I don't like the way that close the PR by a simple commit message without any code change. It doesn't give any positive feedback to PR author, and it could hurt the feeling of the PR author.
Normally, we could close the PR if the PR is merged. Either the PR author or the committer can do this work. If the PR is not merged, the committer should provide some suggestions and remind the PR author the PR could be closed for some solid reasons. It's important that we always encourage the contribution no matter if it's a big one or a small one, and help the contributor to do their job better. If the PR is not good enough, the maintainer could accept part of the change and do some changes on the base of PR as Justin said. Willem Jiang Blog: http://willemjiang.blogspot.com (English) http://jnn.iteye.com (Chinese) Twitter: willemjiang Weibo: 姜宁willem On Mon, Jan 30, 2017 at 12:48 PM, yukon <[email protected]> wrote: > Hi, > > My sincere apology for these two closed pull requests without adequate > explanations, as zhanhui said, we are deeply appreciating > efforts/contributions from community without any doubt. > > And I think we should improve the PR closing process and add the result to > this doc[1], below questions need discussions: > > 1. When we can close a PR? > 2. Who perform the closing operations? Committers or the PR author? > 3. If committers do the closing operations, how? With an empty commit or > other way? > > And as Justin said, committers should accept and fix the defective issue > rather than rejecting or asking the contributor to make it perfect, > actually we don't know we have change permissions about PR, so we always > guide contributors step by step to make PR better. > > [1] http://rocketmq.incubator.apache.org/docs/pull-request/ > > Happy Chinese New Year, > yukon > > On Mon, Jan 30, 2017 at 11:46 AM, Justin Mclean <[email protected]> > wrote: > > > Hi, > > > > I don't see a huge concern here but more conversation and more detailed > > reasons for closing PR is always good. Helping contributors getting their > > PRs into shape is even better. Even if a PR needs a little work, it could > > be accepted and fixed by an existing committer rather than rejected or > > asking the contributor to make it perfect. > > > > Thanks, > > Justin >
