If an encoding is not specified, ReadBytesWrapper uses the file's enc…

…oding attribute, and then falls back to utf-8.

storages/utils.py

@@ -123,9 +123,19 @@ class ReadBytesWrapper:
     """
     A wrapper for a file-like object, that makes read() always returns bytes.
     """
-    def __init__(self, file, encoding="utf-8"):
+    def __init__(self, file, encoding=None):
+        """
+        :param file: The file-like object to wrap.
+        :param encoding: Specify the encoding to use when file.read() returns strings.
+            If not provided will default to file.encoding, of if that's not available,
+            to utf-8.
+        """
         self._file = file
-        self._encoding = encoding
+        self._encoding = (
+            encoding
+            or getattr(file, "encoding", None)
+            or "utf-8"
+        )
         self._needs_encoding = None
 
     def read(self, *args, **kwargs):

tests/test_files/windows-1252-encoded.txt

@@ -0,0 +1 @@
+���

tests/test_utils.py

@@ -1,5 +1,6 @@
 import datetime
 from io import BytesIO, StringIO, SEEK_CUR
+from os.path import dirname, join
 
 from django.conf import settings
 from django.core.exceptions import SuspiciousFileOperation
@@ -175,9 +176,36 @@ def test_with_string_file(self):
         self.assertEqual(2, file_wrapped.seek(0, SEEK_CUR))
         self.assertEqual(b"yz", file_wrapped.read())
 
-    def test_with_string_file_custom_encoding(self):
-        file = StringIO("\u2122\u20AC\u2030")
-        file_wrapped = utils.ReadBytesWrapper(file, encoding="windows-1252")
+    # I chose the characters ™€‰ for the following tests because they produce different
+    # bytes when encoding with utf-8 vs windows-1252 vs utf-16
 
-        # test read() returns custom encoding
-        self.assertEqual(b"\x99\x80\x89", file_wrapped.read())
+    def test_with_string_file_specified_encoding(self):
+        content = "\u2122\u20AC\u2030"
+        file = StringIO(content)
+        file_wrapped = utils.ReadBytesWrapper(file, encoding="utf-16")
+
+        # test read() returns specified encoding
+        self.assertEqual(file_wrapped.read(), content.encode("utf-16"))
+
+    def test_with_string_file_detect_encoding(self):
+        content = "\u2122\u20AC\u2030"
+        with open(
+            file=join(dirname(__file__), "test_files", "windows-1252-encoded.txt"),
+            mode="r",
+            encoding="windows-1252",
+        ) as file:
+            self.assertEqual(file.read(), content)
+            file.seek(0)
+
+            file_wrapped = utils.ReadBytesWrapper(file)
+
+            # test read() returns encoding detected from file object.
+            self.assertEqual(file_wrapped.read(), content.encode("windows-1252"))
+
+    def test_with_string_file_fallback_encoding(self):
+        content = "\u2122\u20AC\u2030"
+        file = StringIO(content)
+        file_wrapped = utils.ReadBytesWrapper(file)
+
+        # test read() returns fallback utf-8 encoding
+        self.assertEqual(file_wrapped.read(), content.encode("utf-8"))