Manually resolve string types into ForwardRef in __init__ of a da…

…taclass
sobolevn · Oct 22, 2021 · Oct 22, 2021 · Oct 22, 2021 · Oct 22, 2021 · Oct 22, 2021
commit 97a34960a875eb28b2745ea5c945c432a327102f
@@ -9,6 +9,7 @@
 import abc
 import _thread
 from types import FunctionType, GenericAlias
+from typing import ForwardRef
 
 
 __all__ = ['dataclass',
@@ -447,10 +448,24 @@ def _field_assign(frozen, name, value, self_name):
     return f'{self_name}.{name}={value}'
 
 
-def _field_init(f, frozen, globals, self_name, slots):
+def _field_init(f, frozen, globals, self_name, slots, module):
     # Return the text of the line in the body of __init__ that will
     # initialize this field.
 
+    if f.init and isinstance(f.type, str):
+        # We need to resolve this string type into a real `ForwardRef` object,
+        # because otherwise we might end up with unsolvable annotations.
+        # For example:
+        #   def __init__(self, d: collections.OrderedDict) -> None:
+        # We won't be able to resolve `collections.OrderedDict`
+        # with wrong `module` param, when placed in a different module. #45524
+        try:
+            f.type = ForwardRef(f.type, module=module, is_class=True)
+        except SyntaxError:
+            # We don't want to fail class creation
+            # when `ForwardRef` cannot be constructed.
+            pass
+
     default_name = f'_dflt_{f.name}'
     if f.default_factory is not MISSING:
         if f.init:
@@ -527,7 +542,7 @@ def _init_param(f):
 
 
 def _init_fn(fields, std_fields, kw_only_fields, frozen, has_post_init,
-             self_name, globals, slots):
+             self_name, globals, slots, module):
     # fields contains both real fields and InitVar pseudo-fields.
 
     # Make sure we don't have fields without defaults following fields
@@ -554,7 +569,7 @@ def _init_fn(fields, std_fields, kw_only_fields, frozen, has_post_init,
 
     body_lines = []
     for f in fields:
-        line = _field_init(f, frozen, locals, self_name, slots)
+        line = _field_init(f, frozen, locals, self_name, slots, module)
         # line is None means that this field doesn't require
         # initialization (it's a pseudo-field).  Just skip it.
         if line:
@@ -906,7 +921,6 @@ def _process_class(cls, init, repr, eq, order, unsafe_hash, frozen,
     # we're iterating over them, see if any are frozen.
     any_frozen_base = False
     has_dataclass_bases = False
-    init_globals = dict(globals)
 
     for b in cls.__mro__[-1:0:-1]:
         # Only process classes that have been processed by our
@@ -918,17 +932,6 @@ def _process_class(cls, init, repr, eq, order, unsafe_hash, frozen,
                 fields[f.name] = f
             if getattr(b, _PARAMS).frozen:
                 any_frozen_base = True
-        if has_dataclass_bases:
-            # If dataclass has other dataclass as a base type,
-            # it might have existing `__init__` method.
-            # We need to change its `globals`, because otherwise
-            # we might end up with unsolvable annotations. For example:
-            # `def __init__(self, d: collections.OrderedDict) -> None:`
-            # We won't be able to resolve `collections.OrderedDict`
-            # with wrong `globals`, when placed in a different module. #45524
-            super_init = getattr(b, '__init__', None)
-            if super_init is not None:
-                init_globals.update(getattr(super_init, '__globals__', {}))
 
     # Annotations that are defined in this class (not in base
     # classes).  If __annotations__ isn't present, then this class
@@ -1045,8 +1048,9 @@ def _process_class(cls, init, repr, eq, order, unsafe_hash, frozen,
                                     # if possible.
                                     '__dataclass_self__' if 'self' in fields
                                             else 'self',
+                                    globals,
                                     slots,
-                                    init_globals,
+                                    cls.__module__,
                           ))
 
     # Get the fields as a list, and include only real fields.  This is

diff --git a/Lib/test/dataclass_textanno2.py b/Lib/test/dataclass_textanno2.py
@@ -13,6 +13,15 @@ class Child(dataclass_textanno.Bar):
     custom: Custom
 
 
+class Foo:  # matching name with `dataclass_testanno.Foo`
+    pass
+
+
+@dataclasses.dataclass
+class WithMatchinNameOverride(dataclass_textanno.Bar):
+    foo: Foo  # we override existing `foo: Foo` with
+
+
 @dataclasses.dataclass(init=False)
 class WithFutureInit(Child):
     def __init__(self, foo: dataclass_textanno.Foo, custom: Custom) -> None:

diff --git a/Lib/test/test_dataclasses.py b/Lib/test/test_dataclasses.py
@@ -3221,6 +3221,37 @@ class FutureInitChild(dataclass_textanno2.WithFutureInit):
                     },
                 )
 
+    def test_dataclass_proxy_modules_matching_name_override(self):
+        # see bpo-45524
+        from test import dataclass_textanno2
+        from dataclasses import dataclass
+
+        @dataclass
+        class Default(dataclass_textanno2.WithMatchinNameOverride):
+            pass
+
+        classes = [
+            Default,
+            dataclass_textanno2.WithMatchinNameOverride
+        ]
+        for klass in classes:
+            with self.subTest(klass=klass):
+                self.assertEqual(
+                    get_type_hints(klass),
+                    {
+                        'foo': dataclass_textanno2.Foo,
+                    },
+                )
+                self.assertEqual(get_type_hints(klass.__new__), {})
+                self.assertEqual(
+                    get_type_hints(klass.__init__),
+                    {
+                        'foo': dataclass_textanno2.Foo,
+                        'return': type(None),
+                    },
+                )
+
+
 
 class TestMakeDataclass(unittest.TestCase):
     def test_simple(self):