On Tue, Sep 23, 2014 at 4:44 PM, Rohit Yadav <rohit.ya...@shapeblue.com> wrote: > Hi David, > > On 23-Sep-2014, at 8:31 pm, David Nalley <da...@gnsa.us> wrote: >> Where was the merge request for this huge merge to master? (it was at >> 50 commit emails, when it stopped sending, ) >> We have passed feature freeze for 4.5.0, so I am confused as why this >> was merged. Is there a reason not to revert all of this? > > This was the request: https://github.com/apache/cloudstack/pull/16 > And JIRA: https://issues.apache.org/jira/browse/CLOUDSTACK-7143 > We all get emails from Github PR (just re-checked) so you may find them on > the ML. >
Yes, GH PR is exactly like the Review Board emails in this particular aspect. My question is why is this merged into master rather than a feature branch, and why no [MERGE] email as per: https://cwiki.apache.org/confluence/display/CLOUDSTACK/Branch+Merge+Expectations > I was reviewing this Github pull request which came in couple of weeks ago. > This is well tested and successfully refactors the way we build systemvms. It > was only merged when it passed build tests and painful manual QA (I did on > KVM). Just want to mention that this is not my code or $dayjob work, I was > just trying to help out the contribution from community member(s). > > Pardon my ignorance but I was not aware about our official status of > master/4.5.0 that we’re already passed feature freeze. Can you confirm the > official stand/status on master/4.5.0 and when do we plan to cut the release > branch, and if we are changing anything with the way we do releases etc? > See: http://markmail.org/message/jctcystkzv4sqrvz Based on that thread, we decided not to branch 4.5.0 and instead keep master = 4.5.0 until much later in the schedule. Hence merging features into master is breaking feature freeze. > I see a lot of changes that comes in every other day on master. The pattern I > see is that developers check-in barely working feature and then find the > excuse to check-in more patches/commits as bug fixes where they actually are > trying to complete an incomplete feature. I think this is a chronic issue > which should be checked and everyone should make an effort to work in their > feature branch, work on their git-foo, rebase/fix-conflicts and send merge > request or Github pull requests or request using reviewboard. I think this > will allow all of us to participate in on-going efforts and would help > improve its quality using QA and code reviewing. > This is a separate issue; but I tend to agree with your analysis of how things happen. If you see this happening though, please speak up. As a committer you have veto power. Speaking frankly, we all have to be on guard for quality issues. --David
