It kind of got lost through my email, but when I talked about force-push, it is almost completely in the context of feature-branch rebase.
We almost never do force-push outside of this context, neither do we encourage it. And like I pointed, the fact that the mainline branches (trunk, branch-2 etc) are already protected against force-pushes, and the latest infra’s change to notify force-pushes as explicitly marked so, are good enough for handling them IMO. Thanks +Vinod > On Jan 14, 2016, at 8:07 PM, Yongjun Zhang <yzh...@cloudera.com> wrote: > > I assume we will have at least a committer review of any force-push. Say, > if the commit history already has x, y, z, now we need to force push with > x', we'd better not torget y and z, and if there is conflict putting in y > and z after x', the conflict is better reviewed.