Fix UnicodeDecodeError: 'ascii' codec can't decode byte #492
Conversation
|
I'm not sure this is the right fix. I don't understand why the conversion to unicode is needed in the first place; and even if it is, why does it make sense to hardwire the encoding to windows-1252? |
|
If the user put non-ascii characters in an 8-bit string, there's nothing we can safely do -- there's no way to assume the encoding that will work for everyone. If that's what the user wants to do, they must pass Unicode strings. I think this should remain an error (though we could improve the error message if we want). |
|
Christoph, your solution will break if a unicode string is supplied as an argument to is_math_text: yields Simply removing the call to unicode() solves the immediate problem--it makes the example work--and I don't see anything wrong with it. But Mike wrote is_math_text() with the call to unicode(), so I would not want to delete it without his concurrence. All this is made more complicated by the huge byte/string/unicode change in Py 3.x. Eric |
|
See pull request #494 for an alternative solution. |
On Windows the example font_table_ttf.py fails with
UnicodeDecodeError: 'ascii' codec can't decode byte 0xc1 in position 0: ordinal not in range(128)