https://github.com/bear/python-twitter
Revision ecf8f65c81e53186491a9f16b204d92ce383e417 authored by Tom Callaway on 26 January 2016, 18:33:05 UTC, committed by Tom Callaway on 26 January 2016, 18:33:05 UTC
+ /usr/bin/python3 setup.py build '--executable=/usr/bin/python3 -s'
Traceback (most recent call last):
  File "setup.py", line 40, in <module>
    read('AUTHORS.rst') + '\n\n' +
  File "setup.py", line 27, in read
    return f.read()
  File "/usr/lib64/python3.4/encodings/ascii.py", line 26, in decode
    return codecs.ascii_decode(input, self.errors)[0]
UnicodeDecodeError: 'ascii' codec can't decode byte 0xc5 in position 412: ordinal not in range(128)

This only fails on Python 3 (I tested with 3.4.3). Since we know the encoding of the files is UTF-8, and we aren't concerned about these open() calls for performance, being explicit with encoding='utf-8' and using io.open is the lazy fix that resolves this issue without breaking Python 2.
1 parent e0773b3
History
Tip revision: ecf8f65c81e53186491a9f16b204d92ce383e417 authored by Tom Callaway on 26 January 2016, 18:33:05 UTC
Unicode + Python = day drinking. That fact aside, there is a Python3 and unicode issue with the setup.py file, specifically, in how it opens and reads AUTHORS.rst:
Tip revision: ecf8f65
File Mode Size
doc
examples
testdata
tests
twitter
.gitignore -rw-r--r-- 363 bytes
.travis.yml -rw-r--r-- 234 bytes
AUTHORS.rst -rw-r--r-- 701 bytes
CHANGES -rw-r--r-- 21.5 KB
COPYING -rw-r--r-- 605 bytes
LICENSE -rw-r--r-- 11.1 KB
MANIFEST.in -rw-r--r-- 93 bytes
Makefile -rw-r--r-- 901 bytes
NOTICE -rw-r--r-- 109 bytes
README.rst -rw-r--r-- 6.3 KB
coveragerc -rw-r--r-- 23 bytes
get_access_token.py -rwxr-xr-x 2.7 KB
python-twitter.spec -rw-r--r-- 1.1 KB
requirements.testing.txt -rw-r--r-- 44 bytes
requirements.txt -rw-r--r-- 34 bytes
setup.cfg -rw-r--r-- 127 bytes
setup.py -rwxr-xr-x 2.1 KB
test.py -rw-r--r-- 152 bytes

README.rst

back to top