Subject: Re: Humanization of ls(1)
To: Bruce J.A. Nourish <bjan+tech-userlevel@bjan.net>
From: David P. Reese Jr. <daver@gomerbud.com>
List: tech-userlevel
Date: 12/28/2003 17:28:40
On Sun, Dec 28, 2003 at 11:03:52AM -0700, Bruce J.A. Nourish wrote:
> On Sun, Dec 28, 2003 at 01:22:28AM -0800, David P. Reese Jr. wrote:
> > On Sat, Dec 27, 2003 at 05:25:25PM -0700, Bruce J.A. Nourish wrote:
> > > On Sun, Dec 28, 2003 at 01:02:58AM +0100, mouss wrote:
> > > > >On Sat, 20 Dec 2003, David P. Reese Jr. wrote:
> > > > >
> > > > > >  When using '-sh', should the file sizes be printed in bytes or
> > > > > >  blocks?
> > > >
> > > > '-s' is for number of blocks, so it should print number of blocks
> > > > (in 512 bytes by default, but since -h is an addition, 1024 may be an 
> > > 
> > > GNU ls prints human readable sizes when -s and -h are on the same command
> > > line. That may be wrong, but if we did it, at least we'd be compatible 
> > > with someone.
> > 
> > SUSv3 says that '-s' is non portable and may be removed from the standard
> > in the future.  I'm not very excited about reading file sizes in gigablocks.
> 
> "-s" is already in NetBSD. The question is how "-h" interacts with it.

I am talking about '-s' and '-h' interaction.  SUSv3 says that '-s' is non
portable due to differences in block sizes.  Thus I don't feel too bad if
'-h' radically effects the behavior of '-s'.

-- 
   David P. Reese, Jr.                                  daver at gomerbud.com
                                               http://www.gomerbud.com/daver/