Re: [PATCH 3/3] git submodule update should give notice when run without init beforehand

From: Chris Packham <judge.packham@gmail.com>
Date: Sep 19, 2013, 10:25 UTC
Message-ID: <523AD117.20109@gmail.com>
In-Reply-To: <CALUzUxrf+ZGX8nQ0DVqYEyWto3Cos16VLTUfjsX99qnDLa=S6w@mail.gmail.com>

On 18/09/13 22:22, Tay Ray Chuan wrote:

Show 10 quoted lines

> Hi Chris,
> 
> I think you mentioned usability issues with git-submodule before on
> the git mailing list, so I thought you might be interested in taking a
> look at this patch. It's attached below, you can also view it at
> 
>   http://thread.gmane.org/gmane.comp.version-control.git/1379266703-29808-3-git-send-email-rctay89@gmail.com
> 
> I would be interested in hearing what you think.
>

The case I had was that an un-init'd submodule was still a directory on the file system so 'cd submodule' would work and it could go unnoticed that we hadn't entered the submodule.

Your change might have helped somewhat but the bigger problem was that the work-tree representation of an un-init'd submodule is a directory. There was never any thought to actually run 'git submodule update' because the developer in question thought they already had.

I think Jens' autoinit may be the right direction. It certainly would be what we'd want for our use-case at $dayjob.