threads / discuss / 28302

Pushing with --mirror over HTTP?

Subject: Pushing with --mirror over HTTP?

## tl;dr

4 messages between Sep 5, 2011 and Dec 19, 2011.

replies: 3people: 2as markdown or json

Eli Barzilay· Sep 5, 2011, 04:05 UTC · lore

Is there anything broken with pushing with mirror over HTTP? I'm trying that with a github url, and I get a broken-looking error message:

  remote part of refspec is not a valid name in :.have

and with the google code, I get:

  error: unable to push to unqualified destination: HEAD

Pushing to both of these work fine without `--mirror'.

(BTW, as a workaround, I'm using
  push --force --tags <url> :
is this achieving the same effect for a repo without weird refs?)

-- 
          ((lambda (x) (x x)) (lambda (x) (x x)))          Eli Barzilay:
                    http://barzilay.org/                   Maze is Life!

Jeff King· Sep 7, 2011, 21:39 UTC · re: Eli Barzilay · lore

Re: Pushing with --mirror over HTTP?

On Mon, Sep 05, 2011 at 12:05:37AM -0400, Eli Barzilay wrote:

Show 5 quoted lines

> Is there anything broken with pushing with mirror over HTTP?  I'm
> trying that with a github url, and I get a broken-looking error
> message:
> 
>   remote part of refspec is not a valid name in :.have

It's probably nothing to do with http, but rather with alternate object databases on the server (which GitHub uses heavily). The server hands out fake ".have" refs telling you it has some other branch tips to base packs off of. So I suspect the "push --mirror" code is simply wrong for trying to update those refs (it may be exacerbated by using http, though, as the remote helper code seems to have some extra checks).

Show 5 quoted lines

> and with the google code, I get:
> 
>   error: unable to push to unqualified destination: HEAD
> 
> Pushing to both of these work fine without `--mirror'.

This one, I'm not sure. It may be related.

> (BTW, as a workaround, I'm using
>   push --force --tags <url> :
> is this achieving the same effect for a repo without weird refs?)

Not quite. I think:

  git push --force <url> refs/*:refs/*

would be closer. But even that's not quite right. I believe that "--mirror" will also delete any remote refs that don't exist locally (which is why you are seeing the ":.have" refspec above, which attempts to delete it).

-Peff

Eli Barzilay· Sep 8, 2011, 02:58 UTC · re: Jeff King · lore

Re: Pushing with --mirror over HTTP?

5 hours ago, Jeff King wrote:

Show 15 quoted lines

> On Mon, Sep 05, 2011 at 12:05:37AM -0400, Eli Barzilay wrote:
> 
> > Is there anything broken with pushing with mirror over HTTP?  I'm
> > trying that with a github url, and I get a broken-looking error
> > message:
> > 
> >   remote part of refspec is not a valid name in :.have
> 
> It's probably nothing to do with http, but rather with alternate
> object databases on the server (which GitHub uses heavily). The
> server hands out fake ".have" refs telling you it has some other
> branch tips to base packs off of. So I suspect the "push --mirror"
> code is simply wrong for trying to update those refs (it may be
> exacerbated by using http, though, as the remote helper code seems
> to have some extra checks).

Ah -- I thought that this was some result of parsing some text message or something like that, maybe if the error was

  remote part of refspec is not a valid name in ":.have"

or even

  remote part of refspec is not a valid name in: :.have

it would have been clearer? Seems like it's a good place for this since some `:foo' is likely to appear there, and the colon can be confused as part of the text.

Also, maybe the man page should say something about `--mirror' not working well with such servers? It looks to me like mirroring to github and to google code would be pretty popular.

Show 17 quoted lines

> > and with the google code, I get:
> > 
> >   error: unable to push to unqualified destination: HEAD
> > 
> > Pushing to both of these work fine without `--mirror'.
> 
> This one, I'm not sure. It may be related.
> 
> > (BTW, as a workaround, I'm using
> >   push --force --tags <url> :
> > is this achieving the same effect for a repo without weird refs?)
> 
> Not quite. I think:
> 
>   git push --force <url> refs/*:refs/*
> 
> would be closer.

Thanks -- I'll use that instead.

> But even that's not quite right.  I believe that "--mirror" will
> also delete any remote refs that don't exist locally (which is why
> you are seeing the ":.have" refspec above, which attempts to delete
> it).

Is there some way of doing that? (We do use a branch during releases that is deleted after the release, so I need to propagate these deletions.)

-- 
          ((lambda (x) (x x)) (lambda (x) (x x)))          Eli Barzilay:
                    http://barzilay.org/                   Maze is Life!

Jeff King· Dec 19, 2011, 17:12 UTC · re: Eli Barzilay · lore

Re: Pushing with --mirror over HTTP?

On Wed, Sep 07, 2011 at 10:58:16PM -0400, Eli Barzilay wrote:

Show 16 quoted lines

> 5 hours ago, Jeff King wrote:
> > On Mon, Sep 05, 2011 at 12:05:37AM -0400, Eli Barzilay wrote:
> > 
> > > Is there anything broken with pushing with mirror over HTTP?  I'm
> > > trying that with a github url, and I get a broken-looking error
> > > message:
> > > 
> > >   remote part of refspec is not a valid name in :.have
> > 
> > It's probably nothing to do with http, but rather with alternate
> > object databases on the server (which GitHub uses heavily). The
> > server hands out fake ".have" refs telling you it has some other
> > branch tips to base packs off of. So I suspect the "push --mirror"
> > code is simply wrong for trying to update those refs (it may be
> > exacerbated by using http, though, as the remote helper code seems
> > to have some extra checks).

Sorry for the very delayed response on your bug, but at least I have good news. :)

It should be fixed by:

  http://article.gmane.org/gmane.comp.version-control.git/187373

(I was trying to fix another bug there, but see my followup for a discussion of .have).

-Peff