Do not error when thumbnailing invalid files (#8236)

If a file cannot be thumbnailed for some reason (e.g. the file is empty), then
catch the exception and convert it to a reasonable error message for the client.

changelog.d/8236.bugfix

@@ -0,0 +1 @@
+Fix a longstanding bug where files that could not be thumbnailed would result in an Internal Server Error.

synapse/rest/media/v1/media_repository.py

@@ -53,7 +53,7 @@ from .media_storage import MediaStorage
 from .preview_url_resource import PreviewUrlResource
 from .storage_provider import StorageProviderWrapper
 from .thumbnail_resource import ThumbnailResource
-from .thumbnailer import Thumbnailer
+from .thumbnailer import Thumbnailer, ThumbnailError
 from .upload_resource import UploadResource
 
 logger = logging.getLogger(__name__)
@@ -460,13 +460,30 @@ class MediaRepository:
         return t_byte_source
 
     async def generate_local_exact_thumbnail(
-        self, media_id, t_width, t_height, t_method, t_type, url_cache
-    ):
+        self,
+        media_id: str,
+        t_width: int,
+        t_height: int,
+        t_method: str,
+        t_type: str,
+        url_cache: str,
+    ) -> Optional[str]:
         input_path = await self.media_storage.ensure_media_is_in_local_cache(
             FileInfo(None, media_id, url_cache=url_cache)
         )
 
-        thumbnailer = Thumbnailer(input_path)
+        try:
+            thumbnailer = Thumbnailer(input_path)
+        except ThumbnailError as e:
+            logger.warning(
+                "Unable to generate a thumbnail for local media %s using a method of %s and type of %s: %s",
+                media_id,
+                t_method,
+                t_type,
+                e,
+            )
+            return None
+
         t_byte_source = await defer_to_thread(
             self.hs.get_reactor(),
             self._generate_thumbnail,
@@ -506,14 +523,36 @@ class MediaRepository:
 
             return output_path
 
+        # Could not generate thumbnail.
+        return None
+
     async def generate_remote_exact_thumbnail(
-        self, server_name, file_id, media_id, t_width, t_height, t_method, t_type
-    ):
+        self,
+        server_name: str,
+        file_id: str,
+        media_id: str,
+        t_width: int,
+        t_height: int,
+        t_method: str,
+        t_type: str,
+    ) -> Optional[str]:
         input_path = await self.media_storage.ensure_media_is_in_local_cache(
             FileInfo(server_name, file_id, url_cache=False)
         )
 
-        thumbnailer = Thumbnailer(input_path)
+        try:
+            thumbnailer = Thumbnailer(input_path)
+        except ThumbnailError as e:
+            logger.warning(
+                "Unable to generate a thumbnail for remote media %s from %s using a method of %s and type of %s: %s",
+                media_id,
+                server_name,
+                t_method,
+                t_type,
+                e,
+            )
+            return None
+
         t_byte_source = await defer_to_thread(
             self.hs.get_reactor(),
             self._generate_thumbnail,
@@ -559,6 +598,9 @@ class MediaRepository:
 
             return output_path
 
+        # Could not generate thumbnail.
+        return None
+
     async def _generate_thumbnails(
         self,
         server_name: Optional[str],
@@ -590,7 +632,18 @@ class MediaRepository:
             FileInfo(server_name, file_id, url_cache=url_cache)
         )
 
-        thumbnailer = Thumbnailer(input_path)
+        try:
+            thumbnailer = Thumbnailer(input_path)
+        except ThumbnailError as e:
+            logger.warning(
+                "Unable to generate thumbnails for remote media %s from %s using a method of %s and type of %s: %s",
+                media_id,
+                server_name,
+                media_type,
+                e,
+            )
+            return None
+
         m_width = thumbnailer.width
         m_height = thumbnailer.height
 

synapse/rest/media/v1/thumbnail_resource.py

@@ -16,6 +16,7 @@
 
 import logging
 
+from synapse.api.errors import SynapseError
 from synapse.http.server import DirectServeJsonResource, set_cors_headers
 from synapse.http.servlet import parse_integer, parse_string
 
@@ -173,7 +174,7 @@ class ThumbnailResource(DirectServeJsonResource):
             await respond_with_file(request, desired_type, file_path)
         else:
             logger.warning("Failed to generate thumbnail")
-            respond_404(request)
+            raise SynapseError(400, "Failed to generate thumbnail.")
 
     async def _select_or_generate_remote_thumbnail(
         self,
@@ -235,7 +236,7 @@ class ThumbnailResource(DirectServeJsonResource):
             await respond_with_file(request, desired_type, file_path)
         else:
             logger.warning("Failed to generate thumbnail")
-            respond_404(request)
+            raise SynapseError(400, "Failed to generate thumbnail.")
 
     async def _respond_remote_thumbnail(
         self, request, server_name, media_id, width, height, method, m_type

synapse/rest/media/v1/thumbnailer.py

@@ -15,7 +15,7 @@
 import logging
 from io import BytesIO
 
-from PIL import Image as Image
+from PIL import Image
 
 logger = logging.getLogger(__name__)
 
@@ -31,12 +31,22 @@ EXIF_TRANSPOSE_MAPPINGS = {
 }
 
 
+class ThumbnailError(Exception):
+    """An error occurred generating a thumbnail."""
+
+
 class Thumbnailer:
 
     FORMATS = {"image/jpeg": "JPEG", "image/png": "PNG"}
 
     def __init__(self, input_path):
-        self.image = Image.open(input_path)
+        try:
+            self.image = Image.open(input_path)
+        except OSError as e:
+            # If an error occurs opening the image, a thumbnail won't be able to
+            # be generated.
+            raise ThumbnailError from e
+
         self.width, self.height = self.image.size
         self.transpose_method = None
         try:

tests/rest/media/v1/test_media_storage.py

@@ -120,12 +120,13 @@ class _TestImage:
     extension = attr.ib(type=bytes)
     expected_cropped = attr.ib(type=Optional[bytes])
     expected_scaled = attr.ib(type=Optional[bytes])
+    expected_found = attr.ib(default=True, type=bool)
 
 
 @parameterized_class(
     ("test_image",),
     [
-        # smol png
+        # smoll png
         (
             _TestImage(
                 unhexlify(
@@ -161,6 +162,8 @@ class _TestImage:
                 None,
             ),
         ),
+        # an empty file
+        (_TestImage(b"", b"image/gif", b".gif", None, None, False,),),
     ],
 )
 class MediaRepoTests(unittest.HomeserverTestCase):
@@ -303,12 +306,16 @@ class MediaRepoTests(unittest.HomeserverTestCase):
         self.assertEqual(headers.getRawHeaders(b"Content-Disposition"), None)
 
     def test_thumbnail_crop(self):
-        self._test_thumbnail("crop", self.test_image.expected_cropped)
+        self._test_thumbnail(
+            "crop", self.test_image.expected_cropped, self.test_image.expected_found
+        )
 
     def test_thumbnail_scale(self):
-        self._test_thumbnail("scale", self.test_image.expected_scaled)
+        self._test_thumbnail(
+            "scale", self.test_image.expected_scaled, self.test_image.expected_found
+        )
 
-    def _test_thumbnail(self, method, expected_body):
+    def _test_thumbnail(self, method, expected_body, expected_found):
         params = "?width=32&height=32&method=" + method
         request, channel = self.make_request(
             "GET", self.media_id + params, shorthand=False
@@ -325,11 +332,23 @@ class MediaRepoTests(unittest.HomeserverTestCase):
         )
         self.pump()
 
-        self.assertEqual(channel.code, 200)
-        if expected_body is not None:
-            self.assertEqual(
-                channel.result["body"], expected_body, channel.result["body"]
-            )
+        if expected_found:
+            self.assertEqual(channel.code, 200)
+            if expected_body is not None:
+                self.assertEqual(
+                    channel.result["body"], expected_body, channel.result["body"]
+                )
+            else:
+                # ensure that the result is at least some valid image
+                Image.open(BytesIO(channel.result["body"]))
         else:
-            # ensure that the result is at least some valid image
-            Image.open(BytesIO(channel.result["body"]))
+            # A 404 with a JSON body.
+            self.assertEqual(channel.code, 404)
+            self.assertEqual(
+                channel.json_body,
+                {
+                    "errcode": "M_NOT_FOUND",
+                    "error": "Not found [b'example.com', b'12345?width=32&height=32&method=%s']"
+                    % method,
+                },
+            )