On May 16, 5:24 pm, Marek Baczynski <[EMAIL PROTECTED]> wrote:
> On 16 Maj, 23:11, John <[EMAIL PROTECTED]> wrote:
>
> > Thanks.  You're right.  I just tried that out.
>
> > I actually have a deeper problem that I was trying to make a simpler
> > recreation.  This wasn't it.
>
> I had a lot of issues with autocrlf which were similar but not quite
> like yours. I'd experiment with adding autocrlf = false to the cloned
> repo's config.

Thanks.  That worked for me.

Out of the box, msysgit sets core.autocrlf = true in system scope: /
etc/gitconfig

I mostly work with linux and am only attempting git on windows because
I think git can solve some of our problems but many of my co-workers
are using windows.

So I guess with autocrlf = true, git checkouts and git status were
translating lf -> crlf (or crlf -> lf) inconsistently, also a bit
unpredictably.  I ran through clone-checkout-status cycle many times
and did not get the same number of 'changed' files every time.

Wish I had a cozier feeling about this.  Seems my co-workers could be
faced with a few issues regarding lf<=>crlf translations.  I think
I'll scan the forum for discussion.

Thanks again.

John

Reply via email to