Guido van Rossum schrieb: > On 7/12/07, thomas.heller <[email protected]> wrote: >> Author: thomas.heller >> Date: Thu Jul 12 13:59:52 2007 >> New Revision: 56313 >> >> Modified: >> python/branches/py3k-struni/Lib/ctypes/test/test_objects.py >> Log: >> Fix a ctypes test. >> >> Modified: python/branches/py3k-struni/Lib/ctypes/test/test_objects.py >> ============================================================================== >> --- python/branches/py3k-struni/Lib/ctypes/test/test_objects.py (original) >> +++ python/branches/py3k-struni/Lib/ctypes/test/test_objects.py Thu Jul 12 >> 13:59:52 2007 >> @@ -22,9 +22,9 @@ >> >> >>> array[4] = 'foo bar' >> >>> array._objects >> -{'4': 'foo bar'} >> +{s'4': b'foo bar'} >> >>> array[4] >> -'foo bar' >> +s'foo bar' >> >>> > > FWIW, when tests show s'...', the preferred fix is to change the > underlying code to return unicode, not to fix the test to accept the > 8-bit string. This is why we made str8.__repr__ add the 's' prefix: so > more tests would fail so we'd be alerted to places still returning > str8 instead of str. > Sure.
Generally, these are quick fixes to just fix the tests before I change the semantics of the ctypes implementation. In this special case, the test is testing the internals of the ctypes objects, and the I will probably change the code to use bytes instead of 8-bit strings. Thomas _______________________________________________ Python-3000-checkins mailing list [email protected] http://mail.python.org/mailman/listinfo/python-3000-checkins
