I just removed the .gitattributes file. Clearly it isn't worth the trouble.
On Fri, Aug 7, 2009 at 10:19 PM, Matt Birkholz<[email protected]> wrote: >> From: Taylor R Campbell <[email protected]> >> Date: Thu, 6 Aug 2009 21:15:02 -0400 >> >> Why does `git diff' show these changes, and how can I make it stop? I >> haven't touched the old RCS keyword lines, but `git pull' doesn't fix >> them either. > > I pondered similar nonsense aloud, on this list, back in June. I > eventually found that creating a .git/info/attributes file containing > the line: "* -ident" allowed me to use master again. The problem only > affects heads like master that contain the new .gitattributes file. > > You may have to "git reset" afterward. ? > > The problem is that the "ident" file attribute causes git-diff to see > "$Id$" where the working file says "$Id: <mumble>$". Unfortunately > the repo's version is not similarly filtered. (Git assumes you had no > $Id$ tag in a file when you first set the ident attribute?) So you > can check out "$Id: <mumble> $", but the working file will not match! > > I believe we need to check in (continue checking in?) diffs like these > > > -$Id: bkpt.h,v 9.40 2008/01/30 20:02:11 cph Exp $ > > +$Id$ > > until there are no "$Id: ... $" tags left -- until they are all > checked in with "$Id$". > > > _______________________________________________ > MIT-Scheme-devel mailing list > [email protected] > http://lists.gnu.org/mailman/listinfo/mit-scheme-devel > _______________________________________________ MIT-Scheme-devel mailing list [email protected] http://lists.gnu.org/mailman/listinfo/mit-scheme-devel
