cvs commit: src Makefile.inc1

[Alexey Dokuchaev]

On Mon, Dec 15, 2003 at 07:37:04PM +1100, Peter Jeremy wrote:
> On Sun, Dec 14, 2003 at 11:36:02AM +0000, Mark Murray wrote:
> >Dag-Erling =?iso-8859-1?q?Sm=F8rgrav?= writes:
> >> It uses Perl directly to compute the numeric mode of a file.  The
> >> version in -CURRENT uses a combination of stat(1), bc(1) and shell
> >> arithmetic which causes a syntax error in 4.x.
> >> 
> >> Personally, I think it would be best if stat(1) could gain an option
> >> (-m perhaps) that makes it simply print its arguments' modes in octal.
> >
> >Does 4.x's stat(1) have the "-s" switch? If so:
> >
> >$ ( eval $(stat -s .profile) ; printf "%o %s\n" ${st_mode} ${st_mode} )
> >100644 0100644
> >
> >Would that be any use?
> 
> Not in the absence of stat(1) :-(.  As an alternative, maybe ls(1)
> could grow an option to spit out the mode in octal - assuming there
> are any spare opton letters left.  The simplest solution would seem
> to be to MFC stat(1) to 4.x.  The only other alternative would be a
> (messy) awk script to convert the mode letters in 'ls -l' output to
> an octal number.

Frankly, adding an option to ls(1) or writing ls(1) -l/awk(1) combo
takes my vote, rather than adding yet another foo(1) utility to the
base; especially provided that its functionality isn't strictly
orthogonal.

Just my $.02.

./danfe