fix for duplicate lines in wheel RECORD, prefer pathlib

Author: dimbleby

src/poetry/core/json/__init__.py

@@ -1,25 +1,25 @@
 from __future__ import annotations
 
 import json
-import os
 
+from pathlib import Path
 from typing import Any
 
 
-SCHEMA_DIR = os.path.join(os.path.dirname(__file__), "schemas")
+SCHEMA_DIR = Path(__file__).parent / "schemas"
 
 
 class ValidationError(ValueError):
     pass
 
 
 def validate_object(obj: dict[str, Any], schema_name: str) -> list[str]:
-    schema_file = os.path.join(SCHEMA_DIR, f"{schema_name}.json")
+    schema_file = SCHEMA_DIR / f"{schema_name}.json"
 
-    if not os.path.exists(schema_file):
+    if not schema_file.exists():
         raise ValueError(f"Schema {schema_name} does not exist.")
 
-    with open(schema_file, encoding="utf-8") as f:
+    with schema_file.open(encoding="utf-8") as f:
         schema = json.loads(f.read())
 
     from jsonschema import Draft7Validator

src/poetry/core/masonry/builders/wheel.py

@@ -163,19 +163,19 @@ def _build(self, wheel: zipfile.ZipFile) -> None:
                 # we assume that the build script will build and copy the files
                 # directly.
                 # That way they will be picked up when adding files to the wheel.
-                current_path = os.getcwd()
+                current_path = Path.cwd()
                 try:
-                    os.chdir(str(self._path))
+                    os.chdir(self._path)
                     self._run_build_script(self._package.build_script)
                 finally:
                     os.chdir(current_path)
             else:
                 with SdistBuilder(poetry=self._poetry).setup_py() as setup:
                     # We need to place ourselves in the temporary
                     # directory in order to build the package
-                    current_path = os.getcwd()
+                    current_path = Path.cwd()
                     try:
-                        os.chdir(str(self._path))
+                        os.chdir(self._path)
                         self._run_build_command(setup)
                     finally:
                         os.chdir(current_path)
@@ -194,9 +194,9 @@ def _build(self, wheel: zipfile.ZipFile) -> None:
                         if pkg.is_dir() or self.is_excluded(pkg):
                             continue
 
-                        rel_path = str(pkg.relative_to(lib))
+                        rel_path = pkg.relative_to(lib)
 
-                        if rel_path in wheel.namelist():
+                        if rel_path.as_posix() in wheel.namelist():
                             continue
 
                         logger.debug(f"Adding: {rel_path}")
@@ -210,7 +210,7 @@ def _copy_file_scripts(self, wheel: zipfile.ZipFile) -> None:
             self._add_file(
                 wheel,
                 abs_path,
-                Path.joinpath(Path(self.wheel_data_folder), "scripts", abs_path.name),
+                Path(self.wheel_data_folder) / "scripts" / abs_path.name,
             )
 
     def _run_build_command(self, setup: Path) -> None:
@@ -268,7 +268,7 @@ def prepare_metadata(self, metadata_directory: Path) -> Path:
                 continue
 
             dest = dist_info / license_file.relative_to(self._path)
-            os.makedirs(dest.parent, exist_ok=True)
+            dest.parent.mkdir(parents=True, exist_ok=True)
             shutil.copy(license_file, dest)
 
         return dist_info
@@ -345,19 +345,15 @@ def tag(self) -> str:
     def _add_file(
         self,
         wheel: zipfile.ZipFile,
-        full_path: Path | str,
-        rel_path: Path | str,
+        full_path: Path,
+        rel_path: Path,
     ) -> None:
-        full_path, rel_path = str(full_path), str(rel_path)
-        if os.sep != "/":
-            # We always want to have /-separated paths in the zip file and in
-            # RECORD
-            rel_path = rel_path.replace(os.sep, "/")
-
-        zinfo = zipfile.ZipInfo(rel_path)
+        # We always want to have /-separated paths in the zip file and in RECORD
+        rel_path_name = rel_path.as_posix()
+        zinfo = zipfile.ZipInfo(rel_path_name)
 
         # Normalize permission bits to either 755 (executable) or 644
-        st_mode = os.stat(full_path).st_mode
+        st_mode = full_path.stat().st_mode
         new_mode = normalize_file_permissions(st_mode)
         zinfo.external_attr = (new_mode & 0xFFFF) << 16  # Unix attributes
 
@@ -375,10 +371,10 @@ def _add_file(
             src.seek(0)
             wheel.writestr(zinfo, src.read(), compress_type=zipfile.ZIP_DEFLATED)
 
-        size = os.stat(full_path).st_size
+        size = full_path.stat().st_size
         hash_digest = urlsafe_b64encode(hashsum.digest()).decode("ascii").rstrip("=")
 
-        self._records.append((rel_path, hash_digest, size))
+        self._records.append((rel_path_name, hash_digest, size))
 
     @contextlib.contextmanager
     def _write_to_zip(

src/poetry/core/version/markers.py

@@ -830,8 +830,10 @@ def dnf(marker: BaseMarker) -> BaseMarker:
         return MarkerUnion.of(
             *[MultiMarker.of(*c) for c in itertools.product(*sub_marker_lists)]
         )
+
     if isinstance(marker, MarkerUnion):
         return MarkerUnion.of(*[dnf(m) for m in marker.markers])
+
     return marker
 
 

tests/masonry/builders/test_complete.py

@@ -1,5 +1,6 @@
 from __future__ import annotations
 
+import csv
 import os
 import platform
 import re
@@ -47,7 +48,7 @@ def clear_samples_dist() -> None:
     or platform.python_implementation().lower() == "pypy",
     reason="Disable test on Windows for Python <=3.6 and for PyPy",
 )
-def test_wheel_c_extension() -> None:
+def test_wheel_c_extension() -> None:  # NOSONAR
     module_path = fixtures_dir / "extended"
     builder = Builder(Factory().create_poetry(module_path))
     builder.build(fmt="all")
@@ -89,20 +90,27 @@ def test_wheel_c_extension() -> None:
             is not None
         )
 
-        records = zip.read("extended-0.1.dist-info/RECORD").decode()
+        record = zip.read("extended-0.1.dist-info/RECORD").decode()
+        records = csv.reader(record.splitlines())
+        record_files = [row[0] for row in records]
 
-        assert re.search(r"\s+extended/extended.*\.(so|pyd)", records) is not None
+        assert re.search(r"\s+extended/extended.*\.(so|pyd)", record) is not None
     finally:
         zip.close()
 
+    # Files in RECORD should match files in wheel.
+    zip_files = sorted(zip.namelist())
+    assert zip_files == sorted(record_files)
+    assert len(set(record_files)) == len(record_files)
+
 
 @pytest.mark.skipif(
     sys.platform == "win32"
     and sys.version_info <= (3, 6)
     or platform.python_implementation().lower() == "pypy",
     reason="Disable test on Windows for Python <=3.6 and for PyPy",
 )
-def test_wheel_c_extension_with_no_setup() -> None:
+def test_wheel_c_extension_with_no_setup() -> None:  # NOSONAR
     module_path = fixtures_dir / "extended_with_no_setup"
     builder = Builder(Factory().create_poetry(module_path))
     builder.build(fmt="all")
@@ -144,20 +152,27 @@ def test_wheel_c_extension_with_no_setup() -> None:
             is not None
         )
 
-        records = zip.read("extended-0.1.dist-info/RECORD").decode()
+        record = zip.read("extended-0.1.dist-info/RECORD").decode()
+        records = csv.reader(record.splitlines())
+        record_files = [row[0] for row in records]
 
-        assert re.search(r"\s+extended/extended.*\.(so|pyd)", records) is not None
+        assert re.search(r"\s+extended/extended.*\.(so|pyd)", record) is not None
     finally:
         zip.close()
 
+    # Files in RECORD should match files in wheel.
+    zip_files = sorted(zip.namelist())
+    assert zip_files == sorted(record_files)
+    assert len(set(record_files)) == len(record_files)
+
 
 @pytest.mark.skipif(
     sys.platform == "win32"
     and sys.version_info <= (3, 6)
     or platform.python_implementation().lower() == "pypy",
     reason="Disable test on Windows for Python <=3.6 and for PyPy",
 )
-def test_wheel_c_extension_src_layout() -> None:
+def test_wheel_c_extension_src_layout() -> None:  # NOSONAR
     module_path = fixtures_dir / "src_extended"
     builder = Builder(Factory().create_poetry(module_path))
     builder.build(fmt="all")
@@ -199,12 +214,19 @@ def test_wheel_c_extension_src_layout() -> None:
             is not None
         )
 
-        records = zip.read("extended-0.1.dist-info/RECORD").decode()
+        record = zip.read("extended-0.1.dist-info/RECORD").decode()
+        records = csv.reader(record.splitlines())
+        record_files = [row[0] for row in records]
 
-        assert re.search(r"\s+extended/extended.*\.(so|pyd)", records) is not None
+        assert re.search(r"\s+extended/extended.*\.(so|pyd)", record) is not None
     finally:
         zip.close()
 
+    # Files in RECORD should match files in wheel.
+    zip_files = sorted(zip.namelist())
+    assert zip_files == sorted(record_files)
+    assert len(set(record_files)) == len(record_files)
+
 
 def test_complete() -> None:
     module_path = fixtures_dir / "complete"
@@ -297,20 +319,23 @@ def test_complete() -> None:
         # vary per operating systems and Python versions.
         # So instead of 1:1 assertion, let's do a bit clunkier one:
 
-        expected_records = [
+        actual_files = [row[0] for row in csv.reader(actual_records.splitlines())]
+        expected_files = [
+            "my_package-1.2.3.data/scripts/script.sh",
+            "my_package-1.2.3.dist-info/LICENSE",
+            "my_package-1.2.3.dist-info/METADATA",
+            "my_package-1.2.3.dist-info/RECORD",
+            "my_package-1.2.3.dist-info/WHEEL",
+            "my_package-1.2.3.dist-info/entry_points.txt",
             "my_package/__init__.py",
             "my_package/data1/test.json",
             "my_package/sub_pkg1/__init__.py",
             "my_package/sub_pkg2/__init__.py",
             "my_package/sub_pkg2/data2/data.json",
-            "my_package-1.2.3.dist-info/entry_points.txt",
-            "my_package-1.2.3.dist-info/LICENSE",
-            "my_package-1.2.3.dist-info/WHEEL",
-            "my_package-1.2.3.dist-info/METADATA",
+            "my_package/sub_pkg3/foo.py",
         ]
 
-        for expected_record in expected_records:
-            assert expected_record in actual_records
+        assert sorted(expected_files) == sorted(actual_files)
 
     finally:
         zip.close()