gh-133956 fix bug where dataclass wouldn't detect ClassVar fields if ClassVar was re-exported from a module other than typing (#140541)

Co-authored-by: Dmitrii Zherbin <zherbin.dima@yandex.ru>

Author: taminomara

Lib/dataclasses.py

@@ -757,22 +757,16 @@ def _is_kw_only(a_type, dataclasses):
     return a_type is dataclasses.KW_ONLY
 
 
-def _is_type(annotation, cls, a_module, a_type, is_type_predicate):
-    # Given a type annotation string, does it refer to a_type in
-    # a_module?  For example, when checking that annotation denotes a
-    # ClassVar, then a_module is typing, and a_type is
-    # typing.ClassVar.
+def _get_type_from_annotation(annotation, cls):
+    # Loosely parse a string annotation and return its type.
 
-    # It's possible to look up a_module given a_type, but it involves
-    # looking in sys.modules (again!), and seems like a waste since
-    # the caller already knows a_module.
+    # We can't perform a full type hint evaluation at the point where @dataclass
+    # was invoked because class's module is not fully initialized yet. So we resort
+    # to parsing string annotation using regexp, and extracting a type before
+    # the first square bracket.
 
     # - annotation is a string type annotation
     # - cls is the class that this annotation was found in
-    # - a_module is the module we want to match
-    # - a_type is the type in that module we want to match
-    # - is_type_predicate is a function called with (obj, a_module)
-    #   that determines if obj is of the desired type.
 
     # Since this test does not do a local namespace lookup (and
     # instead only a module (global) lookup), there are some things it
@@ -803,24 +797,21 @@ def _is_type(annotation, cls, a_module, a_type, is_type_predicate):
     # https://github.com/python/cpython/issues/77634 for details.
     global _MODULE_IDENTIFIER_RE
     if _MODULE_IDENTIFIER_RE is None:
-        _MODULE_IDENTIFIER_RE = re.compile(r'(?:\s*(\w+)\s*\.)?\s*(\w+)')
+        _MODULE_IDENTIFIER_RE = re.compile(r'^\s*(\w+(?:\s*\.\s*\w+)*)')
 
     match = _MODULE_IDENTIFIER_RE.prefixmatch(annotation)
-    if match:
-        ns = None
-        module_name = match[1]
-        if not module_name:
-            # No module name, assume the class's module did
-            # "from dataclasses import InitVar".
-            ns = sys.modules.get(cls.__module__).__dict__
-        else:
-            # Look up module_name in the class's module.
-            module = sys.modules.get(cls.__module__)
-            if module and module.__dict__.get(module_name) is a_module:
-                ns = sys.modules.get(a_type.__module__).__dict__
-        if ns and is_type_predicate(ns.get(match[2]), a_module):
-            return True
-    return False
+    if not match:
+        return None
+
+    # Note: _MODULE_IDENTIFIER_RE guarantees that path is non-empty
+    path = match[1].split(".")
+    root = sys.modules.get(cls.__module__)
+    for path_item in path:
+        root = getattr(root, path_item.strip(), None)
+        if root is None:
+            return None
+
+    return root
 
 
 def _get_field(cls, a_name, a_type, default_kw_only):
@@ -858,17 +849,18 @@ def _get_field(cls, a_name, a_type, default_kw_only):
     # is actually of the correct type.
 
     # For the complete discussion, see https://bugs.python.org/issue33453
+    if isinstance(a_type, str):
+        a_type_annotation = _get_type_from_annotation(a_type, cls)
+    else:
+        a_type_annotation = a_type
 
     # If typing has not been imported, then it's impossible for any
     # annotation to be a ClassVar.  So, only look for ClassVar if
     # typing has been imported by any module (not necessarily cls's
     # module).
     typing = sys.modules.get('typing')
     if typing:
-        if (_is_classvar(a_type, typing)
-            or (isinstance(f.type, str)
-                and _is_type(f.type, cls, typing, typing.ClassVar,
-                             _is_classvar))):
+        if _is_classvar(a_type_annotation, typing):
             f._field_type = _FIELD_CLASSVAR
 
     # If the type is InitVar, or if it's a matching string annotation,
@@ -877,10 +869,7 @@ def _get_field(cls, a_name, a_type, default_kw_only):
         # The module we're checking against is the module we're
         # currently in (dataclasses.py).
         dataclasses = sys.modules[__name__]
-        if (_is_initvar(a_type, dataclasses)
-            or (isinstance(f.type, str)
-                and _is_type(f.type, cls, dataclasses, dataclasses.InitVar,
-                             _is_initvar))):
+        if _is_initvar(a_type_annotation, dataclasses):
             f._field_type = _FIELD_INITVAR
 
     # Validations for individual fields.  This is delayed until now,
@@ -1073,10 +1062,11 @@ def _process_class(cls, init, repr, eq, order, unsafe_hash, frozen,
     dataclasses = sys.modules[__name__]
     for name, type in cls_annotations.items():
         # See if this is a marker to change the value of kw_only.
-        if (_is_kw_only(type, dataclasses)
-            or (isinstance(type, str)
-                and _is_type(type, cls, dataclasses, dataclasses.KW_ONLY,
-                             _is_kw_only))):
+        if isinstance(type, str):
+            a_type_annotation = _get_type_from_annotation(type, cls)
+        else:
+            a_type_annotation = type
+        if _is_kw_only(a_type_annotation, dataclasses):
             # Switch the default to kw_only=True, and ignore this
             # annotation: it's not a real field.
             if KW_ONLY_seen:

Lib/test/test_dataclasses/__init__.py

@@ -4353,10 +4353,17 @@ def test_classvar_module_level_import(self):
         from test.test_dataclasses import dataclass_module_1_str
         from test.test_dataclasses import dataclass_module_2
         from test.test_dataclasses import dataclass_module_2_str
-
-        for m in (dataclass_module_1, dataclass_module_1_str,
-                  dataclass_module_2, dataclass_module_2_str,
-                  ):
+        from test.test_dataclasses import dataclass_module_3
+        from test.test_dataclasses import dataclass_module_3_str
+        from test.test_dataclasses import dataclass_module_4
+        from test.test_dataclasses import dataclass_module_4_str
+
+        for m in (
+            dataclass_module_1, dataclass_module_1_str,
+            dataclass_module_2, dataclass_module_2_str,
+            dataclass_module_3, dataclass_module_3_str,
+            dataclass_module_4, dataclass_module_4_str,
+        ):
             with self.subTest(m=m):
                 # There's a difference in how the ClassVars are
                 # interpreted when using string annotations or
@@ -4660,6 +4667,14 @@ def custom_dataclass(cls, *args, **kwargs):
         self.assertEqual(c.x, 10)
         self.assertEqual(c.__custom__, True)
 
+    def test_empty_annotation_string(self):
+        @dataclass
+        class DataclassWithEmptyTypeAnnotation:
+            x: ""
+
+        c = DataclassWithEmptyTypeAnnotation(10)
+        self.assertEqual(c.x, 10)
+
 
 class TestReplace(unittest.TestCase):
     def test(self):

Lib/test/test_dataclasses/_types_proxy.py

@@ -0,0 +1,8 @@
+# We need this to test a case when a type
+# is imported via some other package,
+# like ClassVar from typing_extensions instead of typing.
+# https://github.com/python/cpython/issues/133956
+from typing import ClassVar
+from dataclasses import InitVar
+
+__all__ = ["ClassVar", "InitVar"]

Lib/test/test_dataclasses/dataclass_module_3.py

@@ -0,0 +1,32 @@
+#from __future__ import annotations
+USING_STRINGS = False
+
+# dataclass_module_3.py and dataclass_module_3_str.py are identical
+# except only the latter uses string annotations.
+
+from dataclasses import dataclass
+import test.test_dataclasses._types_proxy as tp
+
+T_CV2 = tp.ClassVar[int]
+T_CV3 = tp.ClassVar
+
+T_IV2 = tp.InitVar[int]
+T_IV3 = tp.InitVar
+
+@dataclass
+class CV:
+    T_CV4 = tp.ClassVar
+    cv0: tp.ClassVar[int] = 20
+    cv1: tp.ClassVar = 30
+    cv2: T_CV2
+    cv3: T_CV3
+    not_cv4: T_CV4  # When using string annotations, this field is not recognized as a ClassVar.
+
+@dataclass
+class IV:
+    T_IV4 = tp.InitVar
+    iv0: tp.InitVar[int]
+    iv1: tp.InitVar
+    iv2: T_IV2
+    iv3: T_IV3
+    not_iv4: T_IV4  # When using string annotations, this field is not recognized as an InitVar.

Lib/test/test_dataclasses/dataclass_module_3_str.py

@@ -0,0 +1,32 @@
+from __future__ import annotations
+USING_STRINGS = True
+
+# dataclass_module_3.py and dataclass_module_3_str.py are identical
+# except only the latter uses string annotations.
+
+from dataclasses import dataclass
+import test.test_dataclasses._types_proxy as tp
+
+T_CV2 = tp.ClassVar[int]
+T_CV3 = tp.ClassVar
+
+T_IV2 = tp.InitVar[int]
+T_IV3 = tp.InitVar
+
+@dataclass
+class CV:
+    T_CV4 = tp.ClassVar
+    cv0: tp.ClassVar[int] = 20
+    cv1: tp.ClassVar = 30
+    cv2: T_CV2
+    cv3: T_CV3
+    not_cv4: T_CV4  # When using string annotations, this field is not recognized as a ClassVar.
+
+@dataclass
+class IV:
+    T_IV4 = tp.InitVar
+    iv0: tp.InitVar[int]
+    iv1: tp.InitVar
+    iv2: T_IV2
+    iv3: T_IV3
+    not_iv4: T_IV4  # When using string annotations, this field is not recognized as an InitVar.

Lib/test/test_dataclasses/dataclass_module_4.py

@@ -0,0 +1,38 @@
+#from __future__ import annotations
+USING_STRINGS = False
+
+# dataclass_module_4.py and dataclass_module_4_str.py are identical
+# except only the latter uses string annotations.
+
+from dataclasses import dataclass
+import dataclasses
+import typing
+
+class TypingProxy:
+    class Nested:
+        ClassVar = typing.ClassVar
+        InitVar = dataclasses.InitVar
+
+T_CV2 = TypingProxy.Nested.ClassVar[int]
+T_CV3 = TypingProxy.Nested.ClassVar
+
+T_IV2 = TypingProxy.Nested.InitVar[int]
+T_IV3 = TypingProxy.Nested.InitVar
+
+@dataclass
+class CV:
+    T_CV4 = TypingProxy.Nested.ClassVar
+    cv0: TypingProxy.Nested.ClassVar[int] = 20
+    cv1: TypingProxy.Nested.ClassVar = 30
+    cv2: T_CV2
+    cv3: T_CV3
+    not_cv4: T_CV4  # When using string annotations, this field is not recognized as a ClassVar.
+
+@dataclass
+class IV:
+    T_IV4 = TypingProxy.Nested.InitVar
+    iv0: TypingProxy.Nested.InitVar[int]
+    iv1: TypingProxy.Nested.InitVar
+    iv2: T_IV2
+    iv3: T_IV3
+    not_iv4: T_IV4  # When using string annotations, this field is not recognized as an InitVar.

Lib/test/test_dataclasses/dataclass_module_4_str.py

@@ -0,0 +1,38 @@
+from __future__ import annotations
+USING_STRINGS = True
+
+# dataclass_module_4.py and dataclass_module_4_str.py are identical
+# except only the latter uses string annotations.
+
+from dataclasses import dataclass
+import dataclasses
+import typing
+
+class TypingProxy:
+    class Nested:
+        ClassVar = typing.ClassVar
+        InitVar = dataclasses.InitVar
+
+T_CV2 = TypingProxy.Nested.ClassVar[int]
+T_CV3 = TypingProxy.Nested.ClassVar
+
+T_IV2 = TypingProxy.Nested.InitVar[int]
+T_IV3 = TypingProxy.Nested.InitVar
+
+@dataclass
+class CV:
+    T_CV4 = TypingProxy.Nested.ClassVar
+    cv0: TypingProxy.Nested.ClassVar[int] = 20
+    cv1: TypingProxy.Nested.ClassVar = 30
+    cv2: T_CV2
+    cv3: T_CV3
+    not_cv4: T_CV4  # When using string annotations, this field is not recognized as a ClassVar.
+
+@dataclass
+class IV:
+    T_IV4 = TypingProxy.Nested.InitVar
+    iv0: TypingProxy.Nested.InitVar[int]
+    iv1: TypingProxy.Nested.InitVar
+    iv2: T_IV2
+    iv3: T_IV3
+    not_iv4: T_IV4  # When using string annotations, this field is not recognized as an InitVar.

Misc/NEWS.d/next/Library/2025-05-16-01-43-58.gh-issue-133956.5kWDYd.rst

@@ -0,0 +1,4 @@
+Fix bug where :func:`@dataclass <dataclasses.dataclass>`
+wouldn't detect ``ClassVar`` fields
+if ``ClassVar`` was re-exported from a module
+other than :mod:`typing`.