Re: [Quilt-dev] Re: Link check after pop really needed? (was: Another shell re-write of backup-files)

[Jean Delvare]

On Monday 21 March 2011 09:39:18 pm Jean Delvare wrote:
> Hi Andreas,
> 
> On Monday 21 March 2011 08:48:19 pm Andreas Gruenbacher wrote:
> > Jean,
> >
> > On Monday 21 March 2011 15:39:35 Jean Delvare wrote:
> > > So I start wondering if this check is still needed today. It
> > > could as well be that the internal state representation made it
> > > required back then, and no longer.
> >
> > the meta-data format hasn't changed in the meantime, so it can't be
> >  that.  I think that it was a bug where GNU patch did a rename("f",
> >  "g") with "f" and "g" being hardlinks (which is defined to do
> >  nothing!), and it didn't check for hardlinks.  Should be fixed in
> >  GNU patch since quite some time.  I cannot reproduce the problem.
> 
> Thanks for the quick reply.
> 
> What I don't get is: if the bug was in GNU patch, the problem would
>  have happened upon applying the problematic patch, i.e. with "quilt
>  push", without waiting for the user to "quilt pop" down to the patch
>  in question, wouldn't it?
> 
> I'll see if I can reproduce the problem with an older version of GNU
> patch. Maybe asking for a not too ancient version of GNU patch would
>  be the way to go.

I tried with GNU patch 2.4 and 2.5 but couldn't reproduce the problem. 
So either the bug was in even older versions of GNU patch (and we don't 
care) or it is very hard to trigger.

For now, I am inclined to let quilt (actually, backup-files) complain 
loudly if the problem is detected, and ask the user to report if/when it 
happens, so that we can investigate. And if nobody complains in the year 
following the next release of quilt, declare the problem solved and drop 
the useless code. Objection?

-- 
Jean Delvare
Suse L3