Re: [PATCH 1/3] Add "partial commit" tests during a conflicted merge

From: Junio C Hamano <gitster@pobox.com>
Date: Jan 23, 2009, 07:16 UTC
Message-ID: <7vab9i331g.fsf@gitster.siamese.dyndns.org>
In-Reply-To: <49796D0C.5070408@viscovery.net>

Johannes Sixt <j.sixt@viscovery.net> writes:

Show 10 quoted lines

>> +test_expect_success 'reject --only during a merge' '
>> +	git checkout HEAD^0 &&
>> +	git reset --hard the-other-side-says-nitfol &&
>> +	test_must_fail git merge one-side-says-frotz &&
>> +	echo yomin-only >file &&
>> +	test_must_fail git commit -m merge --only file &&
>
> I don't see why this must fail: 'file' is the only file that is different
> from HEAD. Yes, currently we fail; but if something is about to be
> changed, then this can change as well.

Not at all.

Avoiding --only is to prevent a much more dangerous glitch.

Suppose you and the other have two paths diverged, and one merges cleanly and the other results in conflict. When "git merge" gives control back to you, the cleanly merged result is ALREADY IN THE INDEX.

Now you futz with the other path, and say

	git commit --only other

What --only tells git is "I do not care what I've staged in the index. Start from the contents of HEAD commit, and update the index entry at these paths (and these path _ONLY_), and commit the contents registered in the index.

That is why --include is the only sane semantics during a conflicted merge. I thought you should know better, as you were the one who gave the explanation to Nathan, which triggered Nana's response, which resulted in this series.