From: Junio C Hamano <gitster@pobox.com>
Date: Tue, 27 Dec 2016 17:27:22 GMT
Subject: Re: git-apply: warn/fail on *changed* end of line (eol) *only*?
Message-ID: <xmqq8tr1cmf9.fsf@gitster.mtv.corp.google.com>
In-Reply-To: <xmqqvau7cqy1.fsf@gitster.mtv.corp.google.com>

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

> Imagine that the project wants LF line endings, i.e. it considers
> that a line with CRLF ending has an unwanted "whitespace" at the
> end.  Now, you start from this source file:
>
>     1 <CRLF>
>     3 <CRLF>
>     5 <CRLF>
>
> and a patch like this comes in:
>
>      1 <CRLF>
>     -3 <CRLF>
>     +three <CRLF>
>      5 <CRLF>
>
> You think that "3 <CRLF>" was replaced by "three <CRLF>", and the
> claim is "the 'previous' contents already had <CRLF> ending, so the
> change is not making things worse".

To see the problem with "check existing lines", it probably is
easier to extend the above example to start from a file with one
more line, like this:

    1 <CRLF>
    3 <CRLF>
    4 <LF>
    5 <CRLF>

and extend all the example patches to remove "4 <LF>" line as well,
where they remove "3 <CRLF>", making the first example patch like
so:

     1 <CRLF>
    -3 <CRLF>
    -4 <LF>
    +three <CRLF>
     5 <CRLF>

Now, if you take "three <CRLF>" to be replacing "3 <CRLF>", then you
may feel that not warning on the CRLF would be the right thing, but
there is no reason (other than the fact you, a human, understand
what 'three' means) to choose "3 <CRLF>" over "4 <LF>" as the
original.  If you take "three <CRLF>" to be replacing "4 <LF>", you
would need to warn.

A totally uninteresting special case is when the original is all
<CRLF>, but in that case, as you already said in the original
message, the project wants <CRLF> and you can configure it as such.
Then <CR> in the line-end <CRLF> won't be mistaken as if it is a
whitespace character <CR> at the end of a line terminated with <LF>.