So yeah came across this while browsing why's site. Very cool. I hadn't been put into a situation where I needed to match the same thing on multiple lines (what with the nature of directory based information). But still this is damn cool. Give a regular expression to a string#scan and get back all the matching elements.
Means I can slurp in an entire file and pick out the bits I want. Memory overhead goes down the drain, but still for quick snatch-and-grabs, that has be the coolest thing.
re = /pattern./