“gitk doesn’t get any information from remote repositories; it only shows you what’s in your local repository, so the position of origin/master in gitk’s display will be from the last time you updated it with a fetch, pull or push.”

So if I do git fetch , gitk won’t show any information about the state of that remote repository? i.e. I can’t actually see the history of that remote. I certainly do see lots of labels in gitk like “remotes/andy/master”. Here, what gitk is just showing me, then, is information about what remote ‘andy’ pushed to origin/master?

By “doesn’t get any information from remote repositories”, I meant “doesn’t fetch information over the internet when you run it”. It does show you the state of branches from remote repositories (as you’ve noted) from the last time you fetched from them because it’s showing remote-tracking branches like andy/master. (If you want to make sure that all of these are included, you can run gitk --all.)

Another general question/comment : I’ve always thought that “local” means “local to my machine”, and “remote” means on a physical remote server somewhere. But now I am getting this uneasy feeling that this is not always the intended sense in which these terms are used when describing git repositories…

Well, the “remote repository” might be another repository on the same computer, just in a different directory. You should also be aware that “a remote” (i.e. “remote” as a noun) in git terminology is also like an alias or nickname for the URL of a remote repository. For example, “origin” is a remote, as is “andy” in your example above – they might be on the same machine or a different one, depending on the URL.

“gitk doesn’t get any information from remote repositories; it only shows you what’s in your local repository, so the position of origin/master in gitk’s display will be from the last time you updated it with a fetch, pull or push.”

So if I do git fetch , gitk won’t show any information about the state of that remote repository? i.e. I can’t actually see the history of that remote. I certainly do see lots of labels in gitk like “remotes/andy/master”. Here, what gitk is just showing me, then, is information about what remote ‘andy’ pushed to origin/master?

Another general question/comment : I’ve always thought that “local” means “local to my machine”, and “remote” means on a physical remote server somewhere. But now I am getting this uneasy feeling that this is not always the intended sense in which these terms are used when describing git repositories…

So, if I understand this, we actually have an ‘origin/master’ stored on our local machine? And so when I do a fetch, all I am doing is updating the local copy of the “remote” branch “origin/master” (which isn’t really ‘remote’ , i.e. off on some distant server, but is actually a local copy of that distant server branch? Fetch now makes alot more sense to me!

That’s essentially right – if you have a look at the file .git/refs/remotes/origin/master you’ll see it just has a the object name (hash) of a commit in it. “git fetch” makes sure that all the objects necessary for that commit are fetched and in .git/objects before updating that file.

It would be interesting to see an article on which commands actually require an internet connection, and which ones don’t. Apparently, git fetch requires an internet connection.

Of the commands that are commonly used, it’s only really “git fetch” (and thus “git pull” as well), “git push” and “git clone” (if you give it a non-local URL), but there are others that are less frequently used, like “git submodule update”, “git ls-remote”, “git remote update”, etc.

And what about gitk? If I don’t do any fetches, does ‘gitk’ go out to the internet to get updates from remote branches? Or does it only look at what is in your local copy of “origin/master” (for example), e.g. whatever is stored in the .git directory.

gitk doesn’t get any information from remote repositories; it only shows you what’s in your local repository, so the position of origin/master in gitk’s display will be from the last time you updated it with a fetch, pull or push.

Thanks for a great article. One comment you made really stuck out for me , and cleared up (I think) a major confusion I was having with git :

“(I emphasize “locally” since some people see “origin/master” and assume that in some sense this branch is incomplete without access to the remote server – that isn’t the case.)”

So, if I understand this, we actually have an ‘origin/master’ stored on our local machine? And so when I do a fetch, all I am doing is updating the local copy of the “remote” branch “origin/master” (which isn’t really ‘remote’ , i.e. off on some distant server, but is actually a local copy of that distant server branch? Fetch now makes alot more sense to me!

It would be interesting to see an article on which commands actually require an internet connection, and which ones don’t. Apparently, git fetch requires an internet connection.

And what about gitk? If I don’t do any fetches, does ‘gitk’ go out to the internet to get updates from remote branches? Or does it only look at what is in your local copy of “origin/master” (for example), e.g. whatever is stored in the .git directory.