> On May 14, 2014, 5:12 a.m., Dominic Hamon wrote: > > support/mesos-style.py, line 93 > > <https://reviews.apache.org/r/21360/diff/3/?file=581314#file581314line93> > > > > could we just run find_candidates on each of the file_paths in this > > case and avoid the cost of generating the candidates set in the first place?
While this is a good idea, a quick test suggests that generating "candidates" list is very fast in our repo. I propose we do this optimization in a follow up review, if necessary. Sounds good? - Vinod ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/21360/#review42943 ----------------------------------------------------------- On May 14, 2014, 4:24 a.m., Steven Phung wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/21360/ > ----------------------------------------------------------- > > (Updated May 14, 2014, 4:24 a.m.) > > > Review request for mesos and Vinod Kone. > > > Bugs: MESOS-1349 > https://issues.apache.org/jira/browse/MESOS-1349 > > > Repository: mesos-git > > > Description > ------- > > Previously run_lint was being invoked a large subset of the source tree which > was inefficient. With this change it enables mesos-style.py to potentially be > used as a pre-commit hook. > > git diff is used to determine which staged files that have been added or > removed. The intersection between the staged files and all lint candidates > is computed and sent to cpplint. The return value of the mesos-style.py > script is the result of the cpplint process exit or 0 if there the > intersection is empty. > > > Diffs > ----- > > support/mesos-style.py bd7dcdb > > Diff: https://reviews.apache.org/r/21360/diff/ > > > Testing > ------- > > ./support/mesos-style.py > > > Thanks, > > Steven Phung > >
