Changed decode_string to always return unicode.

deluge/common.py

@@ -608,22 +608,34 @@ def xml_encode(string):
 
 def decode_string(s, encoding="utf8"):
     """
-    Decodes a string and re-encodes it in utf8.  If it cannot decode using
-    `:param:encoding` then it will try to detect the string encoding and
-    decode it.
+    Decodes a string and return unicode. If it cannot decode using
+    `:param:encoding` then it will try latin1, and if that fails,
+    try to detect the string encoding. If that fails, decode with
+    ignore.
 
     :param s: string to decode
     :type s: string
     :keyword encoding: the encoding to use in the decoding
     :type encoding: string
+    :returns: s converted to unicode
+    :rtype: unicode
 
     """
+    if not s:
+        return u''
+    elif isinstance(s, unicode):
+        return s
 
-    try:
-        s = s.decode(encoding).encode("utf8", "ignore")
-    except UnicodeDecodeError:
-        s = s.decode(chardet.detect(s)["encoding"], "ignore").encode("utf8", "ignore")
-    return s
+    encodings = [(encoding, 'strict'), ("utf8", 'strict'),
+                 ("iso-8859-1", 'strict'),
+                 (chardet.detect(s)["encoding"], 'strict'),
+                 (chardet.detect(s)["encoding"], 'ignore')]
+    for i in range(len(encodings)):
+        try:
+            return s.decode(encodings[i][0], encodings[i][1])
+        except UnicodeDecodeError:
+            pass
+    return u''
 
 def utf8_encoded(s):
     """
@@ -636,7 +648,10 @@ def utf8_encoded(s):
 
     """
     if isinstance(s, str):
-        s = decode_string(s)
+        try:
+            s = decode_string(s).encode("utf8")
+        except UnicodeEncodeError:
+            log.warn("Error when encoding to utf8: %s" % s)
     elif isinstance(s, unicode):
         s = s.encode("utf8", "ignore")
     return s

deluge/ui/common.py

@@ -51,7 +51,7 @@ except ImportError:
     from sha import sha
 
 from deluge import bencode
-from deluge.common import decode_string, path_join
+from deluge.common import utf8_encoded, path_join
 import deluge.configmanager
 
 log = logging.getLogger(__name__)
@@ -88,9 +88,9 @@ class TorrentInfo(object):
         # Check if 'name.utf-8' is in the torrent and if not try to decode the string
         # using the encoding found.
         if "name.utf-8" in self.__m_metadata["info"]:
-            self.__m_name = decode_string(self.__m_metadata["info"]["name.utf-8"])
+            self.__m_name = utf8_encoded(self.__m_metadata["info"]["name.utf-8"])
         else:
-            self.__m_name = decode_string(self.__m_metadata["info"]["name"], self.encoding)
+            self.__m_name = utf8_encoded(self.__m_metadata["info"]["name"], self.encoding)
 
         # Get list of files from torrent info
         paths = {}
@@ -104,7 +104,7 @@ class TorrentInfo(object):
                 if "path.utf-8" in f:
                     path = os.path.join(prefix, *f["path.utf-8"])
                 else:
-                    path = decode_string(os.path.join(prefix, decode_string(os.path.join(*f["path"]), self.encoding)), self.encoding)
+                    path = utf8_encoded(os.path.join(prefix, utf8_encoded(os.path.join(*f["path"]), self.encoding)), self.encoding)
                 f["index"] = index
                 paths[path] = f
 
@@ -160,7 +160,7 @@ class TorrentInfo(object):
                 if "path.utf-8" in f:
                     path = os.path.join(prefix, *f["path.utf-8"])
                 else:
-                    path = decode_string(os.path.join(prefix, decode_string(os.path.join(*f["path"]), self.encoding)), self.encoding)
+                    path = utf8_encoded(os.path.join(prefix, utf8_encoded(os.path.join(*f["path"]), self.encoding)), self.encoding)
                 self.__m_files.append({
                     'path': path,
                     'size': f["length"],