Raise AttributeError on attempts to access unset oneof fields (da…

…nielgtaylor#510)

# Conflicts:
#	poetry.lock
#	src/betterproto/__init__.py
#	src/betterproto/plugin/parser.py

# Conflicts:
#	poetry.lock
#	pyproject.toml

.pre-commit-config.yaml

@@ -8,9 +8,10 @@ repos:
       - id: isort
 
   - repo: https://github.com/psf/black
-    rev: 22.3.0
+    rev: 23.1.0
     hooks:
       - id: black
+        args: ["--target-version", "py310"]
 
   - repo: https://github.com/PyCQA/doc8
     rev: 0.10.1

pyproject.toml

@@ -13,7 +13,7 @@ packages = [
 
 [tool.poetry.dependencies]
 python = "^3.7"
-black = { version = ">=19.3b0", optional = true }
+black = { version = ">=23.1.0", optional = true }
 dataclasses = { version = "^0.7", python = ">=3.6, <3.7" }
 grpcio = "^1.43.0"
 importlib-metadata = { version = ">=1.6.0", python = "<3.8" }
@@ -62,7 +62,7 @@ cmd  = "mypy src --ignore-missing-imports"
 help = "Check types with mypy"
 
 [tool.poe.tasks.format]
-cmd  = "black . --exclude tests/output_"
+cmd  = "black . --exclude tests/output_ --target-version py310"
 help = "Apply black formatting to source code"
 
 [tool.poe.tasks.docs]

src/betterproto/__init__.py

@@ -693,8 +693,28 @@ def __repr__(self) -> str:
         def __getattribute__(self, name: str) -> Any:
             """
             Lazily initialize default values to avoid infinite recursion for recursive
-            message types
+            message types.
+            Raise :class:`AttributeError` on attempts to access unset ``oneof`` fields.
             """
+            try:
+                group_current = super().__getattribute__("_group_current")
+            except AttributeError:
+                pass
+            else:
+                if name not in {"__class__", "_betterproto"}:
+                    group = self._betterproto.oneof_group_by_field.get(name)
+                    if group is not None and group_current[group] != name:
+                        if sys.version_info < (3, 10):
+                            raise AttributeError(
+                                f"{group!r} is set to {group_current[group]!r}, not {name!r}"
+                            )
+                        else:
+                            raise AttributeError(
+                                f"{group!r} is set to {group_current[group]!r}, not {name!r}",
+                                name=name,
+                                obj=self,
+                            )
+
             value = super().__getattribute__(name)
             if value is not PLACEHOLDER:
                 return value
@@ -761,7 +781,10 @@ def __bytes__(self) -> bytes:
         """
         output = bytearray()
         for field_name, meta in self._betterproto.meta_by_field_name.items():
-            value = getattr(self, field_name)
+            try:
+                value = getattr(self, field_name)
+            except AttributeError:
+                continue
 
             if value is None:
                 # Optional items should be skipped. This is used for the Google
@@ -775,9 +798,7 @@ def __bytes__(self) -> bytes:
             # Note that proto3 field presence/optional fields are put in a
             # synthetic single-item oneof by protoc, which helps us ensure we
             # send the value even if the value is the default zero value.
-            selected_in_group = (
-                meta.group and self._group_current[meta.group] == field_name
-            )
+            selected_in_group = bool(meta.group)
 
             # Empty messages can still be sent on the wire if they were
             # set (or received empty).
@@ -1016,7 +1037,12 @@ def parse(self: T, data: bytes) -> T:
                     parsed.wire_type, meta, field_name, parsed.value
                 )
 
-            current = getattr(self, field_name)
+            try:
+                current = getattr(self, field_name)
+            except AttributeError:
+                current = self._get_field_default(field_name)
+                setattr(self, field_name, current)
+
             if meta.proto_type == TYPE_MAP:
                 # Value represents a single key/value pair entry in the map.
                 current[value.key] = value.value
@@ -1077,7 +1103,10 @@ def to_dict(
         defaults = self._betterproto.default_gen
         for field_name, meta in self._betterproto.meta_by_field_name.items():
             field_is_repeated = defaults[field_name] is list
-            value = getattr(self, field_name)
+            try:
+                value = getattr(self, field_name)
+            except AttributeError:
+                value = self._get_field_default(field_name)
             cased_name = casing(field_name).rstrip("_")  # type: ignore
             if meta.proto_type == TYPE_MESSAGE:
                 if isinstance(value, datetime):
@@ -1209,7 +1238,7 @@ def from_dict(self: T, value: Mapping[str, Any]) -> T:
 
             if value[key] is not None:
                 if meta.proto_type == TYPE_MESSAGE:
-                    v = getattr(self, field_name)
+                    v = self._get_field_default(field_name)
                     cls = self._betterproto.cls_by_field[field_name]
                     if isinstance(v, list):
                         if cls == datetime:
@@ -1480,6 +1509,34 @@ def is_set(self, name: str) -> bool:
         )
         return self.__raw_get(name) is not default
 
+    @classmethod
+    def _validate_field_groups(cls, values):
+        group_to_one_ofs = cls._betterproto_meta.oneof_field_by_group  # type: ignore
+        field_name_to_meta = cls._betterproto_meta.meta_by_field_name  # type: ignore
+
+        for group, field_set in group_to_one_ofs.items():
+            if len(field_set) == 1:
+                (field,) = field_set
+                field_name = field.name
+                meta = field_name_to_meta[field_name]
+
+                # This is a synthetic oneof; we should ignore it's presence and not consider it as a oneof.
+                if meta.optional:
+                    continue
+
+            set_fields = [
+                field.name for field in field_set if values[field.name] is not None
+            ]
+
+            if not set_fields:
+                raise ValueError(f"Group {group} has no value; all fields are None")
+            elif len(set_fields) > 1:
+                set_fields_str = ", ".join(set_fields)
+                raise ValueError(
+                    f"Group {group} has more than one value; fields {set_fields_str} are not None"
+                )
+
+        return values
 
 def serialized_on_wire(message: Message) -> bool:
     """

src/betterproto/grpc/grpclib_server.py

@@ -21,7 +21,6 @@ async def _call_rpc_handler_server_stream(
         stream: grpclib.server.Stream,
         request: Any,
     ) -> None:
-
         response_iter = handler(request)
         # check if response is actually an AsyncIterator
         # this might be false if the method just returns without

src/betterproto/plugin/compiler.py

@@ -21,7 +21,6 @@
 
 
 def outputfile_compiler(output_file: OutputTemplate) -> str:
-
     templates_folder = os.path.abspath(
         os.path.join(os.path.dirname(__file__), "..", "templates")
     )

tests/inputs/google_impl_behavior_equivalence/test_google_impl_behavior_equivalence.py

@@ -50,8 +50,10 @@ def test_bytes_are_the_same_for_oneof():
 
     # None of these fields were explicitly set BUT they should not actually be null
     # themselves
-    assert isinstance(message.foo, Foo)
-    assert isinstance(message2.foo, Foo)
+    assert not hasattr(message, "foo")
+    assert object.__getattribute__(message, "foo") == betterproto.PLACEHOLDER
+    assert not hasattr(message2, "foo")
+    assert object.__getattribute__(message2, "foo") == betterproto.PLACEHOLDER
 
     assert isinstance(message_reference.foo, ReferenceFoo)
     assert isinstance(message_reference2.foo, ReferenceFoo)

tests/inputs/oneof_enum/test_oneof_enum.py

@@ -18,9 +18,8 @@ def test_which_one_of_returns_enum_with_default_value():
         get_test_case_json_data("oneof_enum", "oneof_enum-enum-0.json")[0].json
     )
 
-    assert message.move == Move(
-        x=0, y=0
-    )  # Proto3 will default this as there is no null
+    assert not hasattr(message, "move")
+    assert object.__getattribute__(message, "move") == betterproto.PLACEHOLDER
     assert message.signal == Signal.PASS
     assert betterproto.which_one_of(message, "action") == ("signal", Signal.PASS)
 
@@ -33,9 +32,8 @@ def test_which_one_of_returns_enum_with_non_default_value():
     message.from_json(
         get_test_case_json_data("oneof_enum", "oneof_enum-enum-1.json")[0].json
     )
-    assert message.move == Move(
-        x=0, y=0
-    )  # Proto3 will default this as there is no null
+    assert not hasattr(message, "move")
+    assert object.__getattribute__(message, "move") == betterproto.PLACEHOLDER
     assert message.signal == Signal.RESIGN
     assert betterproto.which_one_of(message, "action") == ("signal", Signal.RESIGN)
 
@@ -44,5 +42,6 @@ def test_which_one_of_returns_second_field_when_set():
     message = Test()
     message.from_json(get_test_case_json_data("oneof_enum")[0].json)
     assert message.move == Move(x=2, y=3)
-    assert message.signal == Signal.PASS
+    assert not hasattr(message, "signal")
+    assert object.__getattribute__(message, "signal") == betterproto.PLACEHOLDER
     assert betterproto.which_one_of(message, "action") == ("move", Move(x=2, y=3))

tests/oneof_pattern_matching.py

@@ -0,0 +1,46 @@
+from dataclasses import dataclass
+
+import pytest
+
+import betterproto
+
+
+def test_oneof_pattern_matching():
+    @dataclass
+    class Sub(betterproto.Message):
+        val: int = betterproto.int32_field(1)
+
+    @dataclass
+    class Foo(betterproto.Message):
+        bar: int = betterproto.int32_field(1, group="group1")
+        baz: str = betterproto.string_field(2, group="group1")
+        sub: Sub = betterproto.message_field(3, group="group2")
+        abc: str = betterproto.string_field(4, group="group2")
+
+    foo = Foo(baz="test1", abc="test2")
+
+    match foo:
+        case Foo(bar=_):
+            pytest.fail("Matched 'bar' instead of 'baz'")
+        case Foo(baz=v):
+            assert v == "test1"
+        case _:
+            pytest.fail("Matched neither 'bar' nor 'baz'")
+
+    match foo:
+        case Foo(sub=_):
+            pytest.fail("Matched 'sub' instead of 'abc'")
+        case Foo(abc=v):
+            assert v == "test2"
+        case _:
+            pytest.fail("Matched neither 'sub' nor 'abc'")
+
+    foo.sub = Sub(val=1)
+
+    match foo:
+        case Foo(sub=Sub(val=v)):
+            assert v == 1
+        case Foo(abc=v):
+            pytest.fail("Matched 'abc' instead of 'sub'")
+        case _:
+            pytest.fail("Matched neither 'sub' nor 'abc'")

tests/test_features.py

@@ -1,4 +1,5 @@
 import json
+import sys
 from copy import (
     copy,
     deepcopy,
@@ -18,6 +19,8 @@
     Optional,
 )
 
+import pytest
+
 import betterproto
 
 
@@ -151,17 +154,18 @@ class Foo(betterproto.Message):
     foo.baz = "test"
 
     # Other oneof fields should now be unset
-    assert foo.bar == 0
+    assert not hasattr(foo, "bar")
+    assert object.__getattribute__(foo, "bar") == betterproto.PLACEHOLDER
     assert betterproto.which_one_of(foo, "group1")[0] == "baz"
 
-    foo.sub.val = 1
+    foo.sub = Sub(val=1)
     assert betterproto.serialized_on_wire(foo.sub)
 
     foo.abc = "test"
 
     # Group 1 shouldn't be touched, group 2 should have reset
-    assert foo.sub.val == 0
-    assert betterproto.serialized_on_wire(foo.sub) is False
+    assert not hasattr(foo, "sub")
+    assert object.__getattribute__(foo, "sub") == betterproto.PLACEHOLDER
     assert betterproto.which_one_of(foo, "group2")[0] == "abc"
 
     # Zero value should always serialize for one-of
@@ -176,6 +180,16 @@ class Foo(betterproto.Message):
     assert betterproto.which_one_of(foo2, "group2")[0] == ""
 
 
+@pytest.mark.skipif(
+    sys.version_info < (3, 10),
+    reason="pattern matching is only supported in python3.10+",
+)
+def test_oneof_pattern_matching():
+    from .oneof_pattern_matching import test_oneof_pattern_matching
+
+    test_oneof_pattern_matching()
+
+
 def test_json_casing():
     @dataclass
     class CasingTest(betterproto.Message):