On 10/03/2014 04:26 PM, CrÃstian Viana wrote:
"make check-local" failed for me with this patch:
/bin/pep8 --version 1.5.6 /bin/pep8 --filename '*.py,*.py.in' --exclude=""src/kimchi/i18n.py,src/kimchi/websocket.py,src/kimchi/websockify.py"" . ./tests/test_config.py:51:80: E501 line too long (90 > 79 characters) ./tests/test_config.py:52:80: E501 line too long (102 > 79 characters) ./tests/test_config.py:73:80: E501 line too long (105 > 79 characters) ./src/kimchi/config.py:96:80: E501 line too long (80 > 79 characters) ./src/kimchi/config.py:140:80: E501 line too long (99 > 79 characters) make: *** [check-local] Error 1 I did see those errors before but as those files are listed in .gitignore, I thought that wouldn't matter. Especially because they basically contain configuration values, not proper source code (i.e.
On Sex, 2014-10-03 at 15:25 -0300, Aline Manera wrote: those values could be in an external config file, like XML, because they're plain values, instead of replacing a Python source file). Looking at it more closely, I see that they are generated by their corresponding .in files and we should actually fix them. But those files - including the bad lines reported above - are generated by injecting text into them automatically and we cannot make sure those paths are PEP8-formatted after appending text inside an existing file. Well, maybe we could, but that would be a big, worthless effort.
The files .py.in are already being checked by PEP8 so the original source code prior to the text replacement is valid. The best solution I see now is to add those two files (i.e. src/kimchi/config.py and tests/test_config.py) to the blacklist as well, while still checking their original .in files. What do you think?
ACK.