You can specify alternatives, and use the
(?{}) construct to keep track of how many times those alternatives were required. You have to keep track of how when you backtrack, so you don't count those multiple times (if your regex has that possibility).
The alternatives might allow the regex engine to match less aggressively than you think it ought to, so you could end up with a higher miss rate than you'd like. In the code below, you get an array of the checkpoints that it had to use the alternative. Play with different values of $str.
use strict;
my $str = 'PolaBexar'; # Missing letter, extra letter, two capitals
my @checkpoints = ();
sub stack_checkpoints {
my $val = shift;
# Remove all checkpoint markers higher than $val
@checkpoints = grep { $val>$_ } @checkpoints;
print "Backtracking to $val\n";
push @checkpoints, $val;
}
print "Matched <$&> with ".@checkpoints." misses\n" if
$str =~ /(?:p|.{0,1}?(?{stack_checkpoints 0}))
(?:o|.{0,1}?(?{stack_checkpoints 1}))
(?:l|.{0,1}?(?{stack_checkpoints 2}))
(?:a|.{0,1}?(?{stack_checkpoints 3}))
(?:r|.{0,1}?(?{stack_checkpoints 4}))
(?:b|.{0,1}?(?{stack_checkpoints 5}))
(?:e|.{0,1}?(?{stack_checkpoints 6}))
(?:a|.{0,1}?(?{stack_checkpoints 7}))
(?:r|.{0,1}?(?{stack_checkpoints 8}))
/x;
(quick update: made the alternatives non-greedy)
We're not really tightening our belts, it just feels that way because we're getting fatter.