Re: [RFC/PATCH 3/4] revert: allow cherry-picking a range of commits

[Johannes Schindelin <Johannes.Schindelin@gmx.de>]

Hi,

On Sat, 29 May 2010, Christian Couder wrote:

> diff --git a/builtin/revert.c b/builtin/revert.c
> index 70372dc..c281a80 100644
> --- a/builtin/revert.c
> +++ b/builtin/revert.c
> @@ -545,6 +542,40 @@ static int revert_or_cherry_pick(int argc, const char **argv)
>  	if (read_cache() < 0)
>  		die("git %s: failed to read the index", me);
>  
> +	dotdot = strstr(commit_name, "..");
> +	if (dotdot) {
> +		struct rev_info revs;
> +		const char *argv[4];
> +		int argc = 0;
> +
> +		argv[argc++] = NULL;
> +		if (action != REVERT)
> +			argv[argc++] = "--reverse";
> +		argv[argc++] = commit_name;
> +		argv[argc++] = NULL;

Maybe "--no-merges"?

> +		init_revisions(&revs, NULL);
> +		setup_revisions(argc - 1, argv, &revs, NULL);
> +		if (prepare_revision_walk(&revs))
> +			die("revision walk setup failed");
> +
> +		if (!revs.commits)
> +			die("empty range passed");
> +
> +		while ( (commit = get_revision(&revs)) ) {

The style more in linet with the rest of the source code would be:

		while ((commit = get_revision(&revs))) {

The rest of the patch series looks very, very good to me. I totally agree 
that we do not have to implement the --abort and --continue for now, as 
well as the HEAD-detaching business we're used to from rebase and am.

Thanks,
Dscho