I'm trying to write a korn shell script and I'm at my wits end as to how to solve it!

The problem I have is that I'm trying to parse a file and edit specific parts of the file.

For example, I have 4 re-occurring lines of data with 2 variables I'm interested in:

blah blah blah VAR1 blah
.......................
......................
blah blah blah blah VAR2

I need the script to match VAR1 and VAR2 to specified critera, which when matched - I then need to sed VAR2 with VAR3

e.g

I live in a town
............
............
and I drive a car
I live in a town
............
............
and I drive a truck

if town is found in line 1, and truck is found in line 4, then replace truck with lorry

I've tried everything I can think of using grep and gnu grep, piping into sed, but can't solve the problem - please help!

I know you guys are proper clever - all help is much appreciated =D

Can the word town be anywhere on the 1st line and the word truck anywhere on the 4th?

If position of the words doesnt matter, this will find the word truck 3 lines after the word town and replace it with lorry:

grep -A3 town infile|tail -1|grep truck|sed -e "s/truck/lorry/g"

Unfortunately this doesn't seem to work for me, the file isn't edited =(.

To make things clearer, I'm using korn on Solaris 9 to script, and here is an example of the actual text in a file I'm looking to edit:

<ManagedElementId string="$VAR1"/>
<primaryType type="xxx"/>
<managedElementType types=""/>
<associatedSite string="Site=$VAR2"/>

I want to leave the file exactly as it is, except for when VAR1 and VAR2 are found in this scenario and then I wish to substitute VAR2 with VAR3