[Python-checkins] [3.10] gh-88233: zipfile: handle extras after a zip64 extra (GH-96161) (#102087)

[ambv]

https://github.com/python/cpython/commit/d445147dfa04bfd52f1cfa7058f4428ca4c7c694
commit: d445147dfa04bfd52f1cfa7058f4428ca4c7c694
branch: 3.10
author: Miss Islington (bot) <31488909+miss-islington at users.noreply.github.com>
committer: ambv <lukasz at langa.pl>
date: 2023-03-28T10:58:51+02:00
summary:

[3.10] gh-88233: zipfile: handle extras after a zip64 extra (GH-96161) (#102087)

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

With many new tests.

Fixes GH-88233
(cherry picked from commit 59e86caca812fc993c5eb7dc8ccd1508ffccba86)

Co-authored-by: Tim Hatch <tim at timhatch.com>

files:
A Misc/NEWS.d/next/Library/2022-09-05-12-17-34.gh-issue-88233.gff9qJ.rst
M Lib/test/test_zipfile.py
M Lib/zipfile.py

diff --git a/Lib/test/test_zipfile.py b/Lib/test/test_zipfile.py
index c7371f59859e..3495fc6548b4 100644
--- a/Lib/test/test_zipfile.py
+++ b/Lib/test/test_zipfile.py
@@ -3221,5 +3221,67 @@ def test_extract_orig_with_implied_dirs(self, alpharep):
         zf.extractall(source_path.parent)
 
 
+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.py b/Lib/zipfile.py
index 23e4ee42d4d9..42e11d7e255c 100644
--- a/Lib/zipfile.py
+++ b/Lib/zipfile.py
@@ -185,6 +185,8 @@ 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 000000000000..806f7011edc3
--- /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."