Not checking UnicodeDecodeError message because it differs between CPython and PyPy.

Checking is too much trouble for very little gain.

CPython:
'utf8' codec can't decode byte 0x80 in position 0: invalid start byte

PyPy:
'utf8' codec can't decode byte 0x80 in position 0: invalid utf-8

Author: rwgk

tests/test_pytypes.py

@@ -262,9 +262,8 @@ def test_pybind11_str_raw_str():
 
     malformed_utf8 = b'\x80'
     if pytest.PY2:
-        with pytest.raises(UnicodeDecodeError) as excinfo:
+        with pytest.raises(UnicodeDecodeError):
             cvt(malformed_utf8)
-        assert "invalid start byte" in str(excinfo)
     else:
         malformed_cvt = cvt(malformed_utf8)
         assert type(malformed_cvt) is unicode if pytest.PY2 else str  # noqa: F821
@@ -420,6 +419,5 @@ def test_pass_bytes_or_unicode_to_string_types():
     assert m.pass_to_std_string(u"Str") == 3
 
     malformed_utf8 = b"\x80"
-    with pytest.raises(UnicodeDecodeError) as excinfo:
+    with pytest.raises(UnicodeDecodeError):
         m.pass_to_pybind11_str(malformed_utf8)
-    assert 'invalid start byte' in str(excinfo.value)