Sed help?

sup /prog/, I need your help for a bash script I'm making.
The last thing I need is a 3-step line.
>1. find the first instance of one less than an existing variable ($X - 1)
>2. search for the first instance of a string above that
>3. take the line of step 2 and save it as a variable

How could I do this using sed/grep?
Thanks.

This is impossible using sed or grep alone.  You might be able to hack together some abomination with awk and sh, but you'd probably be better off just using Perl.

>>2
You can do this with sed and expr or shell arithmetic.
instance of a string above that
If you mean the string is on the previous line, you would need to buffer the sed input with N and hold two lines at once. If a string can span multiple lines its slightly more difficult. You can put a flag in the hold space to determine whether you're in a string or not.
http://sed.sf.net/ has a lot of good resources.

OP here. I can do step 1 and 3 already, it's just step 2 that I'm having trouble with.

#step one
X="15"
Y=$(($X - 1))
Z=$(sed -n '/'$Y'/{=;q;}' file)

#step three
A=$(echo steptwo)

I did it /prog/.

X="15"
Y=$(($X - 1))
Z=$(sed -n '/lol/{=;q;}' file)
T=true

while [ $T == true ]; do
        Z=$(($Z - 1))
        A=$(sed $Z'q;d' file)
        if [ $A == "<item>" ]; then
                T=false
        fi
done
echo $Z

bampu pantsu