I agree with Roman. There’s not much information content in a PR. For example, take https://github.com/pivotalsoftware/quickstep/pull/177 <https://github.com/pivotalsoftware/quickstep/pull/177>.
As a committer, you can do the following: $ git remote add pivotalsoftware https://github.com/pivotalsoftware/quickstep.git <https://github.com/pivotalsoftware/quickstep.git> $ git fetch pivotalsoftware $ git checkout master $ git merge —ff-only pivotalsoftware/use-cache $ git push origin master The pull request tells you the source branch (pivotalsoftware/use-cache) plus the fact that the requester consented to contribute the code to Apache. It doesn’t matter that you put the pull request into apache/quickstep rather than pivotalsoftware/quickstep, as long as you have consent. Julian > On Apr 18, 2016, at 6:48 PM, Roman Shaposhnik <[email protected]> wrote: > > On Mon, Apr 18, 2016 at 6:44 PM, Jignesh Patel <[email protected]> wrote: >> Dear Roman and Julian, >> >> It will be a pain as we have to sync back in the same order as there are >> dependent PRs. > > Please give me an example of something that may be painful. > >> The cloning came somewhat of a surprise and caught us in the middle of a >> sequence >> of PR that are targeting a milestone that is about 10 days off (getting >> TPC-H running >> end-to-end before the students get off to the end-of semester sprint). > > Sure. But the cloning has little to do with PRs. Basically there's a > state of the repo > on ASF and Your side. That state needs to be synchronized (manually) one last > time right before you all decide to cut over and start using ASF Git > as a single source > of truth. That part is super simple and has little to do with PRs. > > Thanks, > Roman.
