# Confused about Git **URL:** \ **Category:** Rails\ **Created:** [July 4, 2008, 10:06pm UTC](https://www.ruby-forum.com/t/confused-about-git/142235 "2008-07-04T22:06:30Z")\ **Posts on this page:** 4\ **Page:** 1
**Author:** ![wemagor](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/w/8edcca/32.png) [@wemagor](https://www.ruby-forum.com/u/wemagor)\ **Post date:** [July 4, 2008, 10:06pm UTC](https://www.ruby-forum.com/t/confused-about-git/142235/1 "2008-07-04T22:06:30Z")
I always try to follow best practices, so I’m in the process of learning Git since the Rails community seems to have favored that over Subversion (I know the basics of svn as it stands right now). I have to admit, I like the idea of working locally and then pushing it up when I’m done, instead of always having to push it to a central server. The one thing that confuses me, however, is that with Subversion you have your central dev/staging server, and you commit and checkout to/from that, which enables you to use continuous integration or similar to make sure that the build is working. I like the distributed model of Git, but I’m disturbed that there’s no “main” server. I guess you can set up the server in the same way and push/pull to/from it just like subversion when you’re ready, but wouldn’t this cause issues if I do several commits on my local version (say, I’m working on implementing a new module) and then push it to my staging server? The revisions would be different because my local branch has more commits (for individual pieces of the module) and the “master” repository only has one (the finished module itself). That sounds like a good idea, since you can do several mini-commits locally, then push it all up to the staging environment when the module/task/feature is complete. I don’t see how this works within Git, unless Git doesn’t care about the revisions. Am I getting myself confused over nothing here? ---
**Author:** ![wemagor](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/w/8edcca/32.png) [@wemagor](https://www.ruby-forum.com/u/wemagor)\ **Post date:** [July 4, 2008, 10:41pm UTC](https://www.ruby-forum.com/t/confused-about-git/142235/2 "2008-07-04T22:41:40Z")
I’ve used Git a little in the past, but use Mercurial now. They are essentially the same in how they work though. When you push to the remote repository, all of your changesets get pushed. So if you’ve made no changes on the remote side, after a push, your git list should look exactly the same. Peace, Phillip ---
**Author:** ![wemagor](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/w/8edcca/32.png) [@wemagor](https://www.ruby-forum.com/u/wemagor)\ **Post date:** [July 4, 2008, 11:15pm UTC](https://www.ruby-forum.com/t/confused-about-git/142235/3 "2008-07-04T23:15:39Z")
So it doesn’t matter if I commit 6 times locally, and then do one “mega commit” to the server? I ask because I want to set up a staging server to test my application, but until I get that done I’ll use my local repository. So, say I’m working on a module for my application. I make sure that my git is up to date from the staging server’s repository, and begin to work: (Local) Commit #1 - Generated model for Foo and created initial migration Commit #2 - Model specifications for Foo Commit #3 - Controller specifications for Foo Commit #4 - Created views for Foo#index, Foo#show, Foo#new At this point, the Foo module is finished. My local repository has had four separate commits, while the staging server hasn’t been touched. It won’t cause any syncing problems if I push these changes up to the staging server’s copy, say with the message “Adding completed Foo module and specifications”? I could have sworn in SVN this would foul things up because the local copy would be at a different revision number (4 commits vs. 1). If this isn’t an issue in Git, then I have nothing to worry about 🙂 ---
**Author:** ![wemagor](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/w/8edcca/32.png) [@wemagor](https://www.ruby-forum.com/u/wemagor)\ **Post date:** [July 4, 2008, 11:25pm UTC](https://www.ruby-forum.com/t/confused-about-git/142235/4 "2008-07-04T23:25:49Z")
You’ve got the idea. Keep in mind that you don’t do a “mega commit” to the remote repository. In DSCM, no repository is any more or less than any other. You do your commits to your local, and when you’re satisfied that what you have is ready to be synchronized with the remote, you do a push. If there are no changesets in the remote repository, it’s just a matter of your local changesets getting pushed out. There won’t even be a message for it. No, no problems. That’s what DSCM is all about. If there were changesets in the remote, Git would ask you to resolve any conflicts. If there are no conflicts, Git will just merge the files for you. Enjoy. Peace, Phillip