Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [egit-dev] EGit / Gerrit / Refspec enhancement


On 12.03.14 10:08, "Sascha Vogt" <FunkyFish@xxxxxxx> wrote:

>Hi all,
>
>sorry for the strange subject, but I didn't know how to put it in a
>single line.
>
>Here is the situation, and I would love to hear how others deal with
>that and what may be done in EGit to enhance it.
>
>We use Gerrit and Mylyn and so the default refspec is
>HEAD:refs/for/master - this is great for normal "review"-operations.
>
>We also have a sandbox namespace in Gerrit where we can put our
>prototypes to sandbox/${username}/*
>
>Now, the HEAD:refs/for/master refspec fails, because obviously sandbox/*
>stuff should not be pushed to for/master but to the remote equivalent.
>So what I would LIKE to use is something like
>refs/heads/sandbox/*:refs/remotes/origin/sandbox/*
>refs/heads/topic/*:refs/for/master

Do you want that the users push all the local branches when they do a push
to
upstream or just the checked out one? Such a ref spec would push all the
branches. 

Or do you want that the push pushes the current branch, but according
to some rule specified in the config? AFAIK this is not possible with a
push ref spec.

What is possible is to set the push.default to upstream and explicitly
create
an entry branch."name".merge = refs/for/master or
refs/remotes/origin/sandbox/"name" for each new branch
(--set-upstream on the command line). But how should this be created
automatically?

But this

Stefan

>
>and maybe even
>refs/heads/maintenance/*:refs/for/*
>
>Obviously for a normal Git server this doesn't make sense at all, you
>can't push multiple branches to the same remote branch, but for the
>virtual for namespace of Gerrit this makes perfect sense.
>
>Would that be useful to others as well? In our case (we have a bunch of
>developers not yet used to Git) this would allow us to distribute a
>general purpose refspec config which then the devs (as long as the
>respect a few rules about the local branch names) wouldn't have to
>tinker with the refspecs at first.
>
>Greetings
>-Sascha-
>_______________________________________________
>egit-dev mailing list
>egit-dev@xxxxxxxxxxx
>https://dev.eclipse.org/mailman/listinfo/egit-dev



Back to the top