[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[gits] Branch Help

For the benefit of everyone else: off-list, what we discovered was that the master branch was getting rebased onto the topic branch (aptui), rather than vice versa. This resulted in extra ?copies? of the commits on master also showing up in the aptui branch. When trying to merge, these duplicated commits were causing problems. The solution was to make a new topic branch that contained only the commits that were supposed to be there, omitting all of the duplicated commits that were already in the master branch.

How exactly this happened, we are not sure.

On Jan 8, 2014, at 3:00 PM, Gary Wong <gtw at flux.utah.edu> wrote:

> On Wed, Jan 08, 2014 at 01:32:43PM -0800, Leigh Stoller wrote:
>>> FYI, I?m working on preparing for class, I?ve handed this over to
>>> Gary. He will push some new branch to emulab-devel that has just the
>>> commits that are supposed to be on the aptui branch
> OK, I've made a first stab at this (just in my public repo for now).
> Anybody who is curious, have a look at branch "temp-aptui" from this
> repository:
>    git-public.flux.utah.edu:/flux/git/users/gtw/testbed.git
> The HEAD is in almost exactly the same state as aptui from emulab-devel,
> (except for the permissions on protogeni/scripts/quickvm.in -- can be
> changed later if anybody cares).  You can do a "git diff aptui temp-aptui"
> to verify that nothing has been lost.
> Also, the new "temp-aptui" branch has a clean history, as a descendent
> from emulab-devel master commit 67569c2b (18th December... the last
> time the old aptui branch was rebased onto emulab-devel master).  In
> other words, everything before that commit is straight out of mainline
> master (containing no aptui work), and everything since that commit
> is specific to aptui (independent of master).
> If everybody is happy with the way it looks, it will be easy to publish
> it somewhere final under a proper name, and it should also merge into
> master without much difficulty.
> Bye for now,
> Gary.
> -- 
>       Gary Wong      gtw at flux.utah.edu      http://www.cs.utah.edu/~gtw/