Hi Mateusz,

Unfortunately that error message hides where the error actually occurred because it can't form the error message.
But presumably you've put in print statements to sys.stderr and have bracketed the exact location where the problem happens.

Once you identify the exact location, the UnicodeDecodeError indicates that you were using a string instead of a unicode string.  If your string only contains ASCII characters, that is fine.  If not, it needs to have a UTF-8 encoding, or you need to use a unicode string. ( See http://docs.python.org/2/howto/unicode.html for details about unicode strings in Python 2.)

For example, if your string has a Polish encoding, a.k.a. ISO/IEC 8859-2, then to get the unicode equivalent:

    unicode_string = string.decode('iso 8859-2')

How you know what encoding to use depends on how the string was obtained.  Since the character that couldn't be decoded in the error message corresponds to a letter in your name, you might be able to fix this bug by adding the proper encoding comment at the top of your .py file and using a unicode string constant whenever that character is used in a string.

To fix the error reporting (this change will be in the next 1.8 release candidate and the daily build), I'm adding the following two lines to .../lib/site-packages/Pmw/Pmw_1_3/lib/PmwBase.py:

    if not isinstance(exc_value, unicode):
       exc_value = exc_value.decode('utf-8', 'replace')

Just before the line:

    msg = msg + u'%s: %s\n' % (exc_type, exc_value)

as seen in the traceback in your screen shot.  FYI, on Windows, to get the text of the debug window instead of doing a screen shot, you right-click on the title bar and use the Edit submenu to Select All and Copy the text.

    HTH,

    Greg

On 05/23/2013 09:44 AM, Mateusz Dobrychłop wrote: