Hi, I'd like to suggest a simple approach - It's more easy to change/understand if you read the code some time later.

#!/usr/bin/perl -w
use strict;

my ($hi_from,$hi_to,$bye_from,$bye_to);


while ( my $line = <DATA> ){
        $line =~ /^(.?)\D*(\d*)\D*(\d*)/;
        next if ( !$1 ); # Empty line

#        print "1: $1  2: $2   3: $3\n";


        if ( $1 eq 'h' ){
                if ( !defined($hi_from) || $hi_from >= $2 ){
                        $hi_from = $2;
                }
                if ( !defined($hi_to) || $hi_to <= $3 ){
                        $hi_to = $3;
                }
        }

            if ( $1 eq 'b' ){
                if ( !defined($bye_from) || $bye_from >= $2 ){
                        $bye_from = $2;
                }
                if ( !defined($bye_to) || $bye_to <= $3 ){
                        $bye_to = $3;
                }
        }
        
}

print "hi_from: $hi_from     hi_to:  $hi_to\n";
print "bye_from: $bye_from     bye_to:  $bye_to\n";


__DATA__
hi:  65 abcdefghijklmnopqrst 85
bye: 12 bcdefghijklmnopqrstu 32

hi:  26 sagfsdgsgwsehbbdgops 106
bye: 33 afasdfdfafasaafadfad 53

hi:  89 sagfsdgsgwsehbbdgops 188
bye: 37 afasdfdfafasaafadfad 89
[download]

It would be possible to do some cool constructs with map, hashes, ...
However, if you have to change something later, something like this is much more easy to understand and debug. (Although unelegant ).

It's also a matter of the needs you have.

• Which size could the file have ?
• Is speed important ?
• Do you need some warnings if the file is misformed ?
• ...

HTH, Michael