Update ufoLib to 2.3.0

[miguelsousa]

No description provided.

[miguelsousa]

The tests won't pass until mutatorMath gets an update. See unified-font-object/ufoLib#160 (comment)

[miguelsousa]

@anthrotype are the data/string differences supposed to happen if ufoLib version 2.3.0 is installed without the lxml option? I was under the impression that you kept the previous behavior by default.

[anthrotype]

I just said that ufoLib now behaves the same way both with or without lxml and in python 2 or python 3. Any of these bytes-that-should-be-unicode should get fixed.

[anthrotype]

Of course they only affect python 2.7, where this confusion can happen

[miguelsousa]

thanks. Can you confirm my assessment that mutatorMath has bytes-that-should-be-unicode things to fix?

[anthrotype]

Can you confirm my assessment that mutatorMath has bytes-that-should-be-unicode things to fix?

I had underestimated this issue.. Requiring all clients of ufoLib to consistently use unicode instead of bytes (==str in python2) anywhere a plist string can occur is too demanding for old python2 code...

I need to revise my approach in the new ufoLib.plistlib. Instead of forcing python2 to behave like python3 (where the text / binary data is clear), I need to implement the use_builtin_types option in plistlib.load/dump functions, and add support for the deprecated plistlib.Data wrapper class for bytes data.

Python 2 is my nemesis.

[miguelsousa]

Thanks for looking into this Cosimo. There will be a time when we don't have to think about Python 2 anymore.

[cjchapman]

@miguelsousa @anthrotype I think we should throw a party when that time comes.

[benkiel]

Python 2 is my nemesis.

And we can get t-shirts with this!

[miguelsousa]

Or: Python 2 is was my nemesis

[benkiel]

Or: Python 2 is was my nemesis

To state the obvious, the is is crossed out by a snake.

[khaledhosny]

Now I got an idea for my next geeky t-shirt, the backtrace of Python 2 encoding exception.

…