bpo-40698: Improve distutils upload hash digests (GH-20260)

- Fix upload test on systems that blocks MD5
- Add SHA2-256 and Blake2b-256 digests based on new Warehous and twine
  specs.

Signed-off-by: Christian Heimes <[email protected]>

Author: tiran

Doc/whatsnew/3.9.rst

@@ -331,6 +331,13 @@ and :meth:`~datetime.datetime.isocalendar()` of :class:`datetime.datetime`
 methods now returns a :func:`~collections.namedtuple` instead of a :class:`tuple`.
 (Contributed by Dong-hee Na in :issue:`24416`.)
 
+distutils
+---------
+
+The :command:`upload` command now creates SHA2-256 and Blake2b-256 hash
+digests. It skips MD5 on platforms that block MD5 digest.
+(Contributed by Christian Heimes in :issue:`40698`.)
+
 fcntl
 -----
 

Lib/distutils/command/upload.py

@@ -16,6 +16,16 @@
 from distutils.spawn import spawn
 from distutils import log
 
+
+# PyPI Warehouse supports MD5, SHA256, and Blake2 (blake2-256)
+# https://bugs.python.org/issue40698
+_FILE_CONTENT_DIGESTS = {
+    "md5_digest": getattr(hashlib, "md5", None),
+    "sha256_digest": getattr(hashlib, "sha256", None),
+    "blake2_256_digest": getattr(hashlib, "blake2b", None),
+}
+
+
 class upload(PyPIRCCommand):
 
     description = "upload binary package to PyPI"
@@ -87,6 +97,7 @@ def upload_file(self, command, pyversion, filename):
             content = f.read()
         finally:
             f.close()
+
         meta = self.distribution.metadata
         data = {
             # action
@@ -101,7 +112,6 @@ def upload_file(self, command, pyversion, filename):
             'content': (os.path.basename(filename),content),
             'filetype': command,
             'pyversion': pyversion,
-            'md5_digest': hashlib.md5(content).hexdigest(),
 
             # additional meta-data
             'metadata_version': '1.0',
@@ -123,6 +133,16 @@ def upload_file(self, command, pyversion, filename):
 
         data['comment'] = ''
 
+        # file content digests
+        for digest_name, digest_cons in _FILE_CONTENT_DIGESTS.items():
+            if digest_cons is None:
+                continue
+            try:
+                data[digest_name] = digest_cons(content).hexdigest()
+            except ValueError:
+                # hash digest not available or blocked by security policy
+                pass
+
         if self.sign:
             with open(filename + ".asc", "rb") as f:
                 data['gpg_signature'] = (os.path.basename(filename) + ".asc",

Lib/distutils/tests/test_upload.py

@@ -130,14 +130,30 @@ def test_upload(self):
 
         # what did we send ?
         headers = dict(self.last_open.req.headers)
-        self.assertEqual(headers['Content-length'], '2162')
+        self.assertGreaterEqual(int(headers['Content-length']), 2162)
         content_type = headers['Content-type']
         self.assertTrue(content_type.startswith('multipart/form-data'))
         self.assertEqual(self.last_open.req.get_method(), 'POST')
         expected_url = 'https://upload.pypi.org/legacy/'
         self.assertEqual(self.last_open.req.get_full_url(), expected_url)
-        self.assertTrue(b'xxx' in self.last_open.req.data)
-        self.assertIn(b'protocol_version', self.last_open.req.data)
+        data = self.last_open.req.data
+        self.assertIn(b'xxx',data)
+        self.assertIn(b'protocol_version', data)
+        self.assertIn(b'sha256_digest', data)
+        self.assertIn(
+            b'cd2eb0837c9b4c962c22d2ff8b5441b7b45805887f051d39bf133b583baf'
+            b'6860',
+            data
+        )
+        if b'md5_digest' in data:
+            self.assertIn(b'f561aaf6ef0bf14d4208bb46a4ccb3ad', data)
+        if b'blake2_256_digest' in data:
+            self.assertIn(
+                b'b6f289a27d4fe90da63c503bfe0a9b761a8f76bb86148565065f040be'
+                b'6d1c3044cf7ded78ef800509bccb4b648e507d88dc6383d67642aadcc'
+                b'ce443f1534330a',
+                data
+            )
 
         # The PyPI response body was echoed
         results = self.get_logs(INFO)
@@ -166,7 +182,7 @@ def test_upload_correct_cr(self):
         cmd.run()
 
         headers = dict(self.last_open.req.headers)
-        self.assertEqual(headers['Content-length'], '2172')
+        self.assertGreaterEqual(int(headers['Content-length']), 2172)
         self.assertIn(b'long description\r', self.last_open.req.data)
 
     def test_upload_fails(self):

Misc/NEWS.d/next/Library/2020-05-20-14-38-04.bpo-40698.zwl5Hc.rst

@@ -0,0 +1,2 @@
+:mod:`distutils` upload creates SHA2-256 and Blake2b-256 digests. MD5
+digests is skipped if platform blocks MD5.