Sync setuptools schema with validate-pyproject (#5157)

Author: abravalheri

newsfragments/5157.misc.rst

@@ -0,0 +1,2 @@
+Sync ``setuptools.schema.json`` with ``validate-pyproject``.
+This will allow using ``package-data`` for stubs only packages.

setuptools/config/_validate_pyproject/NOTICE

@@ -1,7 +1,7 @@
 The code contained in this directory was automatically generated using the
 following command:
 
-    python -m validate_pyproject.pre_compile --output-dir=setuptools/config/_validate_pyproject --enable-plugins setuptools distutils --very-verbose -t setuptools=setuptools/config/setuptools.schema.json -t distutils=setuptools/config/distutils.schema.json
+    python -m validate_pyproject.pre_compile --output-dir=setuptools/config/_validate_pyproject --enable-plugins setuptools distutils --very-verbose -t distutils=setuptools/config/distutils.schema.json -t setuptools=setuptools/config/setuptools.schema.json
 
 Please avoid changing it manually.
 

setuptools/config/_validate_pyproject/error_reporting.py

@@ -1,3 +1,5 @@
+from __future__ import annotations
+
 import io
 import json
 import logging
@@ -6,7 +8,7 @@
 import typing
 from contextlib import contextmanager
 from textwrap import indent, wrap
-from typing import Any, Dict, Generator, Iterator, List, Optional, Sequence, Union
+from typing import Any, Generator, Iterator, Sequence
 
 from .fastjsonschema_exceptions import JsonSchemaValueException
 
@@ -36,7 +38,7 @@
 _NEED_DETAILS = {"anyOf", "oneOf", "allOf", "contains", "propertyNames", "not", "items"}
 
 _CAMEL_CASE_SPLITTER = re.compile(r"\W+|([A-Z][^A-Z\W]*)")
-_IDENTIFIER = re.compile(r"^[\w_]+$", re.I)
+_IDENTIFIER = re.compile(r"^[\w_]+$", re.IGNORECASE)
 
 _TOML_JARGON = {
     "object": "table",
@@ -73,7 +75,7 @@ class ValidationError(JsonSchemaValueException):
     _original_message = ""
 
     @classmethod
-    def _from_jsonschema(cls, ex: JsonSchemaValueException) -> "Self":
+    def _from_jsonschema(cls, ex: JsonSchemaValueException) -> Self:
         formatter = _ErrorFormatting(ex)
         obj = cls(str(formatter), ex.value, formatter.name, ex.definition, ex.rule)
         debug_code = os.getenv("JSONSCHEMA_DEBUG_CODE_GENERATION", "false").lower()
@@ -173,8 +175,8 @@ def _expand_details(self) -> str:
 class _SummaryWriter:
     _IGNORE = frozenset(("description", "default", "title", "examples"))
 
-    def __init__(self, jargon: Optional[Dict[str, str]] = None):
-        self.jargon: Dict[str, str] = jargon or {}
+    def __init__(self, jargon: dict[str, str] | None = None):
+        self.jargon: dict[str, str] = jargon or {}
         # Clarify confusing terms
         self._terms = {
             "anyOf": "at least one of the following",
@@ -207,14 +209,14 @@ def __init__(self, jargon: Optional[Dict[str, str]] = None):
             "multipleOf",
         ]
 
-    def _jargon(self, term: Union[str, List[str]]) -> Union[str, List[str]]:
+    def _jargon(self, term: str | list[str]) -> str | list[str]:
         if isinstance(term, list):
             return [self.jargon.get(t, t) for t in term]
         return self.jargon.get(term, term)
 
     def __call__(
         self,
-        schema: Union[dict, List[dict]],
+        schema: dict | list[dict],
         prefix: str = "",
         *,
         _path: Sequence[str] = (),
@@ -261,15 +263,15 @@ def _is_unecessary(self, path: Sequence[str]) -> bool:
         return any(key.startswith(k) for k in "$_") or key in self._IGNORE
 
     def _filter_unecessary(
-        self, schema: Dict[str, Any], path: Sequence[str]
-    ) -> Dict[str, Any]:
+        self, schema: dict[str, Any], path: Sequence[str]
+    ) -> dict[str, Any]:
         return {
             key: value
             for key, value in schema.items()
             if not self._is_unecessary([*path, key])
         }
 
-    def _handle_simple_dict(self, value: dict, path: Sequence[str]) -> Optional[str]:
+    def _handle_simple_dict(self, value: dict, path: Sequence[str]) -> str | None:
         inline = any(p in value for p in self._guess_inline_defs)
         simple = not any(isinstance(v, (list, dict)) for v in value.values())
         if inline or simple:
@@ -328,7 +330,7 @@ def _child_prefix(self, parent_prefix: str, child_prefix: str) -> str:
         return len(parent_prefix) * " " + child_prefix
 
 
-def _separate_terms(word: str) -> List[str]:
+def _separate_terms(word: str) -> list[str]:
     """
     >>> _separate_terms("FooBar-foo")
     ['foo', 'bar', 'foo']

setuptools/config/_validate_pyproject/extra_validations.py

@@ -3,8 +3,10 @@
 JSON Schema library).
 """
 
+import collections
+import itertools
 from inspect import cleandoc
-from typing import Mapping, TypeVar
+from typing import Generator, Iterable, Mapping, TypeVar
 
 from .error_reporting import ValidationError
 
@@ -19,8 +21,7 @@ class RedefiningStaticFieldAsDynamic(ValidationError):
     """
     __doc__ = _DESC
     _URL = (
-        "https://packaging.python.org/en/latest/specifications/"
-        "pyproject-toml/#dynamic"
+        "https://packaging.python.org/en/latest/specifications/pyproject-toml/#dynamic"
     )
 
 
@@ -31,6 +32,24 @@ class IncludedDependencyGroupMustExist(ValidationError):
     _URL = "https://peps.python.org/pep-0735/"
 
 
+class ImportNameCollision(ValidationError):
+    _DESC = """According to PEP 794:
+
+    All import-names and import-namespaces items must be unique.
+    """
+    __doc__ = _DESC
+    _URL = "https://peps.python.org/pep-0794/"
+
+
+class ImportNameMissing(ValidationError):
+    _DESC = """According to PEP 794:
+
+    An import name must have all parents listed.
+    """
+    __doc__ = _DESC
+    _URL = "https://peps.python.org/pep-0794/"
+
+
 def validate_project_dynamic(pyproject: T) -> T:
     project_table = pyproject.get("project", {})
     dynamic = project_table.get("dynamic", [])
@@ -79,4 +98,54 @@ def validate_include_depenency(pyproject: T) -> T:
     return pyproject
 
 
-EXTRA_VALIDATIONS = (validate_project_dynamic, validate_include_depenency)
+def _remove_private(items: Iterable[str]) -> Generator[str, None, None]:
+    for item in items:
+        yield item.partition(";")[0].rstrip()
+
+
+def validate_import_name_issues(pyproject: T) -> T:
+    project = pyproject.get("project", {})
+    import_names = collections.Counter(_remove_private(project.get("import-names", [])))
+    import_namespaces = collections.Counter(
+        _remove_private(project.get("import-namespaces", []))
+    )
+
+    duplicated = [k for k, v in (import_names + import_namespaces).items() if v > 1]
+
+    if duplicated:
+        raise ImportNameCollision(
+            message="Duplicated names are not allowed in import-names/import-namespaces",
+            value=duplicated,
+            name="data.project.importnames(paces)",
+            definition={
+                "description": cleandoc(ImportNameCollision._DESC),
+                "see": ImportNameCollision._URL,
+            },
+            rule="PEP 794",
+        )
+
+    names = frozenset(import_names + import_namespaces)
+    for name in names:
+        for parent in itertools.accumulate(
+            name.split(".")[:-1], lambda a, b: f"{a}.{b}"
+        ):
+            if parent not in names:
+                raise ImportNameMissing(
+                    message="All parents of an import name must also be listed in import-namespace/import-names",
+                    value=name,
+                    name="data.project.importnames(paces)",
+                    definition={
+                        "description": cleandoc(ImportNameMissing._DESC),
+                        "see": ImportNameMissing._URL,
+                    },
+                    rule="PEP 794",
+                )
+
+    return pyproject
+
+
+EXTRA_VALIDATIONS = (
+    validate_project_dynamic,
+    validate_include_depenency,
+    validate_import_name_issues,
+)

setuptools/config/_validate_pyproject/fastjsonschema_validations.py

@@ -7,7 +7,9 @@
 function with a ``-`` to obtain the format name and vice versa.
 """
 
-import builtins
+from __future__ import annotations
+
+import keyword
 import logging
 import os
 import re
@@ -16,6 +18,8 @@
 from itertools import chain as _chain
 
 if typing.TYPE_CHECKING:
+    import builtins
+
     from typing_extensions import Literal
 
 _logger = logging.getLogger(__name__)
@@ -54,7 +58,9 @@
     (?:\+(?P<local>[a-z0-9]+(?:[-_\.][a-z0-9]+)*))?       # local version
 """
 
-VERSION_REGEX = re.compile(r"^\s*" + VERSION_PATTERN + r"\s*$", re.X | re.I)
+VERSION_REGEX = re.compile(
+    r"^\s*" + VERSION_PATTERN + r"\s*$", re.VERBOSE | re.IGNORECASE
+)
 
 
 def pep440(version: str) -> bool:
@@ -68,7 +74,7 @@ def pep440(version: str) -> bool:
 # PEP 508
 
 PEP508_IDENTIFIER_PATTERN = r"([A-Z0-9]|[A-Z0-9][A-Z0-9._-]*[A-Z0-9])"
-PEP508_IDENTIFIER_REGEX = re.compile(f"^{PEP508_IDENTIFIER_PATTERN}$", re.I)
+PEP508_IDENTIFIER_REGEX = re.compile(f"^{PEP508_IDENTIFIER_PATTERN}$", re.IGNORECASE)
 
 
 def pep508_identifier(name: str) -> bool:
@@ -93,9 +99,9 @@ def pep508(value: str) -> bool:
         """
         try:
             _req.Requirement(value)
-            return True
         except _req.InvalidRequirement:
             return False
+        return True
 
 except ImportError:  # pragma: no cover
     _logger.warning(
@@ -104,7 +110,7 @@ def pep508(value: str) -> bool:
         "To enforce validation, please install `packaging`."
     )
 
-    def pep508(value: str) -> bool:
+    def pep508(value: str) -> bool:  # noqa: ARG001
         return True
 
 
@@ -163,7 +169,7 @@ class _TroveClassifier:
     option (classifiers will be validated anyway during the upload to PyPI).
     """
 
-    downloaded: typing.Union[None, "Literal[False]", typing.Set[str]]
+    downloaded: None | Literal[False] | set[str]
     """
     None => not cached yet
     False => unavailable
@@ -200,7 +206,7 @@ def __call__(self, value: str) -> bool:
             _logger.debug(msg)
             try:
                 self.downloaded = set(_download_classifiers().splitlines())
-            except Exception:
+            except Exception:  # noqa: BLE001
                 self.downloaded = False
                 _logger.debug("Problem with download, skipping validation")
                 return True
@@ -253,21 +259,23 @@ def url(value: str) -> bool:
                 "`scheme` prefix in your URL (e.g. 'http://'). "
                 f"Given value: {value}"
             )
-            if not (value.startswith("/") or value.startswith("\\") or "@" in value):
+            if not (value.startswith(("/", "\\")) or "@" in value):
                 parts = urlparse(f"http://{value}")
 
         return bool(parts.scheme and parts.netloc)
-    except Exception:
+    except Exception:  # noqa: BLE001
         return False
 
 
 # https://packaging.python.org/specifications/entry-points/
 ENTRYPOINT_PATTERN = r"[^\[\s=]([^=]*[^\s=])?"
-ENTRYPOINT_REGEX = re.compile(f"^{ENTRYPOINT_PATTERN}$", re.I)
+ENTRYPOINT_REGEX = re.compile(f"^{ENTRYPOINT_PATTERN}$", re.IGNORECASE)
 RECOMMEDED_ENTRYPOINT_PATTERN = r"[\w.-]+"
-RECOMMEDED_ENTRYPOINT_REGEX = re.compile(f"^{RECOMMEDED_ENTRYPOINT_PATTERN}$", re.I)
+RECOMMEDED_ENTRYPOINT_REGEX = re.compile(
+    f"^{RECOMMEDED_ENTRYPOINT_PATTERN}$", re.IGNORECASE
+)
 ENTRYPOINT_GROUP_PATTERN = r"\w+(\.\w+)*"
-ENTRYPOINT_GROUP_REGEX = re.compile(f"^{ENTRYPOINT_GROUP_PATTERN}$", re.I)
+ENTRYPOINT_GROUP_REGEX = re.compile(f"^{ENTRYPOINT_GROUP_PATTERN}$", re.IGNORECASE)
 
 
 def python_identifier(value: str) -> bool:
@@ -368,11 +376,41 @@ def uint16(value: builtins.int) -> bool:
     return 0 <= value < 2**16
 
 
-def uint(value: builtins.int) -> bool:
+def uint32(value: builtins.int) -> bool:
+    r"""Unsigned 32-bit integer (:math:`0 \leq x < 2^{32}`)"""
+    return 0 <= value < 2**32
+
+
+def uint64(value: builtins.int) -> bool:
     r"""Unsigned 64-bit integer (:math:`0 \leq x < 2^{64}`)"""
     return 0 <= value < 2**64
 
 
+def uint(value: builtins.int) -> bool:
+    r"""Signed 64-bit integer (:math:`0 \leq x < 2^{64}`)"""
+    return 0 <= value < 2**64
+
+
+def int8(value: builtins.int) -> bool:
+    r"""Signed 8-bit integer (:math:`-2^{7} \leq x < 2^{7}`)"""
+    return -(2**7) <= value < 2**7
+
+
+def int16(value: builtins.int) -> bool:
+    r"""Signed 16-bit integer (:math:`-2^{15} \leq x < 2^{15}`)"""
+    return -(2**15) <= value < 2**15
+
+
+def int32(value: builtins.int) -> bool:
+    r"""Signed 32-bit integer (:math:`-2^{31} \leq x < 2^{31}`)"""
+    return -(2**31) <= value < 2**31
+
+
+def int64(value: builtins.int) -> bool:
+    r"""Signed 64-bit integer (:math:`-2^{63} \leq x < 2^{63}`)"""
+    return -(2**63) <= value < 2**63
+
+
 def int(value: builtins.int) -> bool:
     r"""Signed 64-bit integer (:math:`-2^{63} \leq x < 2^{63}`)"""
     return -(2**63) <= value < 2**63
@@ -387,9 +425,9 @@ def SPDX(value: str) -> bool:
         """
         try:
             _licenses.canonicalize_license_expression(value)
-            return True
         except _licenses.InvalidLicenseExpression:
             return False
+        return True
 
 except ImportError:  # pragma: no cover
     _logger.warning(
@@ -398,5 +436,29 @@ def SPDX(value: str) -> bool:
         "To enforce validation, please install `packaging>=24.2`."
     )
 
-    def SPDX(value: str) -> bool:
+    def SPDX(value: str) -> bool:  # noqa: ARG001
         return True
+
+
+VALID_IMPORT_NAME = re.compile(
+    r"""
+    ^                                  # start of string
+        [A-Za-z_][A-Za-z_0-9]+         # a valid Python identifier
+        (?:\.[A-Za-z_][A-Za-z_0-9]*)*  # optionally followed by .identifier's
+    (?:\s*;\s*private)?                # optionally followed by ; private
+    $                                  # end of string
+    """,
+    re.VERBOSE,
+)
+
+
+def import_name(value: str) -> bool:
+    """This is a valid import name. It has to be series of python identifiers
+    (not keywords), separated by dots, optionally followed by a semicolon and
+    the keyword "private".
+    """
+    if VALID_IMPORT_NAME.match(value) is None:
+        return False
+
+    idents, _, _ = value.partition(";")
+    return all(not keyword.iskeyword(ident) for ident in idents.rstrip().split("."))