Exclude top-level .tox|.nox|.venv from sdist (#4603)

Author: abravalheri

MANIFEST.in

@@ -19,4 +19,3 @@ include tox.ini
 include setuptools/tests/config/setupcfg_examples.txt
 include setuptools/config/*.schema.json
 global-exclude *.py[cod] __pycache__
-prune .tox

setuptools/command/egg_info.py

@@ -606,16 +606,6 @@ def _add_referenced_files(self):
             log.debug("adding file referenced by config '%s'", rf)
         self.filelist.extend(referenced)
 
-    def prune_file_list(self):
-        build = self.get_finalized_command('build')
-        base_dir = self.distribution.get_fullname()
-        self.filelist.prune(build.build_base)
-        self.filelist.prune(base_dir)
-        sep = re.escape(os.sep)
-        self.filelist.exclude_pattern(
-            r'(^|' + sep + r')(RCS|CVS|\.svn)' + sep, is_regex=True
-        )
-
     def _safe_data_files(self, build_py):
         """
         The parent class implementation of this method

setuptools/command/sdist.py

@@ -2,6 +2,7 @@
 
 import contextlib
 import os
+import re
 from itertools import chain
 
 from .._importlib import metadata
@@ -156,6 +157,12 @@ def _add_defaults_data_files(self):
         except TypeError:
             log.warn("data_files contains unexpected objects")
 
+    def prune_file_list(self):
+        super().prune_file_list()
+        # Prevent accidental inclusion of test-related cache dirs at the project root
+        sep = re.escape(os.sep)
+        self.filelist.exclude_pattern(r"^(\.tox|\.nox|\.venv)" + sep, is_regex=True)
+
     def check_readme(self):
         for f in self.READMES:
             if os.path.exists(f):

setuptools/tests/test_sdist.py

@@ -9,6 +9,7 @@
 import tempfile
 import unicodedata
 from inspect import cleandoc
+from pathlib import Path
 from unittest import mock
 
 import jaraco.path
@@ -425,6 +426,46 @@ def test_defaults_case_sensitivity(self, source_dir):
         assert 'setup.py' not in manifest, manifest
         assert 'setup.cfg' not in manifest, manifest
 
+    def test_exclude_dev_only_cache_folders(self, source_dir):
+        included = {
+            # Emulate problem in https://github.com/pypa/setuptools/issues/4601
+            "MANIFEST.in": (
+                "global-include LICEN[CS]E* COPYING* NOTICE* AUTHORS*\n"
+                "global-include *.txt\n"
+            ),
+            # For the sake of being conservative and limiting unforeseen side-effects
+            # we just exclude dev-only cache folders at the root of the repository:
+            "test/.venv/lib/python3.9/site-packages/bar-2.dist-info/AUTHORS.rst": "",
+            "src/.nox/py/lib/python3.12/site-packages/bar-2.dist-info/COPYING.txt": "",
+            "doc/.tox/default/lib/python3.11/site-packages/foo-4.dist-info/LICENSE": "",
+            # Let's test against false positives with similarly named files:
+            ".venv-requirements.txt": "",
+            ".tox-coveragerc.txt": "",
+            ".noxy/coveragerc.txt": "",
+        }
+
+        excluded = {
+            # .tox/.nox/.venv are well-know folders present at the root of Python repos
+            # and therefore should be excluded
+            ".tox/release/lib/python3.11/site-packages/foo-4.dist-info/LICENSE": "",
+            ".nox/py/lib/python3.12/site-packages/bar-2.dist-info/COPYING.txt": "",
+            ".venv/lib/python3.9/site-packages/bar-2.dist-info/AUTHORS.rst": "",
+        }
+
+        for file, content in {**excluded, **included}.items():
+            Path(source_dir, file).parent.mkdir(parents=True, exist_ok=True)
+            Path(source_dir, file).write_text(content, encoding="utf-8")
+
+        cmd = self.setup_with_extension()
+        self.assert_package_data_in_manifest(cmd)
+        manifest = {f.replace(os.sep, '/') for f in cmd.filelist.files}
+        for path in excluded:
+            assert os.path.exists(path)
+            assert path not in manifest, (path, manifest)
+        for path in included:
+            assert os.path.exists(path)
+            assert path in manifest, (path, manifest)
+
     @fail_on_ascii
     def test_manifest_is_written_with_utf8_encoding(self):
         # Test for #303.
@@ -915,4 +956,4 @@ def test_sanity_check_setuptools_own_sdist(setuptools_sdist):
 
     # setuptools sdist should not include the .tox folder
     tox_files = [name for name in files if ".tox" in name]
-    assert len(tox_files) == 0
+    assert len(tox_files) == 0, f"not empty {tox_files}"