Subject: Re: ln.diff for bin/19431
To: Jason R. Fink <jrf@adresearch.com>
From: Andy Isaacson <adi@hexapodia.org>
List: tech-userlevel
Date: 12/19/2002 18:07:09
On Thu, Dec 19, 2002 at 06:41:30PM -0500, Jason R. Fink wrote:
> On Thu, Dec 19, 2002 at 11:57:19PM +0100, Klaus Klein wrote:
> > "Jason R. Fink" <jrf@adresearch.com> writes:
> > > +	if (vflag)
> > > +		(void)printf("%s %c> %s\n", source, linkch, target);
> > 
> > Now you've made me curious. :-) I thought the whole point was to
> > resemble the behaviour of GNU fileutils as closely as possible (or
> > desirable?); fileutils ln prints "create {hard,symbolic} link '%s' to
> > '%s'".  Is there prior art using the format you've chosen?
> 
> No, I wanted to use FreeBSD. That is the same reason I kept the
> above, also, cp uses ->. So I wanted to have consistency.
> 
> The standards blurb is in there, as soon as my new checkout is
> done I will generate another diff and send a copy of FreeBSD's
> ln. Now, this does not mean I couldn't go back and change
> cp and ln to say "to", but I would prefer consistancy within
> the BSDs and not with GNU.

From GNU:
% ln -v 29.ps foo.ps
create hard link `foo.ps' to `29.ps'
% cp -v foo.ps zot
`foo.ps' -> `zot'
% cp --version
cp (GNU fileutils) 4.0.36
...

I think both of those are silly for using a grave accent as a quote,
because that practice is deprecated by the Unicode folks.  The ->
behavior seems far more reasonable than the 'create %s link to'
behavior.  I think Jason's patch is right, but perhaps the quotes would
be a good idea.

-andy