Rather than regexes, you could also use the index, split and substr built-ins, as shown in the following session under the debugger:

  DB<1> $str = "this is a test";

  DB<2> $sub = "is a test";

  DB<3> $index = index $str, $sub;

  DB<4> $nr_before = scalar split /\s+/,  substr $str, 0, $index;

  DB<5> p $nr_before;
1
  DB<6> $nr_in = scalar split /\s+/, $sub ;

  DB<7> p $nr_in;
3
  DB<8> print ++$nr_before, " " for 1..$nr_in;
2 3 4
[download]

Note: I know that the call to the scalar built-in is not really useful in a scalar context, but I thought it would clarify the idea.