gh-88233: zipfile: handle extras after a zip64 extra (GH-96161)

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

With many new tests.

Fixes #88233

Automerge-Triggered-By: GH:jaraco

3 files changed, 66 insertions, 0 deletions

diff --git a/Lib/test/test_zipfile/test_core.py b/Lib/test/test_zipfile/test_core.py
index cf41d0e8..e23f5c2 100644
--- a/Lib/test/test_zipfile/test_core.py
+++ b/Lib/test/test_zipfile/test_core.py
@@ -3010,5 +3010,67 @@ class EncodedMetadataTests(unittest.TestCase):
             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()
diff --git a/Lib/zipfile/__init__.py b/Lib/zipfile/__init__.py
index e1833dd..6e6211d 100644
--- a/Lib/zipfile/__init__.py
+++ b/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):
diff --git a/Misc/NEWS.d/next/Library/2022-09-05-12-17-34.gh-issue-88233.gff9qJ.rst b/Misc/NEWS.d/next/Library/2022-09-05-12-17-34.gh-issue-88233.gff9qJ.rst
new file mode 100644
index 0000000..806f701
--- /dev/null
+++ b/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."