Given the incredibly vague problem statement I'll just toss out that this looks awfully like you're trying to parse some variant of XML in which case unless you have a very restricted input that you can guarantee follows a very static format trying to use regexen to parse it is probably going to cause more headaches than it's worth. Use a proper parser (e.g. XML::Twig, XML::Simple) and save yourself the trouble.

Update: And see XML::Twig and Processing Instructions for getting, erm, XML::Twig to handle processing instructions.

To help you with your question, you'll need to give us a bit more detail:

• Where are you searching for this pattern? In a file? In a string stored in a variable?
• When you find this pattern what information do you need to keep about it? Its position in a string/file? The number after "test"?

In the meantime, you may find perlretut helpful. It explains how to use regular expressions to search for patterns in strings.

Best, beth

Maybe what you want is a character set or a non-greedy match. You can read about both of these in perlre.

use strict;
use warnings;
use Data::Dumper;

my $string = '<?test=6565?><?test=6566?>xxx<?test=6567?>';
my @array1 = $string =~ m/(<[^>]*>)/g;
my @array2 = $string =~ m/(<.*?>)/g;

print Dumper(\@array1);
print Dumper(\@array2);
__END__
$VAR1 = [
          '<?test=6565?>',
          '<?test=6566?>',
          '<?test=6567?>'
        ];
$VAR1 = [
          '<?test=6565?>',
          '<?test=6566?>',
          '<?test=6567?>'
        ];
[download]

update: If you need the work "test" to be in the matched string, then something like this:

use strict;
use warnings;
use Data::Dumper;

my $string = '<?test=6565?><?some=6565?><?test=6566?><?other=6566?><?t
+est=6567?><?data=6567?><?test=6568?>xxx<?test=6569?>';
my @array1 = $string =~ m/(<[^>]*test[^>]*>)/g;

print Dumper(\@array1);
__END__
$VAR1 = [
          '<?test=6565?>',
          '<?test=6566?>',
          '<?test=6567?>',
          '<?test=6568?>',
          '<?test=6569?>'
        ];
[download]