Re: Pattern Finding

[ariels]

(In perl -de 17:)

  DB<6> $x = 'helloworldhellohellohihellohiworldhi'

  DB<7> print "$1\n" while $x =~ /(\w+)(?=.*\1.*\1)/g
hello
o
l
hello
h
l
l
o
hi
h
l

The single letters appear, of course, since they're repeated more than once. We can fix that by requiring \1 to be at least 2 characters long...

  DB<8> print "$1\n" while $x =~ /(\w{2,})(?=.*\1.*\1)/g
hello
hello
hi

Of course it still finds `world', which appears 4 times.

To autogenerate the regexp:

sub repeat_finder_re {
  my $n = shift;
  '(\w{2,})(?=' . ('.*\\1'x($n-1)) . ')'
}