rahmu wrote@rolf: Git is not at all as complex as it seems. Sure it takes some getting used to, but there's nothing really complicated about it. As soon as you get away from the centralized mindset, and start realizing how a git repo is nothing more than a tree, the rest is pretty straightforward :)
Did you try to do a push to a remote repo (with a working directory attached) ? :-)
I'm talking of a web developer scenario, where you have a local repo, and a repo on the server, and you want to push your changes to the server repo, then update the attached working directory so that you can see your changes on line (served by apache or whatever).
Here is what happens (I read about half of it...):
http://hans.fugal.net/blog/2008/11/10/git-push-is-worse-than-worthless/
The possible solutions are:
- Have a "hub", that is an intermediary bare repository with no working directory attached. You push to the "hub", then a script automatically pulls from the hub to the final repo and updates it.
- There seems to be post-push script that you can add in the hooks directory on the remote repository (it will be triggered when you do a push), which apparently fixes that behavior.
- If you are blessed with a public IP for your laptop, or whatever device you want to push from, just go on the server and do a pull from there instead.
When I use Mercurial, I do a push to the server, then on the server issue a "hg update" and I see my changes online.
I agree with you, Git is not as complicated as it first seems, it all seems easier to me than when I started, but having such a hackish setup for a push sounds like overkill for my needs.