Installation failed to run on python 3.x when command line encoding is not utf8 and readme_file contains non-ASCII characters

[ghost]

Originally reported by: felixonmars (Bitbucket: felixonmars, GitHub: felixonmars)

---

readme_file.read() will try to use 'ascii' to decode, and thus failed. For example, set LANG=C and get this:

#!python

Traceback (most recent call last):
  File "setup.py", line 93, in <module>
    long_description = readme_file.read() + '\n' + changes_file.read()
  File "/usr/lib/python3.3/encodings/ascii.py", line 26, in decode
    return codecs.ascii_decode(input, self.errors)[0]
UnicodeDecodeError: 'ascii' codec can't decode byte 0xc3 in position 7330: ordinal not in range(128)

I filed a patch for the same problem to python-Levenshtein, hope it helps:
ztane/python-Levenshtein#2

---

• Bitbucket: https://bitbucket.org/pypa/setuptools/issue/144

[ghost]

Original comment by felixonmars (Bitbucket: felixonmars, GitHub: felixonmars):

---

Note: The easiest workaround I'm currently using is:

#!bash
sed -i -e "s/open('\(README\|CHANGES\|CHANGES (links)\).txt')/open('\1.txt', encoding='utf8')/" setup.py

[ghost]

Original comment by jaraco (Bitbucket: jaraco, GitHub: jaraco):

---

Open readme with codecs to enable explicit encoding declaration. Fixes #144.

[ghost]

Original comment by jaraco (Bitbucket: jaraco, GitHub: jaraco):

---

Felix, can you confirm that 2.1.2 fixes the issue in that environment?

[ghost]

Original comment by felixonmars (Bitbucket: felixonmars, GitHub: felixonmars):

---

It works!

Thank you Jason, for the super fast fix :)