If you have an other way of achieving the result I need, I'd be happy to learn about it.

use warnings;
use strict;
use XML::Twig;

my $x = '<map>
    <section1>
        <parameter>Some text</parameter>
        <rstate>CHANGE_THIS</rstate>
    </section1>
    <section2>
        <subsection>
            <parameter>dont change this</parameter>
            <rstate>CHANGE_THIS</rstate>
        </subsection>
        <subsection>
            <parameter>dont change this</parameter>
            <rstate>DONT CHANGE THIS</rstate>
        </subsection>        
        <subsection>
            <parameter>dont change this</parameter>
            <rstate>CHANGE_THIS</rstate>
        </subsection>
    </section2>
    <section3>
        <parameter>dont change this</parameter>
        <rstate>DONT CHANGE THIS</rstate>
    </section3>
</map>
';

my $t = XML::Twig->new(
    twig_handlers => { rstate => \&rstate },
    pretty_print  => 'indented',
);
$t->parse($x);
$t->print();

sub rstate {
    my ($t, $rstate) = @_;
    my $text = $rstate->text();
    $text =~ s/CHANGE_THIS/CHANGED/;
    $rstate->set_text($text);
}

__END__


<map>
  <section1>
    <parameter>Some text</parameter>
    <rstate>CHANGED</rstate>
  </section1>
  <section2>
    <subsection>
      <parameter>dont change this</parameter>
      <rstate>CHANGED</rstate>
    </subsection>
    <subsection>
      <parameter>dont change this</parameter>
      <rstate>DONT CHANGE THIS</rstate>
    </subsection>
    <subsection>
      <parameter>dont change this</parameter>
      <rstate>CHANGED</rstate>
    </subsection>
  </section2>
  <section3>
    <parameter>dont change this</parameter>
    <rstate>DONT CHANGE THIS</rstate>
  </section3>
</map>
[download]

Thanks for your reply! That would work using the parser, given all values need to be changed are the same? If they differ I would need several instances of the parser? If I'm not mistaken the parser approach builds around the idea that I already know the value, which is not my case. I need to read the old value, do some calculations and reinsert the new value. I realize my example was not optimal, sorry for that. In reality all values are not neccecarily the same. I still would like to know how this would be achieved using regexp. Thanks again, Pontus

You can do whatever calculations you like in the XML::Twig tag handler. Trying to do this with regexen is like hammering in a screw---you can do it but it's neither the easiest nor the most adequate way, so just don't.

I realize my example was not optimal

OK, so show us a better example of what you are trying to do, what you have tried, and how it doesn't work for you.

As toolic said, use an XML parser. But, in general terms, when I want to use a regex to change an unknown chunk of text that sits between two known chunks of text, I do it backwards from what you tried to do here. Instead of capturing the part I want to replace, I capture the parts I want to keep, and keep them. For instance:

$string = 'blah blah blah <some_sort_of_tag>CHANGE ME</some_sort_of_ta
+g> blah blah blah';
$newtext = 'I AM NEW';
$string =~ s|^(.*<some_sort_of_tag>).+(</some_sort_of_tag>.*)$|$1$newt
+ext$2|;
[download]

I have understood that using regex for this problem is not the preferred method, at all. It was however my intention to learn more about regexp, so for now, with your help I have come up with (using the backwards strategy)

# First match !
    $data =~ s|^(.*<section1>.*<rstate>).+(</rstate>.*</section1>.*)$|
+$1$newrstate$2|s;

# Second match!
    $data =~ s|^(.*<section2>.*<rstate>).+(</rstate>.*</subsection>.*<
+/subsection>.*</subsection>.*</section2>.*)$|$1$newrstate$2|s;

# Third match!
    $data =~ s|^(.*<section3>.*<rstate>).+(</rstate>.*</section3>.*)$|
+$1$newrstate$2|s;
[download]

Curious as I am, I will read up on the XML parser and try it as well to see how it compares. Thanks for the help!

I'm just curious as to why you're using the many if statements.
I suppose the following should work:

$data=~ s{<rstate>[^<\n]+</rstate>}{<rstate>$new_text</rstate>}g
[download]

Note the use of the 'g' modifier.
If you want to learn more about regular expressions in general, then Friedl's book is a must. It's one the most well-written books I've ever read on any subject and I highly recommend it.