Improve skipping encoding part from language code

We can not blidly ignore anything after ., we have to do at least basic
check that it looks like an encoding.

Issue #27
Issue #760
Signed-off-by: Michal Čihař <michal@cihar.com>

weblate/trans/models/subproject.py

@@ -966,7 +966,9 @@ class SubProject(models.Model, PercentMixin, URLMixin, PathMixin):
         else:
             code = path[len(parts[0]):-len(parts[1])].split('/')[0]
         # Remove possible encoding part
-        return code.split('.')[0]
+        if '.' in code and ('.utf' in code.lower() or '.iso' in code.lower()):
+            return code.split('.')[0]
+        return code
 
     def sync_git_repo(self, validate=False):
         '''