Hmm... Well there are a few issues. First I think we are talking about different things. I am referring to the commit messages that come through in 50 parts.
Secondly, there needs to be JIRA's to track significant changes. And commit messages need to reference those JIRA's or make clear why not. Otherwise, it is impossible to track which changes apply to which JIRA's. Your suggestions for making a JIRA to reflect the pull request and link the JIRA in the commit is an excellent step. I would add that PR's should be merged into master separately rather than with many issues together. On Sat, Apr 4, 2015 at 1:39 AM, Li Yang <[email protected]> wrote: > Hi Ted > > You must mean these merges. Correct me if not. > - > > https://github.com/apache/incubator-kylin/commit/eede2806e96d07ea2c99493f350508741e3f31ec > - > > https://github.com/apache/incubator-kylin/commit/d6e4a70dbef7975f28cf09efd1be86541d363431 > - > > https://github.com/apache/incubator-kylin/commit/6b5eec91c0f31d6b7ad30b2a390cfd6327eb42e3 > - > > https://github.com/apache/incubator-kylin/commit/43cf0a4953dcbe9a867765c105b35917fc5a1770 > - > > https://github.com/apache/incubator-kylin/commit/23bb56302ab01070d8a63f3fd8cbe3655faaaace > > These are my doing to merge below 3 pull requests from the public. They are > all pretty small by the way. > - https://github.com/apache/incubator-kylin/pull/2 (3 lines doc > change) > - https://github.com/KylinOLAP/Kylin/pull/452 (KYLIN-657 > <https://issues.apache.org/jira/browse/KYLIN-657>) > - https://github.com/KylinOLAP/Kylin/pull/451 (KYLIN-657 > <https://issues.apache.org/jira/browse/KYLIN-657>) > > I thought GIT would automatically make clear of the original author and me > the reviewer, since it's merged from pull request. Apparently the result is > confusing.. > > Kindly educate on how these pull requests should be processed in line with > Apache standard. What I can think of is > - Create JIRA for each pull request, copy the discussion from pull request > there. > - Link the JIRA, the original author and reviewer in the commit comment. > > Suggestion? Or is there a good example I can look at? > > Thanks > Yang > > > On Sat, Apr 4, 2015 at 7:34 AM, Ted Dunning <[email protected]> wrote: > > > I am still seeing large merged commits going into the Kylin source. This > > is a serious problem on several grounds: > > > > 1) the source of these commits is obscured by the bulk nature of the > > commits. The source needs to be identified down to the person who wrote > > the code so that the project can be sure that all contributions were > > licensed correctly. > > > > 2) the JIRA's that these commits are associated with is not recorded. > This > > is important because the community needs a record of what code was > written > > for what purpose. > > > > 3) the community is not apparently being involved in reviewing these > > commits. That raises all kinds of flags about development proceeding off > > the mailing list. > > > > These concerns are very serious relative to the Apache standards for > > projects. > > >
