Preserve uploaded file extensions

server/media/fs/filesys.go

@@ -139,8 +139,7 @@ func (fh *fshandler) Upload(fdef *types.FileDef, file io.Reader) (string, int64,
 		return "", 0, err
 	}
 
-	fname := fdef.Id
-	fname += media.ExtensionByMimeType(fdef.MimeType)
+	fname := media.FileURLName(fdef)
 
 	fdef.Location = location
 	// Use file path to create ETag. File paths are unique so will be the ETag.

server/media/media.go

@@ -69,6 +69,58 @@ type AllowedOrigin struct {
 
 var fileNamePattern = regexp.MustCompile(`^[-_A-Za-z0-9]+`)
 
+var uploadExtensionMimeTypes = map[string]string{
+	".aac":  "audio/aac",
+	".m4a":  "audio/x-m4a",
+	".doc":  "application/msword",
+	".docx": "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
+	".ppt":  "application/vnd.ms-powerpoint",
+	".pptx": "application/vnd.openxmlformats-officedocument.presentationml.presentation",
+	".xls":  "application/vnd.ms-excel",
+	".xlsx": "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
+}
+
+// MimeTypeForUpload picks the MIME type to persist for an uploaded file. Some common formats are
+// containers and are misdetected by http.DetectContentType, for example m4a as video/mp4 and Office
+// documents as zip, so the original file extension wins for those cases.
+func MimeTypeForUpload(detected, clientContentType, originalName string) string {
+	ext := ExtensionByFilename(originalName)
+	if extMime := uploadExtensionMimeTypes[ext]; extMime != "" {
+		return extMime
+	}
+	if clientContentType != "" {
+		return clientContentType
+	}
+	return strings.TrimSpace(detected)
+}
+
+// FileURLName returns the public URL file name for a stored file.
+func FileURLName(fdef *types.FileDef) string {
+	if fdef == nil {
+		return ""
+	}
+	fname := fdef.Id
+	if ext := ExtensionByFilename(fdef.OriginalName); ext != "" {
+		return fname + ext
+	}
+	return fname + ExtensionByMimeType(fdef.MimeType)
+}
+
+// ExtensionByFilename returns a safe lower-case extension from a client-provided file name.
+func ExtensionByFilename(name string) string {
+	_, fname := path.Split(strings.TrimSpace(name))
+	ext := strings.ToLower(path.Ext(fname))
+	if len(ext) < 2 || len(ext) > 16 {
+		return ""
+	}
+	for _, r := range ext[1:] {
+		if (r < 'a' || r > 'z') && (r < '0' || r > '9') {
+			return ""
+		}
+	}
+	return ext
+}
+
 // ExtensionByMimeType returns a stable URL extension for uploaded media.
 func ExtensionByMimeType(mimeType string) string {
 	switch strings.ToLower(strings.TrimSpace(strings.Split(mimeType, ";")[0])) {

server/media/media_test.go

@@ -3,6 +3,8 @@ package media
 import (
 	"strings"
 	"testing"
+
+	"github.com/tinode/chat/server/store/types"
 )
 
 func TestMatchCORSOrigin(t *testing.T) {
@@ -237,3 +239,71 @@ func TestExtensionByMimeType(t *testing.T) {
 		}
 	}
 }
+
+func TestMimeTypeForUploadPrefersContainerFilename(t *testing.T) {
+	cases := []struct {
+		name     string
+		detect   string
+		client   string
+		filename string
+		want     string
+	}{
+		{
+			name:     "m4a misdetected as mp4",
+			detect:   "video/mp4",
+			filename: "voice.m4a",
+			want:     "audio/x-m4a",
+		},
+		{
+			name:     "xlsx misdetected as zip",
+			detect:   "application/zip",
+			filename: "iOS任务拆解.xlsx",
+			want:     "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
+		},
+		{
+			name:     "client content type fallback",
+			detect:   "application/octet-stream",
+			client:   "image/jfif",
+			filename: "thumb",
+			want:     "image/jfif",
+		},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			if got := MimeTypeForUpload(tc.detect, tc.client, tc.filename); got != tc.want {
+				t.Fatalf("MimeTypeForUpload()=%q, want %q", got, tc.want)
+			}
+		})
+	}
+}
+
+func TestFileURLNamePrefersOriginalExtension(t *testing.T) {
+	cases := []struct {
+		name string
+		def  *types.FileDef
+		want string
+	}{
+		{
+			name: "voice keeps m4a extension",
+			def:  &types.FileDef{ObjHeader: types.ObjHeader{Id: "file123"}, MimeType: "video/mp4", OriginalName: "voice.m4a"},
+			want: "file123.m4a",
+		},
+		{
+			name: "xlsx keeps xlsx extension",
+			def:  &types.FileDef{ObjHeader: types.ObjHeader{Id: "sheet123"}, MimeType: "application/zip", OriginalName: "iOS任务拆解.xlsx"},
+			want: "sheet123.xlsx",
+		},
+		{
+			name: "falls back to mime extension",
+			def:  &types.FileDef{ObjHeader: types.ObjHeader{Id: "image123"}, MimeType: "image/jpeg"},
+			want: "image123.jpg",
+		},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			if got := FileURLName(tc.def); got != tc.want {
+				t.Fatalf("FileURLName()=%q, want %q", got, tc.want)
+			}
+		})
+	}
+}

server/media/oss/oss.go

@@ -245,8 +245,7 @@ func (oh *ossHandler) Upload(fdef *types.FileDef, file io.Reader) (string, int64
 		return "", 0, err
 	}
 
-	fname := fdef.Id
-	fname += media.ExtensionByMimeType(fdef.MimeType)
+	fname := media.FileURLName(fdef)
 
 	fdef.Location = key
 	fdef.ETag = etagFromLocation(key)

server/media/s3/s3.go

@@ -294,8 +294,7 @@ func (ah *awshandler) Upload(fdef *types.FileDef, file io.Reader) (string, int64
 		return "", 0, err
 	}
 
-	fname := fdef.Id
-	fname += media.ExtensionByMimeType(fdef.MimeType)
+	fname := media.FileURLName(fdef)
 
 	fdef.Location = key
 	if result.ETag != nil {