Re: [PATCH v4] builtin/add.c: replace run_command() with direct apply_all_patches() call

From: Junio C Hamano <gitster@pobox.com>
Date: Aug 26, 2026, 17:16 UTC
Message-ID: <xmqqbjaoiyzx.fsf@gitster.g>
In-Reply-To: <xmqq8q6to4em.fsf@gitster.g>

Junio C Hamano <gitster@pobox.com> writes:

Show 5 quoted lines

> Now the way "apply" API is used in this new code path should be
> pretty much parallel to existing "git apply" and "git am" code
> paths, we should be fine.  I do not use "git add -e", but those who
> do who may care more more deeply about keeping this feature working
> than I do may want to lend an extra pair of eyes on this patch.

And nobody seems to be interested in seeing this topic move forward, unfortunately. After reading the patch again, I think this is safe and correct, and if I merge the topic, one of three things can happen.

 (1) the patch does not regress anything unexpectedly, or

 (2) the patch breaks "add -e" completely but the feature is not
     used by anybody and nobody will notice, or

 (3) the patch breaks "add -e" and the its users will start
     complaining too late.

Reverting a merge would not be too involved as the patch is small-ish, so even in case (3) it won't be too much trouble to deal with fallouts. So let me mark the topic for 'next' for now.

It still is not too late for "add -e" users to interject, though.

Thanks.