Re: [PATCH] libreadline buildworld breakage.

On Thu, Oct 21, 2004 at 12:16:08PM +0200, Ian FREISLICH wrote:
> Andrey Chernov wrote:
> > On Thu, Oct 21, 2004 at 01:33:07PM +0400, Andrey Chernov wrote:
> > > It is old one. New one should be 1.1.1.9, without any $FreeBSD$ tag
> > > Can you compare "cvs co contrib_libreadline" with your contrib/libreadline 
> > > contents?
> > 
> > I just check by cvsupping from another machine, all looks right:
> > 
> > ...
> > Edit src/contrib/libreadline/complete.c
> >   Add delta 1.1.1.9 2004.10.18.07.02.39 ache
> > ...
> 
> I was going to ask in the other mail if some repo surgery had taken
> place.  The last time this happened to me, someone had perpetrated
> a repo surgury and something about when I and the mirror synced
> broke things in a way that screwed me (and I think one or two others)
> over.
> 
Yes, the repo-surgery was done some time ago, and it looks like you
were hit by it.

> [brane-dead] /usr/src/contrib # diff -ud ~ianf/complete.c,v /home/ncvs/src/contrib/libreadline/complete.c,v 
> --- /home/ianf/complete.c,v     Mon Oct 18 09:02:43 2004
> +++ /home/ncvs/src/contrib/libreadline/complete.c,v     Mon Oct 18 09:02:43 2004
> _at__at_ -1,4 +1,5 _at__at_
>  head   1.9;
> +branch 1.1.1;
>  access;
>  symbols
>         v5_0:1.1.1.9
> 
> This is the diff of the ,v file.  The one in ~ianf was moved out
> of CVSROOT after a cvsup that changed nothing.  An immediate
> subsequent cvsup produced a ,v file with this difference.
> 
This explains it.  The surgery returned this file on a vendor branch,
adding this "branch 1.1.1;" line to an RCS file.  Are you by chance
ALWAYS using the -s option to cvsup(1)?


Cheers,
-- 
Ruslan Ermilov
ru_at_FreeBSD.org
FreeBSD committer