| Next |  | 22 | 
The interesting part is:
        /^\(.\)\1\1*$/ { x; 
                         y/abcdefghijklmnopqrstuvwxyz/ABCDEFGHIJKLMNOPQRSTUVWXYZ/;
                                      x; d;
                       }
/^\(.\)\1\1*$/ is a pattern
it matches lines that have 2 or more occurrences of the same character
and nothing else
For these lines, we first retrieve the hold space (x)
(that's the previous line of input)
Translate to upper case (y/.../.../)
Put it back in the hold space as if it had been uppercase all along
And skip to the next line of input, printing nothing d
| Next |  | Next |