Find all possible starting positions of a regular expression match in perl, including overlapping matches?

Question

Is there a way to find all possible start positions for a regex match in perl?

For example, if your regex was "aa" and the text was "aaaa", it would return 0, 1, and 2, instead of, say 0 and 2.

Obviously, you could just do something like return the first match, and then delete all characters up to and including that starting character, and perform another search, but I'm hoping for something more efficient.

Answer 1

Use lookahead:

$ perl -le 'print $-[0] while "aaaa" =~ /a(?=a)/g'

In general, put everything except the first character of the regex inside of the (?=...).