The classic solution of this problem was invented in FORTRAN in early 50 -- it is a backslash at the end of the line.
And, again... exactly how much existing Perl 5 code has backslashes at the end of lines (commented or otherwise)? Approximately none. Boom. It breaks.
Question to you: how many times you corrected missing semicolon in your Perl scripts the last week ?
That's an easy one: Zero.
Which isn't to claim that I never forget semicolons, but I generally only get "possible missing semicolon"-type errors from the compiler once every month or two, and the compiler is often wrong about them. (There's still an error in my code, of course, but it's a different error, rather than the missing semicolon that the compiler suspected.)
Edit: Come to think of it, that closing parenthetical is an even bigger problem with this suggestion. If the compiler thinks I have a missing semicolon, but the real error is actually something else... then that means that, if the compiler says "hey, no problem, I'll just implicitly add that semicolon for you", then the actual error will still be there, undetected, waiting to strike at run-time when I least expect it. That is not an improvement!
