in reply to Re^2: Use of uninitialized value in substitution (s///)
in thread Use of uninitialized value in substitution (s///)

Happened to me just a few hours ago :)

That really ought to be a warning, I can't imagine a use case where you want the bitwise negation of the number of substitutions  $foo = ~s/// or matches  $foo = ~m//

every time I've encountered this in the past, its always been a mistake

warnings should issue a specific warning