On Mon 23 Oct 2006 at 23:11:59 +0000, Jack Cuyler wrote:
> On Mon, 23 Oct 2006 11:13:38 +0000, Duncan wrote:
>
> > "Jack Cuyler"
> > <[EMAIL PROTECTED]> posted
> > [EMAIL PROTECTED], excerpted below, on Sun, 08 Oct 2006 19:13:43
> > +0000:
> >
> >> +TMP=${TMP:-/tmp}
> >> +TEMPFILE=${TMP}/panpost.$(date +%Y%m%d%H%M%S%N)
> I wasn't aware of the mktmp command when I posted, but the date command in
> the variable name will take care of any race conditions, unless two
> instances are run at the same nanosecond. It's possible, I guess, but
> highly unlikely.
IF the time is precise to the nanosecond, which you can't count on. You
should count on no better than 1-second resolution, %N is probably some
Linux extension (repeat after me: All the world is NOT Linux - funny how
that used to be in the past All the world is NOT a VAX and now I have a
VAX running Ultrix and NetBSD). But if you combine the process id ($$)
with second-resolution time, you should be safe.
-Olaf.
--
___ Olaf 'Rhialto' Seibert -- You author it, and I'll reader it.
\X/ rhialto/at/xs4all.nl -- Cetero censeo "authored" delendum esse.
_______________________________________________
Pan-users mailing list
Pan-users@nongnu.org
http://lists.nongnu.org/mailman/listinfo/pan-users
