From: Junio C Hamano <gitster@pobox.com>
Date: Fri, 22 Oct 2010 18:40:41 GMT
Subject: Re: [PATCH] Fix git-apply with -p greater than 1
Message-ID: <7vd3r2m5p2.fsf@alter.siamese.dyndns.org>
In-Reply-To: <20101022053140.GB786@burratino>

Jonathan Nieder <jrnieder@gmail.com> writes:

> Junio C Hamano wrote:
>
>> +test_expect_success 'apply (-p2) diff, mode change only' '
>> +	cat >patch.chmod <<-\EOF &&
>> +	diff --git a/sub/file1 b/sub/file1
>> +	old mode 100644
>> +	new mode 100755
>> +	EOF
>> +	chmod 644 file1 &&
>> +	git apply -p2 patch.chmod &&
>> +	test -x file1
>
> I had thought -p was only supposed to apply to traditional patches.

You are correct that historically it was invented for use with non-git
patches, but I didn't think of a good reason why we should forbid it
offhand.

> Maybe a documentation update would avoid confusion?
>
> 	-p<n>
> 	   Remove <n> leading slashes from traditional diff paths.
> 	   The default is 1.

If we have a good reason to forbid this option on git generated patches,
perhaps yes, and with either error/warning codepath, but first lets find
out a good reason why we should forbid it.  There may indeed be some.