0

I have the following code:

debug=$?
function a {
    su - javi -c "uptime"
    return $debug
}

function b {
    su - javi -c "cat /etc/redhat-release"
    return $debug
}

function c {
    su - javi -c "cat /etc/redhat-release"
    return $debug
}

case $debug in
0)
    a
    if [ $debug == 0 ]; then
        b
        echo "se ejcuta la funcion"
    elif [ $debug == 0 ]
        c   
    elif[].... <-----this

    fi
;;

1)
    echo "se ha producido un error"
;;

esac

Is there any way to debug by removing the if ??I want them to go running a function if it ends well that jumps to the other function and if it does not end well that it leaves the escript, that with 5 functions

  • ìf false ; thenfollowed by # if? – Archemar Feb 6 at 19:39
  • 1
    You also don't need the su operation to run uptime or to read /etc/redhat-release. And why do you return $debug from your functions when it's a global variable, and you ignore the return value anyway? – roaima Feb 6 at 19:48
  • To do what you want, you need to reset $debug with the result of the previous command. Since you're not doing that it looks like all your tests are measuring the initial value of $debug, which incidentally, is $?... not sure what that would be in a new subshell. is it just the value of the last command run? – Tim Kennedy Feb 6 at 19:48
  • 1
    @Tim yes it is. So here Ortiga could run false; ./myscript.sh to enable debug or true; ./myscript.sh to keep it off. But it would also get triggered with two commands such as ls /does/not/exist then ./myscript.sh – roaima Feb 6 at 19:49
  • 1
    Ortiga, you probably want to run this through shellcheck.net You're missing space around the [ and ] characters, for starters. And == is a string comparison. Use -eq for a numeric one. – roaima Feb 6 at 19:51
0

If the goal is simply to run each function, and bail out of there was any errors, maybe something like this will work:

function bail {
    echo "se ha producido un error ($1)"
    exit 1
}

function a {
    uptime
}

function b {
    cat /etc/redhat-release
}

function c { 
    cat /etc/redhat-release
}

for f in 'a' 'b' 'c'; do
    $f || bail "$f: $?"
done

This will run each function, and if an error is encountered, the function name and exit code are sent to the bail function, which will print a line and bail out of any additional commands.

I did remove the use of su because it was easier to test that way. If you want to automate this, I'd recommend creating a sudo profile to run those commands with elevated privileges in such a way that you only need to enter your password once, or no times at all.

0
  1. debug=$? is not evaluated at every command. So the value of debug will be the same throughout the entire script.
  2. The return code of a function is the return code of its last command. So the return statements are pointless.

In summary, something like a && b && c will run b if a succeeds, and will run c if b succeeds. The return code of this compound command is the first non-zero return code of each of the functions.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.