Re: [PATCH] Improved submodule merge support

From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Dec 18, 2007, 20:21 UTC
Message-ID: <Pine.LNX.4.64.0712182018190.23902@racer.site>
In-Reply-To: <20071218195028.GA25510@pvv.org>

Hi,

On Tue, 18 Dec 2007, Finn Arne Gangstad wrote:

Show 19 quoted lines

> diff --git a/merge-recursive.c b/merge-recursive.c
> index 2a58dad..33ccc40 100644
> --- a/merge-recursive.c
> +++ b/merge-recursive.c
> @@ -1463,10 +1467,13 @@ static int process_entry(const char *path, struct stage_data *entry,
>  		mfi = merge_file(&o, &a, &b,
>  				 branch1, branch2);
>  
> +		clean_merge = mfi.clean;
>  		if (mfi.clean)
>  			update_file(1, mfi.sha, mfi.mode, path);
> +		else if (S_ISGITLINK(mfi.mode))
> +			output(1, "CONFLICT (submodule): Merge conflict in %s "
> +			       "- needs %s", path, sha1_to_hex(b.sha1));
>  		else {
> -			clean_merge = 0;
>  			output(1, "CONFLICT (%s): Merge conflict in %s",
>  					reason, path);
>

It took me a little while to find that the unilateral assignment to clean_merge does not break things. But as it was only set to 1 at the beginning of the function, and no loops are involved, this change is correct.

Ciao, Dscho