Re: [PATCH v3] git-add--interactive: manual hunk editing mode

From: Jeff King <peff@peff.net>
Date: Jun 9, 2008, 12:29 UTC
Message-ID: <20080609122938.GA12210@sigill.intra.peff.net>
In-Reply-To: <200806090746.22512.johan@herland.net>

On Mon, Jun 09, 2008 at 07:46:22AM +0200, Johan Herland wrote:

> Is there a good reason against having *both*?
> 
> AFAICS, there's nothing stopping us from having both a "-e"-option to 
> git-add, and an "e"-command inside git-add--interactive.

I agree (and I tried to make that point in an earlier mail).

And I was hoping the right way to do it was to simply build the interactive "e" command on top of Johannes' git-apply work. But I don't think that quite makes sense. His work is about fixing up the hunk header as we apply the patch, but a working "e" command in the hunk selection should probably not actually apply, but simply split into two hunks for the loop.

Show 5 quoted lines

> ("git-add -e" would open the entire diff in an editor, as would "e" from the 
> *main* menu of git-add--interactive. However, "e" from the *single hunk* 
> menu would of course open only that single hunk within the editor. We could 
> even have an "E" command to open all remaining/undecided hunks in an 
> editor.)

I agree with all of this, though I think the big question is what happens to the edited portion. In the interactive command, I think it becomes a new hunk that can be staged or not. In "git add -e" it makes sense to simply stage the result.

-Peff