Re: [PATCH RFC] Use "git fetch" when cloning

From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Nov 21, 2007, 16:04 UTC
Message-ID: <Pine.LNX.4.64.0711211601540.27959@racer.site>
In-Reply-To: <1195659613-5290-1-git-send-email-sbejar@gmail.com>

Hi,

On Wed, 21 Nov 2007, Santi Béjar wrote:

> 1) It always use "git fetch" even if local. So now it checks the validity
>    of the new repo, in particular it does not create a broken repository
>    when there are too many nested alternates.

I doubt that this is a good change. It will make local clones much, much more expensive now.

Show 5 quoted lines

> 2) It connects with the remote repo when fetching the branches/tags
>    and the remote HEAD.
> 
> 3) The quiet mode of "git fetch" prints the changed branches/tags, so "git clone"
>    prints always the new branches/tags.

These changes would merit a split so that we could cherry-pick what we want, and leave what we don't want.

Show 15 quoted lines

> diff --git a/t/t5710-info-alternate.sh b/t/t5710-info-alternate.sh
> index 699df6e..9b7f638 100755
> --- a/t/t5710-info-alternate.sh
> +++ b/t/t5710-info-alternate.sh
> @@ -53,7 +53,7 @@ git prune'
>  
>  cd "$base_dir"
>  
> -test_expect_failure 'creating too deep nesting' \
> +test_expect_success 'creating too deep nesting' \
>  'git clone -l -s C D &&
>  git clone -l -s D E &&
>  git clone -l -s E F &&
> -- 
> 1.5.3.5.next.428.g67f5d-dirty

What on _earth_?

After thinking a bit harder about it, _and_ reading the commit message, I understand why it succeeds now.

But what would you _yourself_ think when reading this part of the code in half a year? Exactly.

Ciao, Dscho