[patch] zip tests

[Paul Smith]

On Wed, 2008-05-07 at 09:53 +0200, Cristian Ionescu-Idbohrn wrote:
> Good observation Paul.  Still.  Are 'true' and 'false' builtins in all
> shells?

> [ $xtrace = n ] || set -x

Well, they have just as much chance of being built-in as test/[ does, so
I don't think this improves the odds much.  If you wanted to be 100%
sure of only using builtins in all shells, you'd need something like:

        case $xtrace in n) set -x ;; esac

Personally I find the readability improvements I get from using
true/false in variables to more than make up for any tiny performance
gain on some systems.

Cheers!