Perl regular expression performance

README.md

In response to an outage on Stack Overflow Rob Pike tweeted that regex parsing failure was avoidable and pointed toward a great post on regex complexity by Russ Cox. The version of perl in that post (5.8.7) is super ancient and in perl 5.10 the regex engine received a ton of optimization work and that optimization work has continued right through to the most recent release - 5.24.0.

Running the script above on 5.24.0 shows:

shadowfax:~ mallen$ time perl re.pl 
match

real	0m0.099s
user	0m0.026s
sys	0m0.039s

If you're looking to explore, debug or optimize your perl regular expressions, let me strongly recommend Damian Conway's RegExp::Debugger You can still make pathological regular expressions of course but it's much more difficult than the patterns given in Russ Cox's post.

re.pl

# tested on 5.24.0 and 5.10.1 - should be true for all versions >5.10 
use 5.024;

my $str = "a" x 100_000;
my $re = qr/a?a?a?aaa/;

if ( $str =~ $re ) {
    say "match";
} 
else {
    say "no match"
}