PR Feedback

ssz/constants.py

@@ -5,9 +5,6 @@
 CHUNK_SIZE = 32  # named BYTES_PER_CHUNK in the spec
 EMPTY_CHUNK = Hash32(b"\x00" * CHUNK_SIZE)
 
-SIZE_PREFIX_SIZE = 4  # named BYTES_PER_LENGTH_PREFIX in the spec
-MAX_CONTENT_SIZE = 2 ** (SIZE_PREFIX_SIZE * 8) - 1
-
 SIGNATURE_FIELD_NAME = "signature"
 
 # number of bytes for a serialized offset

ssz/sedes/base.py

@@ -3,7 +3,6 @@
     abstractmethod,
 )
 import io
-import itertools
 import operator
 from typing import (
     IO,
@@ -176,17 +175,3 @@ def deserialize(self, data: bytes) -> TDeserialized:
     @abstractmethod
     def _deserialize_stream(self, stream: IO[bytes]) -> TDeserialized:
         pass
-
-
-TSerializableElement = TypeVar("TSerializableElement")
-TDeserializedElement = TypeVar("TDeserializedElement")
-
-
-class HomogenousSequence(CompositeSedes[Sequence[Any], Tuple[Any]]):
-    element_sedes: TSedes
-
-    #
-    # Serialization
-    #
-    def _get_item_sedes_pairs(self, value: Sequence[Any]) -> Tuple[Tuple[Any, TSedes], ...]:
-        return tuple(zip(value, itertools.repeat(self.element_sedes)))

ssz/sedes/container.py

@@ -44,8 +44,6 @@ class Container(CompositeSedes[Sequence[Any], Tuple[Any, ...]]):
     def __init__(self, field_sedes: Sequence[TSedes]) -> None:
         if len(field_sedes) == 0:
             raise ValidationError("Cannot define container without any fields")
-        from ssz.sedes.base import BaseSedes
-        assert all(isinstance(sedes, BaseSedes) for sedes in field_sedes)
         self.field_sedes = tuple(field_sedes)
 
     #

ssz/sedes/list.py

@@ -40,7 +40,7 @@
 TSerializable = TypeVar("TSerializable")
 TDeserialized = TypeVar("TDeserialized")
 
-EMPTY_LIST_HASH_TRIES_ROOT = mix_in_length(merkleize(pack([])), 0)
+EMPTY_LIST_HASH_TREE_ROOT = mix_in_length(merkleize(pack([])), 0)
 
 
 class EmptyList(BaseCompositeSedes[Sequence[TSerializable], Tuple[TSerializable, ...]]):
@@ -61,8 +61,8 @@ def deserialize(self, data: bytes) -> Tuple[TDeserialized, ...]:
 
     def hash_tree_root(self, value: Sequence[TSerializable]) -> bytes:
         if len(value):
-            raise ValueError("Cannot compute trie hash for non-empty value using `EmptyList` sedes")
-        return EMPTY_LIST_HASH_TRIES_ROOT
+            raise ValueError("Cannot compute tree hash for non-empty value using `EmptyList` sedes")
+        return EMPTY_LIST_HASH_TREE_ROOT
 
 
 empty_list = EmptyList()
@@ -98,15 +98,24 @@ def _deserialize_stream(self, stream: IO[bytes]) -> Iterable[TDeserialized]:
             element_size = self.element_sedes.get_fixed_size()
             data = stream.read()
             if len(data) % element_size != 0:
-                raise DeserializationError("TODO: INVALID LENGTH")
+                raise DeserializationError(
+                    f"Invalid length. List is comprised of a fixed size sedes "
+                    f"but total serialized data is not an even multiple of the "
+                    f"element size. data length: {len(data)}  element size: "
+                    f"{element_size}"
+                )
             for segment in partition(element_size, data):
                 yield self.element_sedes.deserialize(segment)
         else:
+            stream_zero_loc = stream.tell()
             try:
                 first_offset = s_decode_offset(stream)
             except DeserializationError:
-                # Empty list
-                return
+                if stream.tell() == stream_zero_loc:
+                    # Empty list
+                    return
+                else:
+                    raise
 
             num_remaining_offset_bytes = first_offset - stream.tell()
             if num_remaining_offset_bytes % OFFSET_SIZE != 0:
@@ -134,7 +143,7 @@ def _deserialize_stream(self, stream: IO[bytes]) -> Iterable[TDeserialized]:
     #
     def hash_tree_root(self, value: Iterable[TSerializable]) -> bytes:
         if len(value) == 0:
-            return EMPTY_LIST_HASH_TRIES_ROOT
+            return EMPTY_LIST_HASH_TREE_ROOT
         elif isinstance(self.element_sedes, BasicSedes):
             serialized_items = tuple(
                 self.element_sedes.serialize(element)

ssz/utils.py

@@ -55,7 +55,7 @@ def decode_offset(data: bytes) -> int:
 
 
 def s_decode_offset(stream: IO[bytes]) -> int:
-    data = read_exact(4, stream)
+    data = read_exact(OFFSET_SIZE, stream)
     return decode_offset(data)
 
 

tests/sedes/test_composite_sedes.py

@@ -7,6 +7,9 @@
 import pytest
 
 import ssz
+from ssz.exceptions import (
+    DeserializationError,
+)
 from ssz.sedes import (
     Container,
     List,
@@ -31,6 +34,16 @@ def test_list(value, serialized):
     assert ssz.decode(decode_hex(serialized), sedes) == value
 
 
+def test_invalid_serialized_list():
+    # ensure that an improperly read offset (not enough bytes) does not
+    # incorrectly register as an empty list due to mis-interpreting the failed
+    # stream read as the stream having been empty.
+    data = decode_hex("0x0001")
+    sedes = List(List(uint8))
+    with pytest.raises(DeserializationError):
+        ssz.decode(data, sedes=sedes)
+
+
 @pytest.mark.parametrize(
     ("value", "serialized"),
     (
@@ -58,7 +71,7 @@ def test_tuple_of_static_sized_entries(value, serialized):
         ),
     )
 )
-def test_tuple_of_dynamic_sized_entries(value, serialized):
+def test_list_of_dynamic_sized_entries(value, serialized):
     sedes = Vector(List(uint8), len(value))
     assert encode_hex(ssz.encode(value, sedes)) == serialized
     assert ssz.decode(decode_hex(serialized), sedes) == value