Handle missing Content-Type header when accessing remote media

changelog.d/11200.bugfix

@@ -0,0 +1 @@
+Fixes a long-standing bug wherein a missing Content-Type header when downloading remote media causes Synapse to throw an error.

synapse/rest/media/v1/media_repository.py

@@ -214,7 +214,9 @@ async def get_local_media(
 
         self.mark_recently_accessed(None, media_id)
 
-        media_type = media_info["media_type"]
+        media_type = media_info.get("media_type")
+        if not media_type:
+            media_type = "application/octect-stream"
         media_length = media_info["media_length"]
         upload_name = name if name else media_info["upload_name"]
         url_cache = media_info["url_cache"]
@@ -333,6 +335,9 @@ async def _get_remote_media_impl(
                 logger.info("Media is quarantined")
                 raise NotFoundError()
 
+            if not media_info["media_type"]:
+                media_info["media_type"] = "application/octect-stream"
+
             responder = await self.media_storage.fetch_media(file_info)
             if responder:
                 return responder, media_info
@@ -354,6 +359,8 @@ async def _get_remote_media_impl(
                 raise e
 
         file_id = media_info["filesystem_id"]
+        if not media_info["media_type"]:
+            media_info["media_type"] = "application/octect-stream"
         file_info = FileInfo(server_name, file_id)
 
         # We generate thumbnails even if another process downloaded the media
@@ -445,7 +452,10 @@ async def _download_remote_file(
 
             await finish()
 
-            media_type = headers[b"Content-Type"][0].decode("ascii")
+            if b"Content-Type" in headers:
+                media_type = headers[b"Content-Type"][0].decode("ascii")
+            else:
+                media_type = "application/octet-stream"
             upload_name = get_filename_from_headers(headers)
             time_now_ms = self.clock.time_msec()
 

synapse/rest/media/v1/upload_resource.py

@@ -80,7 +80,7 @@ async def _async_render_POST(self, request: SynapseRequest) -> None:
             assert content_type_headers  # for mypy
             media_type = content_type_headers[0].decode("ascii")
         else:
-            raise SynapseError(msg="Upload request missing 'Content-Type'", code=400)
+            media_type = "application/octet-stream"
 
         # if headers.hasHeader(b"Content-Disposition"):
         #     disposition = headers.getRawHeaders(b"Content-Disposition")[0]

tests/rest/media/v1/test_media_storage.py

@@ -248,7 +248,7 @@ def prepare(self, reactor, clock, hs):
 
         self.media_id = "example.com/12345"
 
-    def _req(self, content_disposition):
+    def _req(self, content_disposition, include_content_type=True):
 
         channel = make_request(
             self.reactor,
@@ -269,10 +269,16 @@ def _req(self, content_disposition):
         )
         self.assertEqual(self.fetches[0][3], {"allow_remote": "false"})
 
-        headers = {
-            b"Content-Length": [b"%d" % (len(self.test_image.data))],
-            b"Content-Type": [self.test_image.content_type],
-        }
+        if include_content_type:
+            headers = {
+                b"Content-Length": [b"%d" % (len(self.test_image.data))],
+                b"Content-Type": [self.test_image.content_type],
+            }
+        else:
+            headers = {
+                b"Content-Length": [b"%d" % (len(self.test_image.data))],
+            }
+
         if content_disposition:
             headers[b"Content-Disposition"] = [content_disposition]
 
@@ -285,6 +291,14 @@ def _req(self, content_disposition):
 
         return channel
 
+    def test_handle_missing_content_type(self):
+        channel = self._req(b"inline; filename=out" + self.test_image.extension, False)
+        headers = channel.headers
+        self.assertEqual(channel.code, 200)
+        self.assertEqual(
+            headers.getRawHeaders(b"Content-Type"), [b"application/octet-stream"]
+        )
+
     def test_disposition_filename_ascii(self):
         """
         If the filename is filename=<ascii> then Synapse will decode it as an