Merge pull request #195 from matrix-org/erikj/content_disposition

Add Content-Disposition headers to media repo downloads

synapse/rest/media/v1/base_resource.py

@@ -27,9 +27,11 @@ from twisted.web.resource import Resource
 from twisted.protocols.basic import FileSender
 
 from synapse.util.async import ObservableDeferred
+from synapse.util.stringutils import is_ascii
 
 import os
 
+import cgi
 import logging
 
 logger = logging.getLogger(__name__)
@@ -37,8 +39,13 @@ logger = logging.getLogger(__name__)
 
 def parse_media_id(request):
     try:
-        server_name, media_id = request.postpath
-        return (server_name, media_id)
+        # This allows users to append e.g. /test.png to the URL. Useful for
+        # clients that parse the URL to see content type.
+        server_name, media_id = request.postpath[:2]
+        if len(request.postpath) > 2 and is_ascii(request.postpath[-1]):
+            return server_name, media_id, request.postpath[-1]
+        else:
+            return server_name, media_id, None
     except:
         raise SynapseError(
             404,
@@ -128,12 +135,21 @@ class BaseMediaResource(Resource):
             media_type = headers["Content-Type"][0]
             time_now_ms = self.clock.time_msec()
 
+            content_disposition = headers.get("Content-Disposition", None)
+            if content_disposition:
+                _, params = cgi.parse_header(content_disposition[0],)
+                upload_name = params.get("filename", None)
+                if upload_name and not is_ascii(upload_name):
+                    upload_name = None
+            else:
+                upload_name = None
+
             yield self.store.store_cached_remote_media(
                 origin=server_name,
                 media_id=media_id,
                 media_type=media_type,
                 time_now_ms=self.clock.time_msec(),
-                upload_name=None,
+                upload_name=upload_name,
                 media_length=length,
                 filesystem_id=file_id,
             )
@@ -144,7 +160,7 @@ class BaseMediaResource(Resource):
         media_info = {
             "media_type": media_type,
             "media_length": length,
-            "upload_name": None,
+            "upload_name": upload_name,
             "created_ts": time_now_ms,
             "filesystem_id": file_id,
         }
@@ -157,11 +173,16 @@ class BaseMediaResource(Resource):
 
     @defer.inlineCallbacks
     def _respond_with_file(self, request, media_type, file_path,
-                           file_size=None):
+                           file_size=None, upload_name=None):
         logger.debug("Responding with %r", file_path)
 
         if os.path.isfile(file_path):
             request.setHeader(b"Content-Type", media_type.encode("UTF-8"))
+            if upload_name:
+                request.setHeader(
+                    b"Content-Disposition",
+                    b"inline; filename=%s" % (upload_name.encode("utf-8"),),
+                )
 
             # cache for at least a day.
             # XXX: we might want to turn this off for data we don't want to

synapse/rest/media/v1/download_resource.py

@@ -32,14 +32,16 @@ class DownloadResource(BaseMediaResource):
     @request_handler
     @defer.inlineCallbacks
     def _async_render_GET(self, request):
-        server_name, media_id = parse_media_id(request)
+        server_name, media_id, name = parse_media_id(request)
         if server_name == self.server_name:
-            yield self._respond_local_file(request, media_id)
+            yield self._respond_local_file(request, media_id, name)
         else:
-            yield self._respond_remote_file(request, server_name, media_id)
+            yield self._respond_remote_file(
+                request, server_name, media_id, name
+            )
 
     @defer.inlineCallbacks
-    def _respond_local_file(self, request, media_id):
+    def _respond_local_file(self, request, media_id, name):
         media_info = yield self.store.get_local_media(media_id)
         if not media_info:
             self._respond_404(request)
@@ -47,24 +49,28 @@ class DownloadResource(BaseMediaResource):
 
         media_type = media_info["media_type"]
         media_length = media_info["media_length"]
+        upload_name = name if name else media_info["upload_name"]
         file_path = self.filepaths.local_media_filepath(media_id)
 
         yield self._respond_with_file(
-            request, media_type, file_path, media_length
+            request, media_type, file_path, media_length,
+            upload_name=upload_name,
         )
 
     @defer.inlineCallbacks
-    def _respond_remote_file(self, request, server_name, media_id):
+    def _respond_remote_file(self, request, server_name, media_id, name):
         media_info = yield self._get_remote_media(server_name, media_id)
 
         media_type = media_info["media_type"]
         media_length = media_info["media_length"]
         filesystem_id = media_info["filesystem_id"]
+        upload_name = name if name else media_info["upload_name"]
 
         file_path = self.filepaths.remote_media_filepath(
             server_name, filesystem_id
         )
 
         yield self._respond_with_file(
-            request, media_type, file_path, media_length
+            request, media_type, file_path, media_length,
+            upload_name=upload_name,
         )

synapse/rest/media/v1/thumbnail_resource.py

@@ -36,7 +36,7 @@ class ThumbnailResource(BaseMediaResource):
     @request_handler
     @defer.inlineCallbacks
     def _async_render_GET(self, request):
-        server_name, media_id = parse_media_id(request)
+        server_name, media_id, _ = parse_media_id(request)
         width = parse_integer(request, "width")
         height = parse_integer(request, "height")
         method = parse_string(request, "method", "scale")

synapse/rest/media/v1/upload_resource.py

@@ -15,7 +15,7 @@
 
 from synapse.http.server import respond_with_json, request_handler
 
-from synapse.util.stringutils import random_string
+from synapse.util.stringutils import random_string, is_ascii
 from synapse.api.errors import SynapseError
 
 from twisted.web.server import NOT_DONE_YET
@@ -84,6 +84,12 @@ class UploadResource(BaseMediaResource):
                 code=413,
             )
 
+        upload_name = request.args.get("filename", None)
+        if upload_name:
+            upload_name = upload_name[0]
+            if upload_name and not is_ascii(upload_name):
+                raise SynapseError(400, "filename must be ascii")
+
         headers = request.requestHeaders
 
         if headers.hasHeader("Content-Type"):
@@ -99,7 +105,7 @@ class UploadResource(BaseMediaResource):
         # TODO(markjh): parse content-dispostion
 
         content_uri = yield self.create_content(
-            media_type, None, request.content.read(),
+            media_type, upload_name, request.content.read(),
             content_length, auth_user
         )
 

synapse/util/stringutils.py

@@ -33,3 +33,12 @@ def random_string_with_symbols(length):
     return ''.join(
         random.choice(_string_with_symbols) for _ in xrange(length)
     )
+
+
+def is_ascii(s):
+    try:
+        s.encode("ascii")
+    except UnicodeDecodeError:
+        return False
+    else:
+        return True