commit bdeb56cd21ef3f4f086c93045d80f2a753823379
author Miss Islington (bot) <31488909+miss-islington@users.noreply.github.com> Mon Dec 03 01:09:11 2018 -0800
committer GitHub <noreply@github.com> Mon Dec 03 01:09:11 2018 -0800
tree 71896c0643ea65514557a34c7e3bffefcdb81dc6
parent 1ef06c62d3c05cbba6448c56af30a09c551c9ec2

bpo-35372: Fix the code page decoder for input > 2 GiB. (GH-10848)

(cherry picked from commit 4013c179117754b039957db4730880bf3285919d)

Co-authored-by: Serhiy Storchaka <storchaka@gmail.com>

Lib/test/test_codecs.py

diff --git a/Lib/test/test_codecs.py b/Lib/test/test_codecs.py
index a59a5e2..5c2de21 100644
--- a/Lib/test/test_codecs.py
+++ b/Lib/test/test_codecs.py
@@ -3185,6 +3185,24 @@
             codec = codecs.lookup('cp123')
             self.assertEqual(codec.name, 'mbcs')
 
+    @support.bigmemtest(size=2**31, memuse=7, dry_run=False)
+    def test_large_input(self):
+        # Test input longer than INT_MAX.
+        # Input should contain undecodable bytes before and after
+        # the INT_MAX limit.
+        encoded = (b'01234567' * (2**28-1) +
+                   b'\x85\x86\xea\xeb\xec\xef\xfc\xfd\xfe\xff')
+        self.assertEqual(len(encoded), 2**31+2)
+        decoded = codecs.code_page_decode(932, encoded, 'surrogateescape', True)
+        self.assertEqual(decoded[1], len(encoded))
+        del encoded
+        self.assertEqual(len(decoded[0]), decoded[1])
+        self.assertEqual(decoded[0][:10], '0123456701')
+        self.assertEqual(decoded[0][-20:],
+                         '6701234567'
+                         '\udc85\udc86\udcea\udceb\udcec'
+                         '\udcef\udcfc\udcfd\udcfe\udcff')
+
 
 class ASCIITest(unittest.TestCase):
     def test_encode(self):