Re: Bug Report: "git submodule deinit" fails right after a clone

From: Heiko Voigt <hvoigt@hvoigt.net>
Date: Oct 6, 2016, 09:20 UTC
Message-ID: <20161006092040.GA38550@book.hvoigt.net>
In-Reply-To: <CAPOqYV+SR8=E24HR18ygJ-J-rV7=FedU+gGpq9mp+a7Mfk2LiQ@mail.gmail.com>

Hi,

please also keep the mailinglist in the CC so everyone can read this.

On Thu, Oct 06, 2016 at 09:11:05AM +0200, Thomas Bétous wrote:

Show 16 quoted lines

> On Wed, Oct 5, 2016 at 3:36 PM, Heiko Voigt <hvoigt@hvoigt.net> wrote:
> 
> >
> > My initial reaction is that this might be a problem with line endings. Did
> > you
> > check whether you get any diff when you do a 'git diff' after the clone?
> >
> > Maybe the variable 'core.autocrlf' is set to 'input' ? Have a look at 'git
> > help
> > config'
> 
> 
> When I do a 'git diff' right after the clone, nothing appears.
> Moreover my global setting for core.autocrlf is true. (This was configured
> on purpose as I work on Windows whereas the repositories are hosted on an
> UNIX server.)

So I guess the same applies to 'git status'?

Show 5 quoted lines

> Nevertheless when I change core.autocrlf to 'input', the error disappears
> and I got the expected behavior for git submodule deinit...
> So I guess it is just a configuration problem but I do not understand why
> core.autocrlf should be set to 'input' to remove this error. Do you have
> any idea?

This is indeed strange. That's why I asked whether 'git diff' shows anything. I was suspecting that the .gitmodules is somehow checked out in UNIX format. And the fact that setting core.autocrlf to 'input' seems to fix it supports that.

I currently do not have access to a windows machine as the moment to test this. Copying the windows mailing list maybe someone over there can reproduce and help with the issue[1].

Cheers Heiko

[1] http://public-inbox.org/git/%3CCAPOqYV+xsrLk7y1hJYHZFY8OfkxVRwPcZBdqhdgrhThqdZysQA@mail.gmail.com%3E/