From 04248e4810cca5008993c7665d8aa02a04934ab2 Mon Sep 17 00:00:00 2001
From: Henrique Dias <hacdias@gmail.com>
Date: Thu, 9 Jun 2022 15:35:53 +0200
Subject: [PATCH] feat: gateway support for tar

---
 core/corehttp/gateway_handler.go     | 12 ++++-
 core/corehttp/gateway_handler_tar.go | 77 ++++++++++++++++++++++++++++
 2 files changed, 88 insertions(+), 1 deletion(-)
 create mode 100644 core/corehttp/gateway_handler_tar.go

diff --git a/core/corehttp/gateway_handler.go b/core/corehttp/gateway_handler.go
index f32fac54ecd6..3b41570eef85 100644
--- a/core/corehttp/gateway_handler.go
+++ b/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)
@@ -873,6 +877,10 @@ func getEtag(r *http.Request, cid cid.Cid) string {
 		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
+		// Since different TAR implementations may produce different byte-for-byte responses, we define a weak Etag.
+		if responseFormat == "application/x-tar" {
+			prefix = "W/" + prefix
+		}
 	}
 	// TODO: include selector suffix when https://github.com/ipfs/go-ipfs/issues/8769 lands
 	return prefix + cid.String() + suffix
@@ -887,11 +895,13 @@ 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") {
diff --git a/core/corehttp/gateway_handler_tar.go b/core/corehttp/gateway_handler_tar.go
new file mode 100644
index 000000000000..47ef5fdef280
--- /dev/null
+++ b/core/corehttp/gateway_handler_tar.go
@@ -0,0 +1,77 @@
+package corehttp
+
+import (
+	"context"
+	"html"
+	"net/http"
+	"time"
+
+	files "github.com/ipfs/go-ipfs-files"
+	"github.com/ipfs/go-ipfs/tracing"
+	ipath "github.com/ipfs/interface-go-ipfs-core/path"
+	"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()
+
+	// Set Cache-Control and read optional Last-Modified time
+	modtime := addCacheControlHeaders(w, r, contentPath, resolvedPath.Cid())
+
+	// Finish early if Etag match
+	if r.Header.Get("If-None-Match") == getEtag(r, resolvedPath.Cid()) {
+		w.WriteHeader(http.StatusNotModified)
+		return
+	}
+
+	// Set Content-Disposition
+	var name string
+	if urlFilename := r.URL.Query().Get("filename"); urlFilename != "" {
+		name = urlFilename
+	} else {
+		name = resolvedPath.Cid().String() + ".tar"
+	}
+	setContentDispositionHeader(w, name, "attachment")
+
+	// Get Unixfs file
+	file, err := i.api.Unixfs().Get(ctx, resolvedPath)
+	if err != nil {
+		webError(w, "ipfs cat "+html.EscapeString(contentPath.String()), err, http.StatusNotFound)
+		return
+	}
+	defer file.Close()
+
+	// Construct the TAR writer
+	tarw, err := files.NewTarWriter(w)
+	if err != nil {
+		webError(w, "could not build TarWriter", 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")
+
+	if err := tarw.WriteFile(file, name); err != nil {
+		// We return error as a trailer, however it is not something browsers can access
+		// (https://github.com/mdn/browser-compat-data/issues/14703)
+		// Due to this, we suggest client always verify that
+		// the received CAR stream response is matching requested DAG selector
+		w.Header().Set("X-Stream-Error", err.Error())
+		return
+	}
+}