On Fri, Jul 18, 2008 at 6:25 AM, Olivier Galibert <galibert@pobox.com> wrote:If the patches have problems, then make a new topic branch and redo them there, or reorder, amend, squash, whatever inside the original branch. The problem comes once you open that branch out for other people to pull from, or if that branch had pulled patches from someone else. At that point you're destroying the history that other people are now already sharing, in a meta-data sense (changing the sha1), as well as invalidating whatever test points that people may have had on those commits. In short, don't rewrite public history. If you want to share out your works-in-progress, just let people know that those branches of yours are volatile and not to be trusted. But you can't pull someone else's public history, and then make it volatile; subtle trouble ensues. --
| Bart Van Assche | Integration of SCST in the mainstream Linux kernel |
| Kamalesh Babulal | [BUG] Linux 2.6.25-rc2 - Kernel Ooops while running dbench |
| Greg Kroah-Hartman | [PATCH 005/196] Chinese: add translation of SubmittingDrivers |
| Paul Jackson | Re: cpuset-remove-sched-domain-hooks-from-cpusets |
git: | |
| Gerrit Renker | [PATCH 0/37] dccp: Feature negotiation - last call for comments |
| Jarek Poplawski | [PATCH] pkt_sched: Destroy gen estimators under rtnl_lock(). |
| Steven Rostedt | Re: -rt scheduling: wakeup bug? |
| David Miller | Re: [GIT]: Networking |
