rohmu: fix Range header syntax for S3 implementation

HTTP specs specify = as separator between unit and the ranges, not space

rohmu/object_storage/s3.py

@@ -322,7 +322,7 @@ def _get_object_stream(self, key: str, byte_range: Optional[tuple[int, int]]) ->
         path = self.format_key_for_backend(key, remove_slash_prefix=True)
         kwargs: dict[str, Any] = {}
         if byte_range:
-            kwargs["Range"] = f"bytes {byte_range[0]}-{byte_range[1]}"
+            kwargs["Range"] = f"bytes={byte_range[0]}-{byte_range[1]}"
         try:
             # Actual usage is accounted for in
             # _read_object_to_fileobj, although that omits the initial

test/test_object_storage_s3.py

@@ -1,4 +1,5 @@
 """Copyright (c) 2022 Aiven, Helsinki, Finland. https://aiven.io/"""
+from botocore.response import StreamingBody
 from dataclasses import dataclass
 from datetime import datetime
 from io import BytesIO
@@ -115,3 +116,20 @@ def test_get_contents_to_fileobj_raises_error_on_invalid_byte_range(infra: S3Inf
             fileobj_to_store_to=BytesIO(),
             byte_range=(100, 10),
         )
+
+
+def test_get_contents_to_fileobj_passes_the_correct_range_header(infra: S3Infra) -> None:
+    transfer = infra.transfer
+    infra.s3_client.get_object.return_value = {
+        "Body": StreamingBody(BytesIO(b"value"), 5),
+        "ContentLength": 5,
+        "Metadata": {},
+    }
+    transfer.get_contents_to_fileobj(
+        key="test_key",
+        fileobj_to_store_to=BytesIO(),
+        byte_range=(10, 100),
+    )
+    infra.s3_client.get_object.assert_called_once_with(
+        Bucket="test-bucket", Key="test-prefix/test_key", Range="bytes=10-100"
+    )