detect_encoding now test utf-8 first

validata_core/source_helpers.py

@@ -14,13 +14,8 @@ def build_tabulator_params(type, name, source):
         # CSV: converts to string
         if format == 'csv':
             scheme = 'text'
-            # Try UTF-8 first, then fallback to detect_encoding,
-            # because detection sometimes fails and returns cp1252 instead of utf-8.
-            try:
-                source = source.decode("utf-8")
-            except UnicodeDecodeError:
-                encoding = helpers.detect_encoding(source)
-                source = source.decode(encoding)
+            encoding = helpers.detect_encoding(source)
+            source = source.decode(encoding)
 
         # Else use custom BytesLoader
         else: