Re: [PATCH] push: don't guess at qualifying remote refs on deletion
- From
Carlos Martín Nieto <cmn@elego.de>
- Date
- Jul 5, 2012, 07:22 UTC
- Message-ID
- <1341472926.10752.5.camel@flaca.cmartin.tk>
- In-Reply-To
- <CABURp0rVPAvxP1sp_nmoNYd+F+OsvWeHgUAeo7-VTnQhdebFeg@mail.gmail.com>
On Wed, 2012-07-04 at 22:23 -0400, Phil Hord wrote:
Show 48 quoted lines
> > On Jul 4, 2012 3:22 AM, "Carlos Martín Nieto" <cmn@elego.de> wrote: > > > > On Tue, 2012-07-03 at 14:40 -0400, Jeff King wrote: > > > On Tue, Jul 03, 2012 at 02:34:59PM -0400, jonsmirl@gmail.com > wrote: > > > > > > > > and now: > > > > > > > > > > $ git push origin :bogus > > > > > error: unable to delete 'bogus': remote ref does not exist > > > > > error: failed to push some refs to '$URL' > > > > > > > > This error return would have made my mistake obvious. > > > > > > Thanks for confirming. > > > > > > > Might want to add a paragraph to the doc saying this is how you > delete > > > > remote branches since it is not an obvious solution. I found it > via > > > > Google and a question asked on stackoverflow.com > > > > > > It's already in git-push(1): > > > > > > OPTIONS > > > ... > > > <refspec> > > > ... > > > Pushing an empty <src> allows you to delete the <dst> ref > from > > > the remote repository. > > > > > > > There is also a flag you can pass, which you can see a few > paragraphs > > under it which. It explains what it does underneath but removes the > need > > to know that an empty source will delete the ref. > > > > --delete > > All listed refs are deleted from the remote repository. This > is > > the same as prefixing all refs with a colon. > > > I do like that, but I agree with OP that 'git push' is not an obvious > to look for delete branch functionality for new users.
Like it? I don't get what you mean, it's a quote of what's already there.
The reason that it's in git-push is because that's the only git command that modifies another repository. There isn't another git command where it would fit. You could try to put it somewhere under git-remote, but then you'd have a single subcommand that affects a remote among a lot of others that don't, introducing an inconsistency in the command.
cmn