Question

What differences, if any, exist between the Kernel#at_exit method and the END (all upper case) keyword? Is the latter merely a more Perlish way of doing things, and the former more Ruby-esque?

I tried doing defined?(END {puts "Bye"}), but got a syntax error.

Answer 1

“The Ruby Programming Language” defines a minor difference in their behavior. at_exit can be called multiple times when within a loop and each iterated call will be executed when the code exits. END will only be called once when inside a loop.

…If an END statement is within a loop and is executed
more than once, then the code associated with it is still only registered once:

a = 4;
if (true)
  END { # This END is executed
  puts "if"; # This code is registered
  puts a # The variable is visible; prints "4"
}
else
  END { puts "else" } # This is not executed
end
10.times {END { puts "loop" }} # Only executed once

The Kernel method at_exit provides an alternative to the END statement; it registers a
block of code to be executed just before the interpreter exits. As with END blocks, the
code associated with the first at_exit call will be executed last. If the at_exit method
is called multiple times within a loop, then the block associated with it will be executed
multiple times when the interpreter exits.

So, running:

2.times {
  END { puts 'END'}
  at_exit { puts 'at_exit' }
}

results in:

at_exit
at_exit
END