pythongh-88233: zipfile: handle extras after a zip64 extra (pythonGH-96161)

Previously, any data _after_ the zip64 extra would be removed.

With many new tests.

Fixes python#88233

Automerge-Triggered-By: GH:jaraco

Author: thatch

Lib/test/test_zipfile/test_core.py

@@ -3010,5 +3010,67 @@ def test_cli_with_metadata_encoding_extract(self):
             self.assertIn(name, listing)
 
 
+class StripExtraTests(unittest.TestCase):
+    # Note: all of the "z" characters are technically invalid, but up
+    # to 3 bytes at the end of the extra will be passed through as they
+    # are too short to encode a valid extra.
+
+    ZIP64_EXTRA = 1
+
+    def test_no_data(self):
+        s = struct.Struct("<HH")
+        a = s.pack(self.ZIP64_EXTRA, 0)
+        b = s.pack(2, 0)
+        c = s.pack(3, 0)
+
+        self.assertEqual(b'', zipfile._strip_extra(a, (self.ZIP64_EXTRA,)))
+        self.assertEqual(b, zipfile._strip_extra(b, (self.ZIP64_EXTRA,)))
+        self.assertEqual(
+            b+b"z", zipfile._strip_extra(b+b"z", (self.ZIP64_EXTRA,)))
+
+        self.assertEqual(b+c, zipfile._strip_extra(a+b+c, (self.ZIP64_EXTRA,)))
+        self.assertEqual(b+c, zipfile._strip_extra(b+a+c, (self.ZIP64_EXTRA,)))
+        self.assertEqual(b+c, zipfile._strip_extra(b+c+a, (self.ZIP64_EXTRA,)))
+
+    def test_with_data(self):
+        s = struct.Struct("<HH")
+        a = s.pack(self.ZIP64_EXTRA, 1) + b"a"
+        b = s.pack(2, 2) + b"bb"
+        c = s.pack(3, 3) + b"ccc"
+
+        self.assertEqual(b"", zipfile._strip_extra(a, (self.ZIP64_EXTRA,)))
+        self.assertEqual(b, zipfile._strip_extra(b, (self.ZIP64_EXTRA,)))
+        self.assertEqual(
+            b+b"z", zipfile._strip_extra(b+b"z", (self.ZIP64_EXTRA,)))
+
+        self.assertEqual(b+c, zipfile._strip_extra(a+b+c, (self.ZIP64_EXTRA,)))
+        self.assertEqual(b+c, zipfile._strip_extra(b+a+c, (self.ZIP64_EXTRA,)))
+        self.assertEqual(b+c, zipfile._strip_extra(b+c+a, (self.ZIP64_EXTRA,)))
+
+    def test_multiples(self):
+        s = struct.Struct("<HH")
+        a = s.pack(self.ZIP64_EXTRA, 1) + b"a"
+        b = s.pack(2, 2) + b"bb"
+
+        self.assertEqual(b"", zipfile._strip_extra(a+a, (self.ZIP64_EXTRA,)))
+        self.assertEqual(b"", zipfile._strip_extra(a+a+a, (self.ZIP64_EXTRA,)))
+        self.assertEqual(
+            b"z", zipfile._strip_extra(a+a+b"z", (self.ZIP64_EXTRA,)))
+        self.assertEqual(
+            b+b"z", zipfile._strip_extra(a+a+b+b"z", (self.ZIP64_EXTRA,)))
+
+        self.assertEqual(b, zipfile._strip_extra(a+a+b, (self.ZIP64_EXTRA,)))
+        self.assertEqual(b, zipfile._strip_extra(a+b+a, (self.ZIP64_EXTRA,)))
+        self.assertEqual(b, zipfile._strip_extra(b+a+a, (self.ZIP64_EXTRA,)))
+
+    def test_too_short(self):
+        self.assertEqual(b"", zipfile._strip_extra(b"", (self.ZIP64_EXTRA,)))
+        self.assertEqual(b"z", zipfile._strip_extra(b"z", (self.ZIP64_EXTRA,)))
+        self.assertEqual(
+            b"zz", zipfile._strip_extra(b"zz", (self.ZIP64_EXTRA,)))
+        self.assertEqual(
+            b"zzz", zipfile._strip_extra(b"zzz", (self.ZIP64_EXTRA,)))
+
+
 if __name__ == "__main__":
     unittest.main()

Lib/zipfile/__init__.py

@@ -207,6 +207,8 @@ def _strip_extra(extra, xids):
         i = j
     if not modified:
         return extra
+    if start != len(extra):
+        buffer.append(extra[start:])
     return b''.join(buffer)
 
 def _check_zipfile(fp):

Misc/NEWS.d/next/Library/2022-09-05-12-17-34.gh-issue-88233.gff9qJ.rst

@@ -0,0 +1,2 @@
+Correctly preserve "extra" fields in ``zipfile`` regardless of their
+ordering relative to a zip64 "extra."