Re: Branch Management

[Bagas Sanjaya <bagasdotme@gmail.com>]

On 11/08/21 05.24, mail.chris.az@gmail.com wrote:
> I have a private git repository (private.git) on a private server
> (PRIVATEREPO.COM).  That repo is updated via development on my local system.
> All works great in that area.  I have 2 branches, "master" and "version".
> 
> I want to update my production system via git from the private git
> repository via command line.  The production system not only contains the
> files from the repo, but also additional files and directories.  I need
> those files to be left alone.
> 

You can specify which files and directories you want to ignore in 
.gitignore. For example, based on your use case:

```
/var/prod/
/etc/ohmyapp/ohmyapp.conf
...
```

> I was able to successful clone just the "master" branch using:
> 
> # git clone -b master git_user@PRIVATEREPO.COM:private.git production.git
> 
> I am now ready to move the new "version" branch.
> 1.  How can I do that?
> 

Use `git checkout version`.

> The "version" branch contains some new files that are not in the "master"
> branch.
> 
> Now suppose that I move to the "version" branch, but determine that I need
> to go back to the "master" branch.
> 
> 2.  How do I go back to the "master" branch and remove any of the new files
> that were uploaded from the "version" branch?  I don't want any remnants
> that the "version" branch was even there.
> 

Practically speaking, `git checkout` updates working directory to 
revision specified in the argument, ignoring `.gitignore`.

I assumed that you have app+default config committed to your repo (not 
including /var-like production data of course). In that case, a simple 
`git checkout` will do most of the job; you need to figure out how to 
handle .gitignore-ed trees (since these maybe version-specific).

-- 
An old man doll... just what I always wanted! - Clara