Comment 2 for bug 452790

As far as I can tell, python should default to whatever C uses, and processing floats should "just work", however that never turns out as planned... I can work around this using 'locale.atof()' instead of 'float()', assuming python interprets your localizations settings properly.

Can you open python on the command line, run:

import locale
locale.localeconv()['decimal_point']

and see if it is set properly?