Fix: read non-utf8 archives index

app/scodoc/sco_archives.py

@@ -47,6 +47,7 @@
  qui est une description (humaine, format libre) de l'archive.
 
 """
+import chardet
 import datetime
 import glob
 import mimetypes
@@ -203,8 +204,16 @@ class BaseArchiver(object):
     def get_archive_description(self, archive_id):
         """Return description of archive"""
         self.initialize()
-        with open(os.path.join(archive_id, "_description.txt")) as f:
-            descr = f.read()
+        filename = os.path.join(archive_id, "_description.txt")
+        try:
+            with open(filename) as f:
+                descr = f.read()
+        except UnicodeDecodeError:
+            # some (old) files may have saved under exotic encodings
+            with open(filename, "rb") as f:
+                data = f.read()
+                descr = data.decode(chardet.detect(data)["encoding"])
+
         return descr
 
     def create_obj_archive(self, oid: int, description: str):