Certainly git can do the same "try a merge of branch A into master, and then try a merge of branch B into that and use that as the diff". And if rbt can take just a "diff" to post, then we should be able to publish that. But I don't know where the data of *what* branches are involved would be stored, as Reviewboard seems to just think in terms of diffs. (rbt -u just guesses what review is associated to your branch by looking at the summary messages, there is no direct link to a branch AFAICT)
John =:-> On Thu, Sep 18, 2014 at 2:45 PM, Ian Booth <[email protected]> wrote: > > > On 18/09/14 20:39, John Meinel wrote: > > So I did get "rbt post" to work with "rbt post -r 54 --parent=REVID", I > had > > to be careful because my actual parent branch *didn't* merge the current > > tip of master but the one I was proposing had. > > So instead I ended up rebasing both commits, and then specifying the > PARENT > > as the specific rebased commit revision id. > > > > Launchpad had a place to record a prerequisite branch, which I'm sure is > > what lbox was using. Launchpad also computed the diff by starting with > > "master" merging the prerequisite branch into it, and then your diff was > > the diff of merging your branch into the combined diff. Which meant it > > handled all these side cases where your branch merges more of master than > > your parent branch, etc. > > > > And that, ladies and gentlemen, is why I *really* miss Launchpad. Doing > what > John did sounds error prone and I'm sure I'd mess it up if I tried. > > Eric, can we do something on the review board and/or github side to make > this > sort of thing easier? Implement a script which follows the Launchpad > workflow or > something? > >
-- Juju-dev mailing list [email protected] Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/juju-dev
