bpo-39017 Fix infinite loop in the tarfile module

Add a check for length = 0 in the _proc_pax function to avoid running into an infinite loop
python · encukou · Jul 15, 2020 · Jul 12, 2020 · Jul 12, 2020 · Jul 14, 2020
commit 1fa6ef2bc7cee1c8e088dd8b397d9b2d54036dbc
diff --git a/Lib/tarfile.py b/Lib/tarfile.py
@@ -1249,6 +1249,8 @@ def _proc_pax(self, tarfile):
 
             length, keyword = match.groups()
             length = int(length)
+            if length == 0:
+                raise InvalidHeaderError("invalid header")
             value = buf[match.end(2) + 1:match.start(1) + length - 1]
 
             # Normally, we could just use "utf-8" as the encoding and "strict"

diff --git a/Lib/test/recursion.tar b/Lib/test/recursion.tar
diff --git a/Lib/test/test_tarfile.py b/Lib/test/test_tarfile.py
@@ -429,6 +429,11 @@ def test_premature_end_of_archive(self):
                 with self.assertRaisesRegex(tarfile.ReadError, "unexpected end of data"):
                     tar.extractfile(t).read()
 
+    def test_length_zero_header(self):
+        with self.assertRaisesRegex(tarfile.ReadError, "file could not be opened successfully"):
+            with tarfile.open(support.findfile('recursion.tar')) as tar:
+                tar.getmembers()
+
 class MiscReadTestBase(CommonReadTest):
     def requires_name_attribute(self):
         pass