Re: [PATCH/RFC v2 1/4] Add "core.eolStyle" variable to control end-of-line conversion

From: Linus Torvalds <torvalds@linux-foundation.org>
Date: May 9, 2010, 18:11 UTC
Message-ID: <alpine.LFD.2.00.1005091046570.3711@i5.linux-foundation.org>
In-Reply-To: <A5422145-63E1-4AF4-9184-7A6D15E9C2B6@gmail.com>

On Sun, 9 May 2010, Eyvind Bernhardsen wrote:

Show 5 quoted lines

> 
> I'm sorry.  Are you the same Linus Torvalds who wrote this:
> 
> > Btw, since we're discussing this, I do think that our current "crlf=input" 
> > syntax for .gitattributes is pretty dubious.

Yes, it's dubious. But as with the kernel, we need to support backwards compatibility for things that have reasonably been used (and "input" has).

I really brought it up as an example of things that weren't necessarily all that well designed.

That said, it looks like people actually do want per-file line-ending settings, ie not just a global "I want CRLF vs LF". So it looks like crlf=input is actually useful in a .gitattributes files, if only because some people seem to want to mix CRLF and just LF in the same repository.

It also sounds like people actually want to have the reverse (ie not just "input", but have a mode where LF may be the default, but then some particular files must always be CRLF even if most files are normal text).

So I suspect we want to really have support for all four combinations _both_ in the .git/config file, _and_ in the .gitattributes file.

The four cases would be "none" ("binary" or "-crlf"), "lf" ("input" or "crlf=input"), "system default", and "force crlf".

Honestly, I would personally have preferred to have just a repo-wide "this is the line ending". Not some path-specific endings like "crlf=input" in the .gitattributes. But people do seem to want it.

			Linus