git-remotes.Rmd

# Remotes {#git-remotes}

Remote repositories are versions of your project that are hosted on the
Internet or another network. A single project can have 1, 2 or even hundreds of
remotes. You pull others changes from remotes and push your changes to remotes.

```{r setup, include = FALSE}
has_bash <- Sys.which('bash') != '' && .Platform$OS.type != 'windows'
```

## Listing what remotes exist

`git remote` lists the names of available remotes, but usually it is more
useful to see what URLs each note corresponds to (with `-v`).

```{bash eval = has_bash}
git remote -v
```

## Adding a new remote

`git clone` automatically adds a new remote, so often you do not need to do
this manually initially. However, after the initial clone, it is often useful to
add additional remotes.

Use `git remote add` to add a new remote

```shell
git remote add happygit https://github.com/jennybc/happy-git-with-r.git
```

Note: when you add a remote you give it a nickname (here `happygit`), which you can use in git commands in place of the entire URL.

```shell
git fetch happygit
```

Sidebar on nicknames: there is a strong convention to use `origin` as the nickname of your main remote. At this point, it is common for the main remote of a repo to be hosted on GitHub (or GitLab or Bitbucket). It is tempting to use a more descriptive nickname (such as `github`), but you might find that following convention is worth it. It makes your setup easier for others to understand and for you to transfer information that you read in documentation, on Stack Overflow, or in blogs.

A common reason to add a  second remote is when you have done a "fork and clone" of a repo and your personal copy is set up as the `origin` remote. Eventually you will want to pull changes from the original repository. It is common to use `upstream` as the nickname for this remote.

## Fetching data from remotes

To get new data from a remote use `git fetch <remote_name>`. This retrieves the
data locally, but importantly it does _not_ change the state of your repository
or your files in any way. To incorporate the data into your repository, you need to merge or rebase your project with the remote project.

```shell
# Fetch the data
git fetch happygit

# Now merge it with our local master
git merge happygit/master master

# git pull is a shortcut which does the above in one command
git pull happygit master
```

For more detail on `git pull` workflows, see \@ref(pull-tricky).

## Pushing to remotes

Use `git push <remote> <branch>` to push your local changes to the `<branch>`
branch on the `<remote>` remote.

```shell
# push my local changes to the origin remote's master branch
git push origin master

# push my local changes to the happygit remote's test branch
git push happygit test
```

## Renaming and changing remotes

`git remote rename` can be used to rename a remote

```shell
git remote rename happygit hg
```

`git remote set-url` can be used to change the URL for a remote. This is
sometimes useful if you initially set up a remote using https, but now want to
use the SSH URL instead (or vise versa).

```shell
git remote set-url happygit git@github.com:jennybc/happy-git-with-r.git
```

One fairly common workflow is you initially cloned a repository on GitHub
locally (without forking it), but now want to create your own fork and push
changes to it. As described earlier, it is common to call the main repository `upstream` and to call your fork `origin`. So, in this case, you need to first rename the existing remote (from `origin` to `upstream`). Then add your fork as a new remote, with the name `origin`.

```shell
git remote rename origin upstream
git remote add origin git@github.com:jimhester/happy-git-with-r.git
```

## Upstream tracking branches

It is possible to set the branch on the remote each of your local remotes
corresponds to. `git clone` sets this up automatically, so for your own master
branch this is not something you will run into. However by default if you
create a new branch and try to push to it you will see something like this.

```shell
git checkout -b mybranch
git push
# fatal: The current branch foo has no upstream branch.
# To push the current branch and set the remote as upstream, use
#
#     git push --set-upstream origin foo
```

You can do as the error message says and explicitly set the upstream branch
with `--set-upstream`. However I would recommend instead changing the default
behavior of `push` to automatically set the upstream branch to the branch with
the same name on the remote.

You can do this by changing the git `push.default` option to `current`.

```shell
git config --global push.default current
```

See also Working with Remotes:

<https://git-scm.com/book/en/v2/Git-Basics-Working-with-Remotes>