Hi,
On Sat, 24 Aug 2013, Ian Jackson wrote:
> Did you know that dpkg-source doesn't notice if sensible-editor exits
> nonzero ?
I saw it when looking up the exact details to write my former reply. But
since at that point all the changes are already applied (and the editor
only allows you to change the automatic header in something more sensible),
I'm not sure what a proper failure would bring.
If we want to abort the whole commit operation, like git would do, it
will need a lot more work for very little gain IMO.
> > And I guess that you're not allowed to push anything in the dgit/*
> > branches (on dgit.debian.net) except when you do a real upload.
>
> Those branches aren't in refs/heads/ precisely to stop git pushing
> them by itself.
Oh, this is pretty confusing... because we do have dgit/<suite> branches
in refs/heads/ after a "dgit clone".
What you mean is that on the server side the dgit-branches containing the
uploads are in refs/dgit/ and not in refs/heads/. But there will also
be refs/heads/dgit/<suite> branches on the server side since that's where
people will naturally push stuff:
$ dgit clone dgit
$ cat .git/HEAD
ref: refs/heads/dgit/sid
$ cat .git/config
…
[remote "dgit"]
fetch = +refs/dgit/*:refs/remotes/dgit/dgit/*
[remote "origin"]
url = git+ssh://git.debian.org/git/dgit-repos/dgit.git
fetch = +refs/heads/*:refs/remotes/origin/*
$ git push --dry-run
To git+ssh://git.debian.org/git/dgit-repos/dgit.git
* [new branch] dgit/sid -> dgit/sid
So there's a refs/head/dgit/<suite> and it can be pushed to the server.
But it will be different to the refs/dgit/<suite> that gets merged by
dgit pull...
Speaking of braindamage... :-)
Cheers,
--
Raphaël Hertzog ◈ Debian Developer
Discover the Debian Administrator's Handbook:
→ http://debian-handbook.info/get/
--
To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20130824222618.ga13...@x230-buxy.home.ouaza.com
