Merge pull request luizalabs#40 from luizalabs/fix-json-serialization…

…-readme

Update serialization docs

README.md

@@ -49,14 +49,21 @@ We use [pickle](https://docs.python.org/3/library/pickle.html) as default to rea
 You can create a custom serializer by implementing the `dumps` and `loads` methods.
 
 ```python
+NULL_BYTE: Final = b"\x00"
+
+
 class JSONSerializer:
     def dumps(self, obj: dict) -> bytes:
-        return json.dumps(obj).encode()
+        return json.dumps(obj).encode() + NULL_BYTE
 
     def loads(self, data: bytes) -> dict:
+        data = data.split(NULL_BYTE, 1)[0]
         return json.loads(data)
+
 ```
 
+Note: A null byte is used to separate the dictionary contents from the bytes that are in memory.
+
 To use the custom serializer you must set it when creating a new shared memory dict instance:
 
 ```python

shared_memory_dict/serializers.py

@@ -18,7 +18,7 @@ def dumps(self, obj: dict) -> bytes:
         return json.dumps(obj).encode() + NULL_BYTE
 
     def loads(self, data: bytes) -> dict:
-        data = bytes(data).split(NULL_BYTE, 1)[0]
+        data = data.split(NULL_BYTE, 1)[0]
         return json.loads(data)