py.builtin._totext does not work with python 2.x

[pytestbot]

• Bitbucket: https://bitbucket.org/pytest-dev/py/issue/75
• Originally reported by: @sirex
• Originally created at: 2016-01-04T16:56:52.988

Steps to reproduce:

Run this in Python 2.7:

# coding: utf-8

import py

py.builtin._totext('bū')

Ant this will raise UnicodeDecodeRerror.

[pytestbot]

Original comment by @sirex

Sorry for the AttributeError, fixed the description. Yes, if uncode receives str it tries to decode it with ascii.

If _totext is equivalent for u'' literal, it should decode str type from utf-8. At least for Python 3.0 it does the correct thing.

[pytestbot]

Original comment by @The-Compiler

That will raise AttributeError 😉

py.buitin._totext('bū') will be equal to unicode('bū'), which indeed does the same.
I'm guessing because python 2 tries to encode the string as ascii and then decode it as utf-8? I'm not very familiar with the str/unicode model on python 2 though.

[RonnyPfannschmidt]

totext works as intended - its not a replacement for u'' strings - its casting to text be it unicode on python2 or native on python3