On Mon, Dec 24, 2001 at 03:09:29AM -0800, Jeremy C. Reed wrote:
> On Sun, 23 Dec 2001, Eric Haszlakiewicz wrote:
> 
> > 	I remember a while ago someone saying that a stat command might be
> > useful.
> 
> I started writing one myself, but I never could decide on the format it
> should output. I think I saw two others mentioned also.
	I used the order of the stat structure with each column separated by
tabs.  (since some fields, like the date, contain spaces)  After I wrote it
I found the linux stat command (but not any others).  Perhaps it would be
useful to match that.

> > I didn't see anyone say they were writing one, so I did.  I'm not done with
> 
> Can you send some usage examples and output?
Sure.  I tried to make the short options as obvious as possible.  Long options
(--foo) use a field name from the stat structure.  The order of the columns is
the order of the stat structure:

Usage: stat [-LqQansbftugl] [-mA | -m[u][g][o][s]] [-TA | -T[a][m][c]] file [...]
       stat [-LqQ] [--fieldname] file [...]

So very quiet output using lstat to get blocks used, uid and user modes would be:
stat -LQbu -mu foo
resulting in:
rw-     erh     120288

Without -Q:
     MODE     UID     BLOCKS
foo: rw-      erh     120288

Or add a -n to get numerical output: (stat -nLbu -mu foo)
0x8184          101     120288
(whoops, didn't mask that mode right.  like I said, not finished yet)

default output: (stat foo)
     TYPE     MODE            UID     GID     SIZE
foo: REG      rw-r--r--       erh     wheel   61559863

Or, specify fields of the stat structure: (stat --st_nlink --st_size foo)
     LINKS    SIZE
foo:     1    61559863

> Since, it didn't build for me:
> stat.c:317: warning: implicit declaration of function `getprogname'
> (I guess I should have tried this under a newer NetBSD, especially since
> this was posted to the current list.)
	uh, yeah. :) Hasn't getprogname() been present for a while now?  I'm running
current as of Dec 1 and as of yet I haven't tried anything earlier.

eric