On 02/25/2015 02:25 AM, Kosta Arvanitis wrote:
Maintaining a list of black files seems like more housekeeping which could get overlooked. Have you looked into the current suite of test tools?
First off, I plan to systematically go through the tree and fix the style issues (though others are welcome to do so). When I do so, I will remove the files from the blacklist. I don't expect it will take very long until the blacklist is gone, but it still might be a matter of weeks.
During that period, I do not want others to accidentally cause regressions in the files that I have already cleaned up, thus I want this tool to be put into place as soon as possible. That requires a blacklist to prevent others from being immediately forced to clean up files that they are working on (though such efforts would be welcome).
As for performance, it can be run with the list of files that have been touched (i.e. from a 'diffstat -l'), so it can limit its check to files in a patch (and not check the whole tree). Style checking the whole tree currently takes less than 2 minutes on my underpowered netbook.
As for coverage, the pep8 tool is much more thorough than the existing style checks. Dan and I discussed this approach on IRC tonight, and these patches were driven directly out of that discussion.