> I disagree. This only true if you use [t]csh, and are calling it in
> the recommended way ( in fact, it uses source from the bash shell
> codebase to extract user info ).
The behaviour of which even on Linux systems used to be as I described,
it has always been a stand-alone binary, and the way it traditionally
worked is no good for giving you a reliable answer to the question for
which you are calling it.
You are correct in that these days, which is part of the util-linux
package, and functions in a way that actually works: getenv("PATH") and
a stat() on all the directories and all the argv.
There's still no reason not to use the internal function of your
interactive shell, which gets its answer straight from the horse's mouth
instead of an external program whos reliability you would have to
establish first.
Volker
--
Volker Kuhlmann is list0570 with the domain in header
http://volker.dnsalias.net/ Please do not CC list postings to me.
