feat(gateway): TAR response format (#9029)

Implementation of IPIP-288 (ipfs/specs#288)

Co-authored-by: Marcin Rataj <lidel@lidel.org>

This commit was moved from ipfs/kubo@a210abd

gateway/core/corehttp/gateway_handler.go

@@ -430,6 +430,10 @@ func (i *gatewayHandler) getOrHeadHandler(w http.ResponseWriter, r *http.Request
 		carVersion := formatParams["version"]
 		i.serveCAR(r.Context(), w, r, resolvedPath, contentPath, carVersion, begin)
 		return
+	case "application/x-tar":
+		logger.Debugw("serving tar file", "path", contentPath)
+		i.serveTAR(r.Context(), w, r, resolvedPath, contentPath, begin, logger)
+		return
 	default: // catch-all for unsuported application/vnd.*
 		err := fmt.Errorf("unsupported format %q", responseFormat)
 		webError(w, "failed respond with requested content type", err, http.StatusBadRequest)
@@ -842,9 +846,10 @@ func getEtag(r *http.Request, cid cid.Cid) string {
 	responseFormat, _, err := customResponseFormat(r)
 	if err == nil && responseFormat != "" {
 		// application/vnd.ipld.foo → foo
-		f := responseFormat[strings.LastIndex(responseFormat, ".")+1:]
-		// Etag: "cid.foo" (gives us nice compression together with Content-Disposition in block (raw) and car responses)
-		suffix = `.` + f + suffix
+		// application/x-bar → x-bar
+		shortFormat := responseFormat[strings.LastIndexAny(responseFormat, "/.")+1:]
+		// Etag: "cid.shortFmt" (gives us nice compression together with Content-Disposition in block (raw) and car responses)
+		suffix = `.` + shortFormat + suffix
 	}
 	// TODO: include selector suffix when https://github.com/ipfs/kubo/issues/8769 lands
 	return prefix + cid.String() + suffix
@@ -859,14 +864,17 @@ func customResponseFormat(r *http.Request) (mediaType string, params map[string]
 			return "application/vnd.ipld.raw", nil, nil
 		case "car":
 			return "application/vnd.ipld.car", nil, nil
+		case "tar":
+			return "application/x-tar", nil, nil
 		}
 	}
 	// Browsers and other user agents will send Accept header with generic types like:
 	// Accept:text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,*/*;q=0.8
-	// We only care about explciit, vendor-specific content-types.
+	// We only care about explicit, vendor-specific content-types.
 	for _, accept := range r.Header.Values("Accept") {
 		// respond to the very first ipld content type
-		if strings.HasPrefix(accept, "application/vnd.ipld") {
+		if strings.HasPrefix(accept, "application/vnd.ipld") ||
+			strings.HasPrefix(accept, "application/x-tar") {
 			mediatype, params, err := mime.ParseMediaType(accept)
 			if err != nil {
 				return "", nil, err

gateway/core/corehttp/gateway_handler_tar.go

@@ -0,0 +1,92 @@
+package corehttp
+
+import (
+	"context"
+	"html"
+	"net/http"
+	"time"
+
+	files "github.com/ipfs/go-ipfs-files"
+	ipath "github.com/ipfs/interface-go-ipfs-core/path"
+	"github.com/ipfs/kubo/tracing"
+	"go.opentelemetry.io/otel/attribute"
+	"go.opentelemetry.io/otel/trace"
+	"go.uber.org/zap"
+)
+
+var unixEpochTime = time.Unix(0, 0)
+
+func (i *gatewayHandler) serveTAR(ctx context.Context, w http.ResponseWriter, r *http.Request, resolvedPath ipath.Resolved, contentPath ipath.Path, begin time.Time, logger *zap.SugaredLogger) {
+	ctx, span := tracing.Span(ctx, "Gateway", "ServeTAR", trace.WithAttributes(attribute.String("path", resolvedPath.String())))
+	defer span.End()
+
+	ctx, cancel := context.WithCancel(ctx)
+	defer cancel()
+
+	// Get Unixfs file
+	file, err := i.api.Unixfs().Get(ctx, resolvedPath)
+	if err != nil {
+		webError(w, "ipfs cat "+html.EscapeString(contentPath.String()), err, http.StatusBadRequest)
+		return
+	}
+	defer file.Close()
+
+	rootCid := resolvedPath.Cid()
+
+	// Set Cache-Control and read optional Last-Modified time
+	modtime := addCacheControlHeaders(w, r, contentPath, rootCid)
+
+	// Weak Etag W/ because we can't guarantee byte-for-byte identical
+	// responses, but still want to benefit from HTTP Caching. Two TAR
+	// responses for the same CID will be logically equivalent,
+	// but when TAR is streamed, then in theory, files and directories
+	// may arrive in different order (depends on TAR lib and filesystem/inodes).
+	etag := `W/` + getEtag(r, rootCid)
+	w.Header().Set("Etag", etag)
+
+	// Finish early if Etag match
+	if r.Header.Get("If-None-Match") == etag {
+		w.WriteHeader(http.StatusNotModified)
+		return
+	}
+
+	// Set Content-Disposition
+	var name string
+	if urlFilename := r.URL.Query().Get("filename"); urlFilename != "" {
+		name = urlFilename
+	} else {
+		name = rootCid.String() + ".tar"
+	}
+	setContentDispositionHeader(w, name, "attachment")
+
+	// Construct the TAR writer
+	tarw, err := files.NewTarWriter(w)
+	if err != nil {
+		webError(w, "could not build tar writer", err, http.StatusInternalServerError)
+		return
+	}
+	defer tarw.Close()
+
+	// Sets correct Last-Modified header. This code is borrowed from the standard
+	// library (net/http/server.go) as we cannot use serveFile without throwing the entire
+	// TAR into the memory first.
+	if !(modtime.IsZero() || modtime.Equal(unixEpochTime)) {
+		w.Header().Set("Last-Modified", modtime.UTC().Format(http.TimeFormat))
+	}
+
+	w.Header().Set("Content-Type", "application/x-tar")
+	w.Header().Set("X-Content-Type-Options", "nosniff") // no funny business in the browsers :^)
+
+	// The TAR has a top-level directory (or file) named by the CID.
+	if err := tarw.WriteFile(file, rootCid.String()); err != nil {
+		w.Header().Set("X-Stream-Error", err.Error())
+		// Trailer headers do not work in web browsers
+		// (see https://github.com/mdn/browser-compat-data/issues/14703)
+		// and we have limited options around error handling in browser contexts.
+		// To improve UX/DX, we finish response stream with error message, allowing client to
+		// (1) detect error by having corrupted TAR
+		// (2) be able to reason what went wrong by instecting the tail of TAR stream
+		_, _ = w.Write([]byte(err.Error()))
+		return
+	}
+}