On Sat, Mar 05, 2011 at 07:44:28PM +0100, Pascal Stumpf wrote:
> On Sat, Mar 05, 2011 at 03:59:16PM +0000, Jason McIntyre wrote:
> > On Sat, Mar 05, 2011 at 03:21:00PM +0100, Pascal Stumpf wrote:
> > > ls(1) still has the old semantics for -g in one spot.
> > > 
> > > Index: ls.1
> > > ===================================================================
> > > RCS file: /cvs/src/bin/ls/ls.1,v
> > > retrieving revision 1.63
> > > diff -u -r1.63 ls.1
> > > --- ls.1        4 Mar 2011 21:03:19 -0000       1.63
> > > +++ ls.1        5 Mar 2011 14:17:15 -0000
> > > @@ -275,9 +275,9 @@
> > >  is displayed for each file:
> > >  mode,
> > >  number of links,
> > > -owner,
> > > -group (though not for
> > > +owner (though not for
> > >  .Fl g ) ,
> > > +group,
> > >  size in bytes,
> > >  time of last modification
> > >  .Pq Dq mmm dd HH:MM ,
> > > 
> > 
> > fixed, thanks. but i'm confused - if originally -g requested group info
> > "be included", why did the man page say (of group info) "though not for
> > -g"? should that passage read "excluded" instead? anyone know?
> 
> Sorry -- after checking the commit logs it apparently got introduced
> right after -g was made POSIX-compliant (r1.49). It's neither correct
> for traditional BSD nor POSIX behaviour.
> 

so, can anyone clarify what -g orignally did, and whether the wording is
ok?

jmc

Reply via email to