feat: remove support for legacy ipfs-404.html

CHANGELOG.md

@@ -19,6 +19,7 @@ The following emojis are used to highlight certain changes:
 ### Changed
 
 ### Removed
+* 🛠 `boxo/gateway`: removed support for undocumented legacy `ipfs-404.html`. Use [`_redirects`](https://specs.ipfs.tech/http-gateways/web-redirects-file/) instead.
 
 ### Fixed
 

gateway/gateway_test.go

@@ -93,46 +93,6 @@ func TestGatewayGet(t *testing.T) {
 	}
 }
 
-func TestPretty404(t *testing.T) {
-	ts, backend, root := newTestServerAndNode(t, nil, "pretty-404.car")
-	t.Logf("test server url: %s", ts.URL)
-
-	host := "example.net"
-	backend.namesys["/ipns/"+host] = newMockNamesysItem(path.FromCid(root), 0)
-
-	for _, test := range []struct {
-		path   string
-		accept string
-		status int
-		text   string
-	}{
-		{"/ipfs-404.html", "text/html", http.StatusOK, "Custom 404"},
-		{"/nope", "text/html", http.StatusNotFound, "Custom 404"},
-		{"/nope", "text/*", http.StatusNotFound, "Custom 404"},
-		{"/nope", "*/*", http.StatusNotFound, "Custom 404"},
-		{"/nope", "application/json", http.StatusNotFound, fmt.Sprintf("failed to resolve /ipns/example.net/nope: no link named \"nope\" under %s\n", root.String())},
-		{"/deeper/nope", "text/html", http.StatusNotFound, "Deep custom 404"},
-		{"/deeper/", "text/html", http.StatusOK, ""},
-		{"/deeper", "text/html", http.StatusOK, ""},
-		{"/nope/nope", "text/html", http.StatusNotFound, "Custom 404"},
-	} {
-		testName := fmt.Sprintf("%s %s", test.path, test.accept)
-		t.Run(testName, func(t *testing.T) {
-			req := mustNewRequest(t, "GET", ts.URL+test.path, nil)
-			req.Header.Add("Accept", test.accept)
-			req.Host = host
-			resp := mustDo(t, req)
-			defer resp.Body.Close()
-			require.Equal(t, test.status, resp.StatusCode)
-			body, err := io.ReadAll(resp.Body)
-			require.NoError(t, err)
-			if test.text != "" {
-				require.Equal(t, test.text, string(body))
-			}
-		})
-	}
-}
-
 func TestHeaders(t *testing.T) {
 	t.Parallel()
 

gateway/handler.go

@@ -770,16 +770,6 @@ func (i *handler) handleWebRequestErrors(w http.ResponseWriter, r *http.Request,
 		}
 	}
 
-	// if Accept is text/html, see if ipfs-404.html is present
-	// This logic isn't documented and will likely be removed at some point.
-	// Any 404 logic in _redirects above will have already run by this time, so it's really an extra fall back
-	// PLEASE do not use this for new websites,
-	// follow https://docs.ipfs.tech/how-to/websites-on-ipfs/redirects-and-custom-404s/ instead.
-	if i.serveLegacy404IfPresent(w, r, immutableContentPath, logger) {
-		logger.Debugw("served legacy 404")
-		return path.ImmutablePath{}, false
-	}
-
 	err = fmt.Errorf("failed to resolve %s: %w", debugStr(contentPath.String()), err)
 	i.webError(w, r, err, http.StatusInternalServerError)
 	return path.ImmutablePath{}, false

gateway/handler_unixfs__redirects.go

@@ -240,73 +240,3 @@ func hasOriginIsolation(r *http.Request) bool {
 
 	return false
 }
-
-// Deprecated: legacy ipfs-404.html files are superseded by _redirects file
-// This is provided only for backward-compatibility, until websites migrate
-// to 404s managed via _redirects file (https://github.com/ipfs/specs/pull/290)
-func (i *handler) serveLegacy404IfPresent(w http.ResponseWriter, r *http.Request, imPath path.ImmutablePath, logger *zap.SugaredLogger) bool {
-	resolved404File, resolved404FileSize, ctype, err := i.searchUpTreeFor404(r, imPath)
-	if err != nil {
-		return false
-	}
-	defer resolved404File.Close()
-
-	logger.Debugw("using pretty 404 file", "path", imPath)
-	w.Header().Set("Content-Type", ctype)
-	w.Header().Set("Content-Length", strconv.FormatInt(resolved404FileSize, 10))
-	w.WriteHeader(http.StatusNotFound)
-	_, err = io.CopyN(w, resolved404File, resolved404FileSize)
-	return err == nil
-}
-
-func (i *handler) searchUpTreeFor404(r *http.Request, imPath path.ImmutablePath) (io.ReadCloser, int64, string, error) {
-	filename404, ctype, err := preferred404Filename(r.Header.Values("Accept"))
-	if err != nil {
-		return nil, 0, "", err
-	}
-
-	pathComponents := strings.Split(imPath.String(), "/")
-
-	for idx := len(pathComponents); idx >= 3; idx-- {
-		pretty404 := gopath.Join(append(pathComponents[0:idx], filename404)...)
-		parsed404Path, err := path.NewPath("/" + pretty404)
-		if err != nil {
-			break
-		}
-		imparsed404Path, err := path.NewImmutablePath(parsed404Path)
-		if err != nil {
-			break
-		}
-
-		_, getResp, err := i.backend.Get(r.Context(), imparsed404Path)
-		if err != nil {
-			continue
-		}
-		if getResp.bytes == nil {
-			// Close the response here if not returning bytes, otherwise it's the caller's responsibility to close the io.ReadCloser
-			getResp.Close()
-			return nil, 0, "", fmt.Errorf("found a pretty 404 but it was not a file")
-		}
-		return getResp.bytes, getResp.bytesSize, ctype, nil
-	}
-
-	return nil, 0, "", fmt.Errorf("no pretty 404 in any parent folder")
-}
-
-func preferred404Filename(acceptHeaders []string) (string, string, error) {
-	// If we ever want to offer a 404 file for a different content type
-	// then this function will need to parse q weightings, but for now
-	// the presence of anything matching HTML is enough.
-	for _, acceptHeader := range acceptHeaders {
-		accepted := strings.Split(acceptHeader, ",")
-		for _, spec := range accepted {
-			contentType := strings.SplitN(spec, ";", 1)[0]
-			switch contentType {
-			case "*/*", "text/*", "text/html":
-				return "ipfs-404.html", "text/html", nil
-			}
-		}
-	}
-
-	return "", "", fmt.Errorf("there is no 404 file for the requested content types")
-}