Merge pull request #4001 from asottile/fix_bytes_repr_text_mix_python_2

Fix UnicodeDecodeError in assertion with mixed non-ascii bytes repr + text

Author: RonnyPfannschmidt

changelog/3999.bugfix.rst

@@ -0,0 +1 @@
+Fix ``UnicodeDecodeError`` in python2.x when a class returns a non-ascii binary ``__repr__`` in an assertion which also contains non-ascii text.

src/_pytest/assertion/rewrite.py

@@ -8,6 +8,7 @@
 import os
 import re
 import six
+import string
 import struct
 import sys
 import types
@@ -466,10 +467,14 @@ def _saferepr(obj):
 
     """
     r = py.io.saferepr(obj)
-    if isinstance(r, six.text_type):
-        return r.replace(u"\n", u"\\n")
-    else:
-        return r.replace(b"\n", b"\\n")
+    # only occurs in python2.x, repr must return text in python3+
+    if isinstance(r, bytes):
+        # Represent unprintable bytes as `\x##`
+        r = u"".join(
+            u"\\x{:x}".format(ord(c)) if c not in string.printable else c.decode()
+            for c in r
+        )
+    return r.replace(u"\n", u"\\n")
 
 
 from _pytest.assertion.util import format_explanation as _format_explanation  # noqa

testing/test_assertrewrite.py

@@ -1,3 +1,4 @@
+# -*- coding: utf-8 -*-
 from __future__ import absolute_import, division, print_function
 
 import glob
@@ -57,7 +58,7 @@ def getmsg(f, extra_ns=None, must_pass=False):
     except AssertionError:
         if must_pass:
             pytest.fail("shouldn't have raised")
-        s = str(sys.exc_info()[1])
+        s = six.text_type(sys.exc_info()[1])
         if not s.startswith("assert"):
             return "AssertionError: " + s
         return s
@@ -608,6 +609,21 @@ def __repr__(self):
 
         assert r"where 1 = \n{ \n~ \n}.a" in util._format_lines([getmsg(f)])[0]
 
+    def test_custom_repr_non_ascii(self):
+        def f():
+            class A(object):
+                name = u"ä"
+
+                def __repr__(self):
+                    return self.name.encode("UTF-8")  # only legal in python2
+
+            a = A()
+            assert not a.name
+
+        msg = getmsg(f)
+        assert "UnicodeDecodeError" not in msg
+        assert "UnicodeEncodeError" not in msg
+
 
 class TestRewriteOnImport(object):
     def test_pycache_is_a_file(self, testdir):