ServeHTTP(wOrig http.ResponseWriter, r *http.Request) { // It is not safe to copy the provided token // into a cookie unless the current vhost // (origin) serves only a single collection or - // we are in trustAllContent mode. + // we are in TrustAllContent mode. statusCode = http.StatusBadRequest return } @@ -167,13 +204,22 @@ func (h *handler) ServeHTTP(wOrig http.ResponseWriter, r *http.Request) { // resulting page. http.SetCookie(w, &http.Cookie{ - Name: "api_token", - Value: auth.EncodeTokenCookie([]byte(t)), + Name: "arvados_api_token", + Value: auth.EncodeTokenCookie([]byte(formToken)), Path: "/", - Expires: time.Now().AddDate(10, 0, 0), HttpOnly: true, }) - redir := (&url.URL{Host: r.Host, Path: r.URL.Path}).String() + + // Propagate query parameters (except api_token) from + // the original request. + redirQuery := r.URL.Query() + redirQuery.Del("api_token") + + redir := (&url.URL{ + Host: r.Host, + Path: r.URL.Path, + RawQuery: redirQuery.Encode(), + }).String() w.Header().Add("Location", redir) statusCode, statusText = http.StatusSeeOther, redir @@ -200,15 +246,16 @@ func (h *handler) ServeHTTP(wOrig http.ResponseWriter, r *http.Request) { if credentialsOK { reqTokens = auth.NewCredentialsFromHTTPRequest(r).Tokens } - tokens = append(reqTokens, anonymousTokens...) + tokens = append(reqTokens, h.Config.AnonymousTokens...) } if len(targetPath) > 0 && targetPath[0] == "_" { // If a collection has a directory called "t=foo" or - // "_", it can be served at //dl.example/_/t=foo/ or - // //dl.example/_/_/ respectively: //dl.example/t=foo/ - // won't work because t=foo will be interpreted as a - // token "foo". + // "_", it can be served at + // //collections.example/_/t=foo/ or + // //collections.example/_/_/ respectively: + // //collections.example/t=foo/ won't work because + // t=foo will be interpreted as a token "foo". targetPath = targetPath[1:] } @@ -216,7 +263,7 @@ func (h *handler) ServeHTTP(wOrig http.ResponseWriter, r *http.Request) { collection := make(map[string]interface{}) found := false for _, arv.ApiToken = range tokens { - err := arv.Get("collections", targetId, nil, &collection) + err := arv.Get("collections", targetID, nil, &collection) if err == nil { // Success found = true @@ -262,7 +309,7 @@ func (h *handler) ServeHTTP(wOrig http.ResponseWriter, r *http.Request) { // someone trying (anonymously) to download public // data that has been deleted. Allow a referrer to // provide this context somehow? - w.Header().Add("WWW-Authenticate", "Basic realm=\"dl\"") + w.Header().Add("WWW-Authenticate", "Basic realm=\"collections\"") statusCode = http.StatusUnauthorized return } @@ -273,6 +320,12 @@ func (h *handler) ServeHTTP(wOrig http.ResponseWriter, r *http.Request) { statusCode, statusText = http.StatusInternalServerError, err.Error() return } + if kc.Client != nil && kc.Client.Transport != nil { + // Workaround for https://dev.arvados.org/issues/9005 + if t, ok := kc.Client.Transport.(*http.Transport); ok { + defer t.CloseIdleConnections() + } + } rdr, err := kc.CollectionFileReader(collection, filename) if os.IsNotExist(err) { statusCode = http.StatusNotFound @@ -283,8 +336,10 @@ func (h *handler) ServeHTTP(wOrig http.ResponseWriter, r *http.Request) { } defer rdr.Close() - // One or both of these can be -1 if not found: basenamePos := strings.LastIndex(filename, "/") + if basenamePos < 0 { + basenamePos = 0 + } extPos := strings.LastIndex(filename, ".") if extPos > basenamePos { // Now extPos is safely >= 0. @@ -292,11 +347,34 @@ func (h *handler) ServeHTTP(wOrig http.ResponseWriter, r *http.Request) { w.Header().Set("Content-Type", t) } } - w.Header().Set("Content-Length", fmt.Sprintf("%d", rdr.Len())) + if rdr, ok := rdr.(keepclient.Reader); ok { + w.Header().Set("Content-Length", fmt.Sprintf("%d", rdr.Len())) + } + + applyContentDispositionHdr(w, r, filename[basenamePos:], attachment) - w.WriteHeader(http.StatusOK) - _, err = io.Copy(w, rdr) + modstr, _ := collection["modified_at"].(string) + modtime, err := time.Parse(time.RFC3339Nano, modstr) if err != nil { - statusCode, statusText = http.StatusBadGateway, err.Error() + modtime = time.Now() + } + http.ServeContent(w, r, path.Base(filename), modtime, rdr) +} + +func applyContentDispositionHdr(w http.ResponseWriter, r *http.Request, filename string, isAttachment bool) { + disposition := "inline" + if isAttachment { + disposition = "attachment" + } + if strings.ContainsRune(r.RequestURI, '?') { + // Help the UA realize that the filename is just + // "filename.txt", not + // "filename.txt?disposition=attachment". + // + // TODO(TC): Follow advice at RFC 6266 appendix D + disposition += "; filename=" + strconv.QuoteToASCII(filename) + } + if disposition != "inline" { + w.Header().Set("Content-Disposition", disposition) } }