On Tue, 27 Oct 1998 00:26:16 -0500  "Perry E. Metzger" wrote:

> Simon Burge writes:
> > > shouldn't this simply be consistant with ls(1)?
> > 
> > Yes and no - I looked at this, but ls(1) has some smarts to autosize
> > columns which is really not practical in find(1) without parsing the
> > whole tree and then redisplaying everything.  So I just went for column
> > widths that suit the types of files we have around here...
> 
> I mention it only because one of my big pet peves is ls(1) and 
> find -ls being out of sync. Maybe we should libraryize the ls(1) code
> or something? I'd really prefer to see these consistant...

You'd still have the problem of find(1) having to parse everything
before it starts printing anything.  If you really want consistancy
between ls(1) and find -ls, it'd probably be easier to add a fixed-field
output type to ls and defaults to the options "-dgils" (that find(1)
uses).  There's still a few unused letters in ls(1)'s option list :-)

If you really have something that wants to rely on the consistancy of
output styles, simple post-processing with awk(1) is probably your best
bet..

Simon.

> > > Simon Burge writes:
> > > > Module Name:	src
> > > > Committed By:	simonb
> > > > Date:		Tue Oct 27 04:53:03 UTC 1998
> > > > 
> > > > Modified Files:
> > > > 	src/usr.bin/find: ls.c
> > > > Log Message:
> > > > For -ls output, change field widths to allow inodes up to 9,999,999 and
> > > > file sizes up to 999,999,999 bytes (and 999,999 blocks) without wobbly
> > > > lines.  Also change device minor/major to be 3/5 digits (current maximum
> > > > is 4/7 digits - the 3/5 split is arbitary).