Date: Fri, 25 Mar 2005 10:33:30 +0100 From: "David J. Weller-Fahy" <dave-lists-freebsd-questions@weller-fahy.com> To: freebsd-questions@freebsd.org Subject: Re: Scripting oddness (sh) - SOLVED (kind of) Message-ID: <20050325093308.GB10950@weller-fahy.com> In-Reply-To: <20050323223723.GC75716@weller-fahy.com> References: <20050323193626.GH7474@weller-fahy.com> <20050323193716.GI7474@weller-fahy.com> <4241EDB9.5000107@toldme.com> <20050323223723.GC75716@weller-fahy.com>
next in thread | previous in thread | raw e-mail | index | archive | help
After some experimenting I believe that I've discovered how to fix the problem, and the reason for it - I'd been thinking of the '&&' shortcut as functionally identical to the if-then-fi construct. That's obviously (now ;) not the case. If, as in the previous message, the following '&&' shortcut is used: #v+ [ -z "$RB_TEMP" ] && PKGLIST="$PKGLIST $PKG" #v- Then, if RB_TEMP is not of zero length, the exit status of that command is one. Since that command is the last performed in the list_required_by function, the exit status of the function is one. Because I used 'set -e' at the top of the script, any command (a function being a complex command) that exits with a value of one halts the execution of the script. However, if we change that line to: #v+ if [ -z "$RB_TEMP" ] ; then PKGLIST="$PKGLIST $PKG" fi #v- Then, following the fi, the exit status is zero for both the command and the function, and all following commands are executed. The following also works: #v+ [ ! -z "$RB_TEMP" ] || PKGLIST="$PKGLIST $PKG" #v- Because the exit status of the first command on the line is zero if RB_TEMP is not zero length. So, there's a fix for it. Figured I'd post this FTR, just in case someone else has a lack of brain bytes similar to mine. ;] Regards, -- dave [ please don't CC me ]
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050325093308.GB10950>