[Core] Work around libtorrent size limit

When loading a torrent_info by file path, libtorrent has a hardcoded
size limit that some torrents exceed. This means that the torrent can be
added normally to deluge, but on restart it fails to load the file from
the state directory and the torrent is lost.

Work around the hardcoded buffer size limit by reading in the file
directly and passing the buffer.

See: https://libtorrent-discuss.narkive.com/wYACQDb7/libtorrent-loading-a-big-torrent-fails
Fixes: https://forum.deluge-torrent.org/viewtopic.php?t=56856
Closes: #483

Author: marcan

deluge/core/torrentmanager.py

@@ -15,6 +15,7 @@
 import pickle
 import time
 from base64 import b64encode
+from pathlib import Path
 from tempfile import gettempdir
 from typing import Dict, List, NamedTuple, Tuple
 
@@ -338,8 +339,8 @@ def get_torrent_info_from_file(self, filepath):
         if log.isEnabledFor(logging.DEBUG):
             log.debug('Attempting to extract torrent_info from %s', filepath)
         try:
-            torrent_info = lt.torrent_info(filepath)
-        except RuntimeError as ex:
+            torrent_info = lt.torrent_info(Path(filepath).read_bytes())
+        except (RuntimeError, OSError) as ex:
             log.warning('Unable to open torrent file %s: %s', filepath, ex)
         else:
             return torrent_info