still unclear on setting up a repository

I'm really unclear here if it's me, the code, or the docs...

first, create the "remote"

        % cg-admin-setuprepo remote.git

that works fine, now set up the "local"

        % mkdir local.git
        % cd local-git
        % cg-init -m "junk message why should I have to do this one, it's dumb"

Now here's where I'm unclear. I need to "push" this into the remote, so I try this:

        % cg-branch-add origin "$(cd ..; pwd)/remote.git#master"
        % cg-branch-ls

This seems to report the right thing. I have a branch "origin" that points at the remote repository. So far, I think I'm understanding what's happening. But now, go boom:

        % cg-push
        cg-push: where to push to?

Huh? How do I do the initial push? Not this way? Is it me, docs, or code that is broken here?

UPDATE...

aha... "cg-push origin" works. But "cg-push" is documented to default to "origin". Why doesn't it?

And then, another "go-boom" problem:

    % cg-fetch
    Hard links don't work - using copy
    Fetching head...
    cp: illegal option -- d
    usage: cp [-R [-H | -L | -P]] [-f | -i | -n] [-pv] src target
           cp [-R [-H | -L | -P]] [-f | -i | -n] [-pv] src1 ... srcN directory
    cg-fetch: unable to get the head pointer of branch master

Any chance you could make this more POSIX-like? I know you have a big warning in README.osx, but are you *really* needing all those weird GNU-isms? Maybe you could fall back to using rsync to copy exactly what you need?

Randal L. Schwartz wrote:

I think this was resolved some weeks ago. Perhaps you could try the latest cogito from git://git.kernel.org/pub/scm/cogito/cogito.git ?

There aren't that many. This particular one was a typo, iirc.

As for rsync; From the output you posted above, you were setting up both the repositories on your local machine. Using rsync for those cases would be pretty stupid.

On a side-note, could you please turn off your spamvertising auto-reply? It sends me some info about your perl-expertise and contact-numbers every now and then. It's getting a bit annoying.

Andreas> Randal L. Schwartz wrote:

Andreas> I think this was resolved some weeks ago. Perhaps you could try the Andreas> latest cogito from git://git.kernel.org/pub/scm/cogito/cogito.git ?

I'm grabbing nearly hourly.

    localhost:~/MIRROR/cogito-GIT % cg-status
    Heads:
       >master      22ff47e9b3c5fc8aa2efbc5ac8690b06b868ef6f
      R origin      22ff47e9b3c5fc8aa2efbc5ac8690b06b868ef6f

Error with this version.

Andreas> As for rsync; From the output you posted above, you were setting up Andreas> both the repositories on your local machine. Using rsync for those Andreas> cases would be pretty stupid.

rsync instead of cp if you really need "-d".

Andreas> On a side-note, could you please turn off your spamvertising Andreas> auto-reply? It sends me some info about your perl-expertise and Andreas> contact-numbers every now and then. It's getting a bit annoying.

If you email me directly, you'll get that note no more than once per month. Some day, I'll create a system to whitelist people that never want to see it again. Damn TUIT shortage. :)

Randal L. Schwartz wrote:

You're right. It was the '-b' option that was removed earlier (which you reported as well, ten days ago).

There are other problems with rsync. It would actually be faster to use the git-daemon on localhost than using rsync for operations between two local repositories.

It's not the amount of mail that annoys me. It's the principle of the thing. Perhaps you can use a different email when participating in list discussions? It's not exactly right to ask others to tell you they don't want your spam after all.

Dear diary, on Sun, Nov 20, 2005 at 02:14:47AM CET, I got a letter where "Randal L. Schwartz" <merlyn@stonehenge.com> said that...

Hmm, right. I changed the -d to -P, which is basically equivalent for our situation, and -P is supposed to be POSIX. Thanks.

If you don't want to receive mailing list traffic related to your mails directly, you can setup your mailer to adjust the 'mail-followup-to' header appropriately and you should stop getting it twice.

Dear diary, on Sun, Nov 20, 2005 at 01:40:20AM CET, I got a letter where "Randal L. Schwartz" <merlyn@stonehenge.com> said that...

The code was broken, cg-push wouldn't default to origin unless you already had an origin head, which you obviously can't have when you are going to push for the first time. Thanks, fixed.