Merge PR #100 - avoid decoding of invalid Unicode output

Use PYTHONIOENCODING to enable utf-8 stdout for the nsjail pipe,
and handle the potential case where this is bypassable.

1 files changed, 17 insertions, 0 deletions

diff --git a/tests/test_nsjail.py b/tests/test_nsjail.py
index 4d4676b..46193b2 100644
--- a/tests/test_nsjail.py
+++ b/tests/test_nsjail.py
@@ -100,6 +100,23 @@ class NsJailTests(unittest.TestCase):
         self.assertEqual(result.stdout, "ValueError: embedded null byte")
         self.assertEqual(result.stderr, None)
 
+    def test_print_bad_unicode_encode_error(self):
+        result = self.nsjail.python3("print(chr(56550))")
+        self.assertEqual(result.returncode, 1)
+        self.assertIn("UnicodeEncodeError", result.stdout)
+        self.assertEqual(result.stderr, None)
+
+    def test_unicode_env_erase_escape_fails(self):
+        result = self.nsjail.python3(dedent("""
+            import os
+            import sys
+            os.unsetenv('PYTHONIOENCODING')
+            os.execl(sys.executable, 'python', '-c', 'print(chr(56550))')
+        """).strip())
+        self.assertEqual(result.returncode, None)
+        self.assertEqual(result.stdout, "UnicodeDecodeError: invalid Unicode in output pipe")
+        self.assertEqual(result.stderr, None)
+
     @unittest.mock.patch("snekbox.nsjail.DEBUG", new=False)
     def test_log_parser(self):
         log_lines = (